TECHNICAL SESSIONS Monday, 9:00-10:20 - Euro 2010 Lisbon

Monday, 9:00-10:20 

� MA-01 


Aula Magna 

Keynote Talk 1 

Stream: Keynote Speakers 

Invited session 

Chair: Gerhard-Wilhelm Weber, Institute of Applied Mathematics, 

Middle East Technical University, ODTÜ, 06531, Ankara, Turkey, 

gweber@metu.edu.tr 

1 - Safe tractable approximations of chance constraints 

Arkadi Nemirovski, School of Industrial and Systems 

Engineering, Georgia Institute of Technology, 765 Ferst Drive, 

NW, GA 30332-0205, Atlanta, Germany, 

nemirovs@isye.gatech.edu 

When optimizing under stochastic uncertainty, the entity of primary importance 

is a chance constraint Prob qsi->P f(x;qsi) in Q >= 1 - epsilon, for all P in PP 

where x is the decision vector, qsi is a random perturbation with distribution P 

known to belong to a given family PP, Q is a given target set, and epsilon « 1 

is a given tolerance. Aside of a handful of special cases, chance constrains are 

computationally intractable: first, it is difficult to check efficiently whether the 

constraint is satisfied at a given x, and second, the feasible set of a chance constraint 

typically is nonconvex, which makes it problematic to optimize under 

the constraint. Given these difficulties, a natural way to process a chance constraint 

is to replace it with its safe tractable approximation a tractable convex 

constraint with the feasible set contained in the one of the chance constraint. 

In the talk, we overview some recent results in this direction, with emphasis on 

chance versions of well-structured convex constraints (primarily, affinely perturbed 

scalar linear and linear matrix inequalities) and establish links between 

this topic and Robust Optimization. 

� MA-02 


3.2.14 

Combinatorial Optimization I 

Stream: Combinatorial Optimization 


Chair: Ivana Ljubic, Department of Statistics and Decision Support 

Systems, University of Vienna, Bruennerstr. 72, 1210, Vienna, 

Austria, ivana.ljubic@univie.ac.at 

1 - Spanning Trees with Node Degree Dependent Costs 

and Knapsack Reformulations 

Pedro Moura, DEIO, University of Lisbon, Campo Grande, 

Bloco C6, 1749-016, Lisbon, Portugal, pmoura@fc.ul.pt, Luis 

Gouveia 

The Degree constrained Minimum Spanning Tree Problem (DMSTP) consists 

in finding a minimal cost spanning tree satisfying the condition that every node 

has degree no greater than a fixed value. We consider a generalization of the 

DMSTP with a more general objective function including modular costs associated 

to the degree of each node, which have a technological motivation in 

the context of telecommunications networks. We present LP models together 

with some valid inequalities and compare their respective linear programming 

relaxations using cable and wireless network based instances. 

2 - Solving the Stochastic Steiner Tree Problem by Branchand-Cut 

Bernd Zey, TU Dortmund, Frankfurter Weg 3, 59439, 

Holzwickede, Germany, bernd.zey@tu-dortmund.de, Immanuel 

Bomze, Markus Chimani, Michael Juenger, Ivana Ljubic, Petra 

Mutzel 

TECHNICAL SESSIONS 

We consider the NP-hard Steiner tree problem under a two-stage stochastic 

model with recourse and finitely many scenarios. We discuss undirected, semidirected, 

and directed cut-set based integer programming models, and suggest 

a branch-and-cut approach based on Benders decomposition and a derived 

Integer-L-shaped algorithm. We compare our different models both theoretically, 

namely from a polyhedral point of view, and computationally. 

3 - Exact and Heuristic Algorithms for the Bounded Cycle 

Cover Problem 

Irene Loiseau, Departamento de Computación-, Facultad de 

Ciencias Exactas y Naturales, Universidad de Buenos Aires, 

Pabellón I - Ciudad Universitaria, 1428, Buenos Aires, 

Argentina, irene@dc.uba.ar 

We will present a new exact algorithm and heuristics for the Bounded Cycle 

Cover Problem. BCCP requires to determine a minimum cost cycle cover of 

a graph, with cycles bounded in length and number of edges. This problem 

arises in several situations related to telecommunications networks design, as, 

for example, when we want to design fiber-optic telecommunications networks 

that employ multiple self-healing rings to provide routing for communication 

traffic, even in the event of a fiber cut or other kinds of failures. 

4 - Reverse Multistar Inequalities and Vehicle Routing 

Problems with lower bound capacities 

Luis Gouveia, DEIO, University of Lisbon, Campo Grande, 

Bloco C6, 1749-016, Lisbon, Portugal, legouveia@fc.ul.pt, Juan 

José Salazar González 

In this talk we discuss, present and test models for the Capacitated Vehicle 

Routing Problem with arc lower bounds. We introduce the so-called reversed 

multistar inequalities and two other related families of inequalities, and show 

that they are relevant for modelling this routing problem. We present results 

from a branch-and-cut algorithm which uses the new inequalities for solving 

instances with up to 50 customers. 

� MA-03 


3.2.15 

TSP 

Stream: Metaheuristics 


Chair: José A. Moreno-Pérez, Estadística, I.O. y Computación, 

University of La Laguna, La Laguna, Spain, jamoreno@ull.es 

Chair: Noraini Mohd Razali, Mechanical & Manufacturing 

Engineering, Dublin City University, Dublin 9, Dublin, Ireland, 

norainimbr@ump.edu.my 

1 - An algorithm for solving the traveling salesman problem 

(tsp) based on multimodal transport, using a secuence 

of linear problems (lp) 

Luis Moreno, Systems, Universidad Nacional de Colombia, 

Facultad de Minas cra. 80 65-223, Bloque M8 oficina 207, 1, 

Medellin, Antioquia, Colombia, lfmoreno@unal.edu.co, Javier 

Diaz, Gloria Pena 

A hybrid algorithm is proposed based on heuristics and linear programming. A 

relaxed linear problem (LP) is solved. If the solution has several non connected 

circuits a new heuristic LP is used to connect the circuits, assuming the cities 

in each circuit in the relaxed solution are close between them. The circuits are 

connected as would do somebody after visiting by car several cities in a region 

that takes a plane to visit optimally another set of regions with additional cities. 

If when connecting the circuit of circuits there are again several non connected 

circuits the heuristic LP is solved in an iterative way that decreases the linear 

problem size until all the cities become shortly and easily connected. 

2 - A Modified Electromagnetism-like Algorithm for Travelling 

Salesman Problems with Precedence Constraints 

Alkin Yurtkuran, Industrial Engineering Department, Uludag 

University, Uludag University, Industrial Engineering 

1

MA-04 EURO 24 - Lisbon 2010 

Department, 16059, Bursa, Turkey, alkin@uludag.edu.tr, Erdal 

Emel 

Traveling Salesman Problem with Precedence Constraints(TSPPC) is an important 

variant of Traveling Salesman Problem. TSPPC belongs to the class of 

NP-Hard problems where there exits a precedence relationship between customers. 

In this study, a Modified Electromagnetism-like Algorithm(EMA) is 

applied to solve TSPPC problems. The key concept of the proposed algorithm 

is the projection of the particle space onto a new coordinate space where each 

precedence constraint is ensured. The computational results show that Modified 

EMA gives promising results within acceptable computational times. 

3 - Performance comparison between different GA selection 

strategies in solving TSP instance. 

Noraini Mohd Razali, Mechanical & Manufacturing 

Engineering, Dublin City University, Dublin 9, Dublin, Ireland, 

norainimbr@ump.edu.my, John Geraghty 

This study presents the comparison of genetic algorithm performance on solving 

TSP using two different stochastic selection methods which are tournament 

and roulette wheel. The GA is mainly composed of three genetic operations 

which are selection, crossover and mutation. With the same crossover and 

mutation operation, the study is focussed on comparing the effect of different 

selection strategy to the performance of convergence that gives optimum 

solution. Numerical experiments show that GA with tournament selection converges 

much faster than roulette wheel selection. 

4 - A Data Mining Based Heuristic Approach for Solution 

of Travelling Salesman Problem 

Hande Gulkac, Computer Engineering, Okan University, Turkey, 

hande.gulkac@okan.edu.tr, Semiye Gönülol, Ahmet Cihan, 

Halenur ¸Sahin, Alpaslan Figlali 

Travelling Salesman Problem (TSP) is a well known NP-hard optimization 

problem. Many methods are applied including heuristics, mathematical programming 

and metaheuristics for obtaining good solutions. In this study a data 

mining based heuristic approach is applied for the solution of TSP. Procedure 

uses a repetitive and forced random tour generation approach. The set of best 

random tours are analyzed via data mining tools to obtain the relations. Using 

the rules derived from the relations, the TSP tour is obtained. The results are 

promising while considering it’s simplicity. 

� MA-04 


3.2.13 

Scheduling with metaheuristics 



Chair: Lionel Amodeo, Charles Delaunay Institute, University of 

Technology of Troyes, 12 Rue Marie Curie BP2060, 10000, Troyes, 

France, lionel.amodeo@utt.fr 

Chair: Farouk Yalaoui, Institut Charles Delaunay, ICD LOSI, 

University of Technology of Troyes, 12, rue marie curie BP 2060, 

10000, Troyes, France, farouk.yalaoui@utt.fr 

1 - Single Machine Scheduling with Rejection: Minimizing 

total Weighted Completion Time and Rejection Cost 

Atefeh Moghaddam, Charles Delaunay Institute, University of 

Technology of Troyes, 12 rue Marie Curie„ Troyes, France, 

10000, Troyes, France, atefeh.moghaddam@utt.fr, Farouk 

Yalaoui, Lionel Amodeo 

It is always assumed that we have to process all jobs. However, we can break 

the assumption by rejecting certain jobs. In this paper, we consider that the 

jobs can be either scheduled on a single machine or be rejected at the cost of 

a penalty. Two objective functions are considered: minimizing total weighted 

completion times and minimizing total penalties. We apply two-phase method 

to find all Pareto-optimal solutions. We also propose bi-objective simulated 

annealing algorithm to find estimated Pareto-optimal solutions. By comparing 

the solutions, we show that the results are reasonably good. 

2 - A Tabu Search Algorithm for Order Acceptance and 

Scheduling 

2 

Bahriye Cesaret, Industrial Engineering, Koc University, Koc 

University, Rumelifeneri yolu Sariyer, 34450, Istanbul, Turkey, 

b.cesaret@gmail.com, Ceyda Oguz, Sibel Salman 

We develop a Tabu Search (TS) algorithm with a probabilistic local search after 

each iteration to solve the order acceptance and scheduling problem on a single 

machine with sequence dependent setup times. We compare the performance 

of the TS algorithm to a greedy constructive heuristic from the literature, using 

upper bounds based on a mixed integer programming formulation. Computational 

studies show that the TS algorithm gives significantly better solutions 

than those of the constructive heuristic in terms of objective function value in 

all instances tested with a small increase in run time. 

3 - Approximate methods for solving the operating room 

planning problem 

Jose M. Molina-Pariente, University of Seville, Spain, 

jmolinap@esi.us.es, Jose M Framinan, Paz Perez Gonzalez 

In this communication, we address the operating room planning problem for a 

surgery unit. This problem involves determining a surgery schedule that specifies 

the number of surgical cases to be scheduled in a given planning horizon 

together with the date of the intervention and the specific operating room in 

which each surgical case will be performed. To solve this problem, we present 

new constructive heuristics and a local search method. We evaluate the performance 

of these methods by generating a large set of instances based on an 

analysis of the literature. 

4 - A GRASP approach to the Multi-Task Employee 

Timetabling Problem 

Pilar Tormos, STATISTICS AND OPERATIONS RESEARCH, 

UNIVERSIDAD POLITECNICA DE VALENCIA, CAMINO 

DE VERA S/N, VALENCIA, 46022, VALENCIA, SPAIN, 

Spain, ptormos@eio.upv.es, Antonio Lova 

Employee Timetabling Problem (ETP) is the operation of assigning employees 

to tasks in a set of shifts during a period of time while satisfying the existing 

constraints and preferences. An extension of this problem, the Multi-Task Employee 

Timetabling Problem (MTETP) implies the assignment of the sequence 

of tasks to be performed by each employee every working day of the planning 

horizon and it is especially relevant for commercial companies. A greedy 

randomized adaptive search procedure (GRASP) is developed to solve it and 

embedded in a computer-aided system (OPTIHPER). A customized version of 

it is in use with very satisfactory results by a leading Spanish distribution company. 

� MA-05 


3.2.16 

Theory 



Chair: Gustavo Melo, Computer Science, Universidade Estadual do 

Ceará, Rua césar fonseca 410 ap 301 bairro papicu, 60176-110, 

Fortaleza, Ceará, Brazil, gustavo.sikora@gmail.com 

Chair: Zahira Benkhellat, informatique, sciences exactes, Bejaia 

university Qlgerie, teacher, Algeria, benkhellat_zahira@yahoo.fr 

1 - Analysis of software for the NGS technology: The Survival 

Guide 

Bruno Vieira, Computational Biology & Population Genomics 

Group, Centro de Biologia Ambiental, Departamento de 

Biologia Animal, Faculdade de Ciências da Universidade de 

Lisboa, Campo Grande, 1749-016, Lisbon, Portugal, 

mail@bmpvieira.com, Francisco Pina Martins, Sofia Seabra, 

Octavio Paulo 

Next Generation Sequencing (NGS) technologies allow the generation of large 

amounts of data in a short time span and for a relatively low cost. A multitude 

of software was recently developed to address the difficulties generated 

by NGS, such as assembling the millions of reads, contig generation and the 

follow up annotation. The cover rate is also relevant to the results accuracy and 

the detection of genetic variation, either in the form of SNPs or CNVs. In this 

communication we address several critical bioinformatics steps and compare 

current software and algorithms to tackle these problems.

� MA-06 


8.2.30 

DEA Methodology I 

Stream: DEA and Performance Measurement 


Chair: Dimitris Despotis, Department of Informatics, University of 

Piraeus, 80, Karaoli & Dimitriou Street, 18534, Piraeus, Greece, 

despotis@unipi.gr 

1 - DEA approach for evaluating performance considering 

institution goal 

Sheu-hua Chen, Distribution Manegement, National Chin-Yi 

University of Technology, Taipin, Taichung, Taiwan, ROC, 411, 

Taichung, Taiwan, Taiwan, shchen@ncut.edu.tw, Hong Tau Lee 

In the data envelopment analysis (DEA) approach some decision-making units 

may reach performance efficiency by their outstanding performance on some 

relative unimportant output items. In this research, we try to add the strictly 

predefined relationships of output items in the existed model that expressed the 

relative importance of those input or output items. According to this approach, 

only the DMUs that really match the predefined requirements and have good 

performance can be regarded as efficient units. That is to say the meaning of 

performance depends on the goals the organization pursues. This is meaningful 

for managerial practices. A case of performance evaluation of faculty in different 

types of university with specific development orientation is provided to 

illustrate the proposed idea. 

2 - Assessing robustness in additive DEA with interval 

measurements 

Maria Gouveia, ISCAC, Quinta Agricola - Bencanta, 3040-316 , 

Coimbra, Portugal, mgouveia@iscac.pt, Luis C. Dias, Carlos 

Henggeler Antunes 

This study addresses the problem of finding the range of efficiency for each 

Decision Making Unit (DMU) considering uncertain data. Uncertainty in the 

DMU coefficients in each factor (input or output) is captured through interval 

coefficients (i.e. these are uncertain but bounded). A two-phase additive Data 

Envelopment Analysis (DEA) model for performance evaluation is used, which 

is adapted to include the concept of super-efficiency to provide a robustness 

analysis of the DMUs in face of uncertain information. 

3 - A proposition of the minimum distance model in Network 

DEA 

Tohru Ueda, Faculty of Science and Technology, Seikei 

University, 3-3-1 Kichijoji-Kitamachi, 180-8633, 

Musashino-Shi, Tokyo, Japan, ueda@st.seikei.ac.jp, Hirofumi 

Amatatsu 

Minimization of objective function in SBM model results in maximization of 

slacks sum. This maximization corresponds to finding a point in the production 

possibility set that is the farthest point from each Decision Making Unit to be 

evaluated. To overcome this shortage, we proposed the unified DEA model. 

Traditional linking constraints where continuity between input and output is 

kept may be too severe to evaluate efficiencies. Considering linking constraints 

and the unified DEA model with minimum distance, we propose a new network 

DEA model and discuss efficiencies of prefectures in Japan. 

4 - An improving approach for estimating return to scale in 

DEA 

Maryam Allahyar, mathematics, science and research branch 

islamic azad university, tehran-ashrafi esfahani highway-to 

hesarak, 0098, tehran, tehran, Iran, Islamic Republic Of, 

mayar4584@yahoo.com, Mohsen Rostamy-malkhalifeh 

In this article a new method will be suggested for the determination of the 

right and left return to scale (RTS). The new approach is different form that of 

Golany and Yu (1997) and doesn’t have its shortcomings. Our approach is able 

to evaluate the right and left RTS in all conditions for any unit. 

EURO 24 - Lisbon 2010 MA-07 

� MA-07 


8.2.47 

Recent Developments in Mathematical 

Programming 

Stream: Mathematical Programming [c] 

Contributed session 




Chair: Jerzy Filar, Mathematics and Statistics, University of South 

Australia, Mawson Lakes Blvd, 5095, Mawson Lakes, SA, Australia, 

j.filar@unisa.edu.au 

1 - A Relax-and-Fix Lagrangean Relaxation Based Algorithm 

for a Class of Multiple-Choice Integer Problems 

Abdelkader Sbihi, Information Systems and Decision Making 

Science, Audencia-Nantes School of Management, 8 route de la 

Jonelière, BP 31222, 44312 , Nantes Cedex 3, France, 

asbihi@audencia.com 

We propose a Lagrangean relaxation based-algorithm for some hard combinatorial 

problems. The idea is to relax a certain type of constraints then to fix 

variables to their optimal value. We consider: (i) the Multiple-Choice Knapsack 

Problem (MCKP) and (ii) the Multiple-Choice Subset Sum Problem (MC- 

SSP) which can be considered as a special case of the MCKP. We used MCSSP 

as an auxiliary problem to tighten the capacity constraint. The obtained results 

showed a high quality of the computed upper bounds. The benchmark has 

demonstrated a high efficiency of the approach. 

2 - An Optimization Method for Solving Assemly Line Balancing 

Problem 

Sukran Seker, Industrial Engineering Department, Yildiz 

Technical University, Barboros Street Yildiz Technical 

University Industrial Engineering Department 343409 Besiktas, 

Istanbul, Turkey, sukranseker@yahoo.com, Mesut Özgürler 

Assembly line balancing or simply line balancing is the problem of assigning 

operations to workstations along an assembly line in such a way that the 

assignment be optimal in some sense. Assign tasks to work stations observing 

balancing restriction so as to minimize balance delay while keeping station 

work content for every station cycle time. There have been a large number 

of proposals for theoretical and practical methods for solving the line balance 

problem. This paper use one of the optimization solution approach to solve 

assembly line balancing problem. 

3 - Geo-spatial data mining by model-based clustering 

methods 

Francisco Figueiredo, UNIDE, ISCTE-IUL, Lisbon, Portugal, 

francisco.m.fig@gmail.com, José G. Dias 

Most of the clustering techniques are inadequate for geo-spatial data mining as 

they tend to ignore that spatially closer areas tend to be more similar than the 

others. Geo-spatial clustering aims to find groups of similar objects that are 

spatially close. We propose a clustering algorithm for spatial segmentation of 

count data under a regression framework, which combines the Neighborhood 

EM (NEM) and Hybrid EM (HEM) algorithms. The geo-spatial data mining 

approach is illustrated with georeferenced political data. 

4 - Alternating Proximal Algorithms and Hierarchical Selection 

of Optima in Games, Control and PDE’s 

Juan Peypouquet, Mathematics, Universidad Tecnica Federico 

Santa Maria, Av Espana 1680, 2340000, Valparaiso, Valparaiso, 

Chile, juan.peypouquet@usm.cl, Hedy Attouch, Marco 

Czarnecki 

We study alternating and diagonal proximal algorithms combining resolvent 

iterations and a penalization scheme. The resulting sequence of iterates and, 

under less restrictive conditions, their averages converge weakly to a point with 

special properties. We also analize a splitting method for structured variational 

problems and comment on the robustness of these methods. The results enable 

us to solve constrained or bilevel optimization problems. This method is applied 

to best response dynamics with cost to change, optimal control problems 

and domain decomposition for PDE’s. 

3


� MA-08 


6.1.36 

Project Management Software and 

Applications 

Stream: Project Management and Scheduling 


Chair: Norbert Trautmann, Department of Business Administration, 

University of Bern, IFM, AP Quantitative Methoden, 

Schützenmattstrasse 14, 3012, Bern, BE, Switzerland, 

norbert.trautmann@pqm.unibe.ch 

Chair: Christoph Schwindt, Institute of Management and Economics, 

Clausthal University of Technology, Julius-Albert-Str. 2, 38678, 

Clausthal-Zellerfeld, Germany, christoph.schwindt@tu-clausthal.de 

1 - Heuristic improvement of Microsoft Project’s resourceallocation 

capabilities 

Norbert Trautmann, Department of Business Administration, 

University of Bern, IFM, AP Quantitative Methoden, 

Schützenmattstrasse 14, 3012, Bern, BE, Switzerland, 

norbert.trautmann@pqm.unibe.ch, Philipp Baumann 

Microsoft Project is widely used for temporal scheduling and resource allocation 

of projects. We show that its integrated resource-allocation procedure 

uses neither the serial nor the parallel schedule-generation scheme, and that 

the procedure performs relatively poor. We present a bi-directional scheduleimprovement 

heuristic. Computational results for a construction project and 

for the J30, J60, and J120 instances from PSPLIB indicate that this heuristic 

shortens the project duration considerably. 

2 - Exact methods for resource levelling problems 

Jürgen Zimmermann, Operations Research, TU Clausthal, 

Julius-Albert Str. 2, 38678, Clausthal-Zellerfeld, Germany, 

juergen.zimmermann@tu-clausthal.de, Thorsten Gather 

We present exact solution methods for resource levelling problems with minimum 

and maximum time lags among the project activities. In particular, we 

consider a time window based enumeration method and two tree-based branchand-bound 

procedures both with a sophisticated constructive lower bound. Furthermore, 

we propose a mixed integer linear programming formulation that can 

be solved by standard solvers such as CPLEX. All approaches are compared in 

a comprehensive computational study using well known test sets from literature. 

Instances with up to 30 activities could be solved to optimality. 

3 - Temporal scheduling of concurrent engineering 

projects 

Christoph Schwindt, Institute of Management and Economics, 

Clausthal University of Technology, Julius-Albert-Str. 2, 38678, 

Clausthal-Zellerfeld, Germany, 

christoph.schwindt@tu-clausthal.de, Philipp Benke 

The concurrent engineering approach is intended to shorten the cycle time 

of development projects by parallelizing consecutive development phases. 

We consider the tradeoff between the time savings enabled by overlapping 

precedence-related project activities and the increase in the activity durations 

that is typically incurred by additional integration and coordination efforts. We 

investigate structural properties of the temporal scheduling problem and explain 

how earliest and latest start and completion times of the activities can be 

determined efficiently based on label-correcting algorithms. 

4 - Integrated Scheduling and Staffing IT-Projects 

Rainer Kolisch, TUM School of Management, Technische 

Universitaet Muenchen, Arcisstr. 21, 80333, Muenchen, 

Germany, rainer.kolisch@wi.tum.de, Christian Heimerl 

In this paper we present an optimization model to address the problem of 

scheduling the activities of multiple IT-projects with serial structures and assigning 

the project work to multi-skilled internal and external human resources 

with static and heterogeneous efficiencies. The mixed-binary linear program 

is solved using ILOG CPLEX and a hybrid metaheuristic. The latter employs 

an efficient evaluation function exploiting the network structure of the staffing 

subproblem. We assess the impacts of several problem parameters on computation 

time and solution gaps. 

4 

� MA-09 


6.2.53 

Challenges of Mathematical Programming 

by Modern Applications 

Stream: Mathematical Programming 


Chair: Zuzana Oplatkova, Dept. of Applied Informatics, Tomas Bata 

University in Zlin, Nad Stranemi 4511, 76005, Zlin, Czech Republic, 

oplatkova@fai.utb.cz 




1 - An optimization approach for prediction of microbial 

growth strategies 

Pinar Ozturk, Systems Bioinformatics, Vrije University, 

Amsterdam, De Boelelaan 1085, 1081HV, Amsterdam, 

Netherlands, pinar.ozturk@falw.vu.nl, Douwe Molenaar 

In limited nutrient conditions, microorganisms regulate cellular activities to 

maintain efficient growth. Efficiency is regarded as correlated with fast reproduction. 

Via an optimization model of the whole cell with an objective to 

maximize growth rate, we predict macro features of microbes at given nutrient 

concentrations by representing essential pathways with modules. This, as 

known to us, is the first time growth rate, size and shape are predicted having 

only intrinsic physical properties of molecules as constraints. 

2 - Robust model development for non-linear models 

Magderie van der Westhuizen, School of Computer, Statistical 

and Mathematical Sciences, North-West University, Private Bag 

X6001, 2520, Potchefstroom, South Africa, 

magderie.vanderwesthuizen@nwu.ac.za, Giel Hattingh, Hennie 

Kruger 

The predictive capability of regression models relies heavily on the applicability 

of the assumptions made by the model builder. In addition, the presence 

of outliers may also lead to models that are not reliable. This study reports 

on robust techniques applied to minimal assumption regression models in an 

effort to improve predictive capability. The approach is based on mathematical 

programming techniques combined with smoothing and piecewise linear 

techniques. Different cases from the literature are considered and presented as 

illustrative examples. 

3 - A Bilevel Competitive Facility Location Model with 

Competitor’s Response 

Hande Kucukaydin, Industrial Engineering, Bogazici University, 

Bogazici University Industrial Engineering Department, 

Bebek-Istanbul-Turkey, 34342, Istanbul, Turkey, 

hande.kucukaydin@boun.edu.tr, Necati Aras, I. Kuban Altinel 

We are concerned with a problem in which a new entrant leader firm aims 

at finding the location and attractiveness of each new facility to maximize its 

profit where there are existing facilities belonging to a competitor. The competitor 

reacts to the leader by adjusting the attractiveness levels of its existing 

facilities to maximize its profit. We first formulate a bilevel mixed-integer nonlinear 

programming model. Then, we convert it into an equivalent single level 

mixed-integer nonlinear program and solve it using global optimization methods. 

� MA-10 


6.2.56 

Graphs and Networks I 

Stream: Graphs and Networks 


Chair: Reinhardt Euler, Informatique, Université de Brest, 20 av. Le 

Gorgeu, B P 817, 29285, Brest, France, reinhardt.euler@univ-brest.fr

1 - On the Power of Decomposition for the Maximum Independent 

Set Problem 

Andreas Brandstädt, Universität Rostock, 18055, Rostock, 

Germany, ab@informatik.uni-rostock.de 

(joint work with C.T.Hoang, V.B.Le, V.V.Lozin, and R. Mosca) In a finite undirected 

graph G=(V,E) a vertex set S is ’independent’ (or ’stable’) if the vertices 

in S are mutually nonadjacent. For given G, the MAXIMUM INDEPENDENT 

SET (MIS) Problem asks fo an independent vertex set of maximum size in G. 

The MWS problem asks for an independent set with maximum vertex weight; 

the MIS problem is the MWS problem with unit weights. It is well known 

that MWS (MIS,respectively) is intractable and hard to approximate. We discuss 

various decomposition techniques such as clique separator decomposition, 

modular decomposition and split decomposition for solving the MWS problem 

efficiently on various particular graph classes. It is well known that the 

above mentioned decompositions are helpful tools for solving the MWS problem. 

One of our results allows to combine clique separator decomposition and 

modular decomposition. This implies various improvements of known results, 

among them a polynomial time algorithm for MWS on apple-free graphs which 

are a common generalization of chordal graphs, cographs and claw-free graphs. 

Finally we mention some open problems. 

2 - Solving efficiently the weighted stable-set problem in 

claw-free graphs using a reduction operation 

Paolo Nobili, Mathematics, University of Lecce, Via Arnesano, 

73100, Lecce, Italy, paolo.nobili@unile.it, Antonio Sassano 

Maximum weight stable sets can be computed in polynomial time for claw-free 

graphs (Minty, Nakamura et al., Schrijver, Oriolo et al.). In this paper we define 

the strongly reducible cliques, extending to the weighted case a reduction operation 

of Lovasz and Plummer. We use the operation for obtaining maximum 

weight alternating paths through matching computations. We embed the procedure 

in an iterative approach that contructs a given claw-free graph G node 

by node, maintaining the associated maximum weight stable set. The resulting 

algorithm has computational complexity O(n4log(n)). 

3 - Reconstruction of Permutations Respect to some Generator 

Sets of the Symmetric Group 

Alpar Vajk Kramer, DEIO, FCUL, Portugal, vajki@web.de 

We will consider the reconstruction of permutations regarding some special 

generator sets of the symmetric group. The generator sets considered are particular 

subsets of involutions such as the reversals, prefix reversals, bubble reversals 

or Coxeter generators and transpositions. The common property of all 

this generator sets is that their corresponding Cayley graph does not contain 

triangles. 

� MA-11 


8.2.38 

Emerging Data Mining Applications in 

Biomedics and Biotech 

Stream: Emerging Applications of OR 


Chair: Honora Smith, School of Mathematics, University of 

Southampton, Highfield, SO17 1BJ, Southampton, Hampshire, 

United Kingdom, honora.smith@soton.ac.uk 




Chair: Veronica Biga, Department of Automatic Control and Systems 

Engineering, The University of Sheffield, Mappin Street, S1 3JD, 

Sheffield, Afghanistan, v.biga@sheffield.ac.uk 

1 - The Information Effect of the Infectious Diseases Outbreak 

on Biotechnology Stock Performance 

Yi-Hsien Wang, Department of Banking & Finance, Chinese 

Culture University, 55, Hwa-Kang Road, Yang-Ming-Shan., 

Taipei, Taiwan 11114, R.O.C, 11114, Taipei, Taiwan, 

holland@mail2000.com.tw, Fu-Ju Yang, Kuang-Husn Shih, 

Li-Je Chen 


Statutory infectious diseases breakout not only affects people’s health and lives, 

but also stagnates the economic growth. The prevalence of infectious diseases 

also provides the development opportunities of biotechnology corporations. 

The effect of the statutory infectious diseases outbreak on Taiwanese biotechnology 

stock price movements is examined. The empirical results point out that 

there exists a significantly positive abnormal return of Taiwan’s biotechnology 

industry because of the statutory infectious epidemic. 

2 - A hybrid classification method: Using a support vector 

machine for rule extraction on diabetes diagnosis 

Chien-hsin Yang, Department of Industrial Engineering and 

Management, Overseas Chinese University, 100,Chiao Kwang 

Rd., Taichung 40721, Taiwan, Taiwan, 

chyang.iem91g@nctu.edu.tw, Chun-Chin Hsu 

Many of the factors related to diabetes mellitus (DM) have been discovered 

from a succession of studies in epidemiology. However, it seems that the usability 

and explainability of methods are inferior to rules extraction. A support 

vector machine based hybrid classification method is employed to extract rules 

for DM diagnosis. To evaluate performance, C5 and back-propagation neural 

networks were used as benchmarks. Results from the hybrid approach demonstrate 

high accuracy and fidelity, and the rules can help for preventive medicine 

in DM diagnosis. 

3 - Application of artificial neural network and support vector 

machine to classify the risk of death of hospitalized 

patients with acute coronary syndrome 

Rodrigo Collazo, Operational Research, CASNAV/UFRJ, Costa 

Doria St, 17, 21910-170, Rio de Janeiro, Rio de Janeiro, Brazil, 

rodrigocollazo@uol.com.br, Basílio Pereira, Laura Bahiense, 

Amália Faria dos Reis, Amália Faria dos Reis 

This study developed an artificial neural network model and a support vector 

machine model to classify the risk of death of hospitalized patients with acute 

coronary syndrome at high and low. It was used the mutual information feature 

selector under uniform information distribution criteria (MIFS-U) for selection 

of the most important input variables. The computational results show a better 

performance of the support vector machine model compared with the artificial 

neural network model and indicate the input variables age, creatinine and any 

prior revascularization as the most relevant. 

4 - Prediction in Medicine: Statistical Models versus Artificial 

Neural Networks 

Ana Papoila, Bioestatística e Informática, Faculdade Ciências 

Médicas da Universidade Nova de Lisboa, CEAUL, Portugal, 

apapoila@hotmail.com, Carlos Geraldes, Patricia Xufre 

Artificial Neural Networks are often used in Biomedical Sciences and 

Medicine. A main goal is to predict a clinical outcome after taking into account 

a set of independent explanatory variables. ANNs arise as an alternative 

to logistic regression. This study compare Generalized Linear Models with 

binary response, with the performance of ANNs, in what concerns their predictive 

and discriminative power. For both approaches, validation techniques 

were applied. These methodologies were used to predict mortality of patients 

admitted to an Intensive Care Unit located in Lisbon. 

5 - A Study in Different Channels’ Consumer on the Purchasing 

Intention and Behavior of Bio-technology Products 

Yuanchau Liour, Logistics Management, Takming University of 

Science and Technology, 11451, Taipei, Taiwan, 

ycliour@takming.edu.tw, Chiao-Ling Huang, Chie-bein Chen 

Biotechnology has been playing an important role in modern financial society; 

recently, as improvement of economy environment and the development 

of technology, the government is committed to the implementation of biotechnology. 

We explore the impact of consumer attitude, consumer’s purchasing 

intention, promotions and product involvement, and perceive risk on purchasing 

intention. We take Northern Taiwan area’s consumer as the research objects, 

use SEM to analyze and make suggestions to bio-technology health food 

companies in accordance with the empirical conclusions. 

5


� MA-12 


8.2.39 

AHP 01 

Stream: Analytic Hierarchy Processes, Analytic Network 

Processes 


Chair: Grzegorz Ginda, Dept. of Operational Research in 

Management, Opole University of Technology, Faculty of 

Management and Production Engineering, ul. Warynskiego 4, 

45-057, Opole, Poland, g.ginda@po.opole.pl 

1 - Aviation and the Belgian Climate Policy (ABC) - A Multicriteria 

Analysis (MCA) for the evaluation of policy options 

to mitigate the total aviation climate change impact 

Annalia Bernardini, MOSI-T, Vrije Universiteit Brussel, 

Pleinlaan 2, 1050, Brussels, Vlaamse Brabant, Belgium, 

Annalia.bernardini@vub.ac.be, Tom Van Lier, Annelies 

Heemeryck, Ellen Van Hoeck, Cathy Macharis 

The ABC project analyses the different climate policy options aimed to reduce 

the climate change impacts of the aviation sector. In view to compare the different 

alternative policies (finan.&econ. tools, R&D, operat. procedures) the 

MCA method is applied. The performances of those policies are evaluated in 

relation to some appropriate criteria (env. performances, social-economic impacts 

aviation sector). A combination of the Analytic Hierarchy Process and the 

PROMETHEE method allows to come to a detailed analysis of the advantages 

and disadvantages of each of the proposed policy measures. 

2 - A consistency-based method for aggregating preference 

information from multiple pairwise comparison 

matrices 

Esther Dopazo, Lenguajes y Sistemas Informáticos, Universidad 

Politecnica Madrid, Facultad Informática, Campus de 

Montegancedo, 28660, Boadilla del Monte, Madrid, Spain, 

edopazo@fi.upm.es, Mauricio Ruiz-Tagle 

We consider a group decision problem, where decision makers estimate their 

preferences of a set of alternatives into the form of pairwise comparison matrices 

(a well-established technique in this field). In this scenario, a fundamental 

problem is the generation of a priority vector for the alternatives from the 

pairwise matrices which represents the consensus opinion for the group. We 

propose a weighted logarithmic goal programming method for aggregating individual 

opinions into an optimal group priority vector, where the consistency 

of each expert is taken into consideration. 

3 - An agreement-based approach for generating priority 

vectors from multiple pairwise comparison matrices 

Mauricio Ruiz-Tagle, Facultad de Cs. de la Ingeniería, 

Universidad Austral de Chile, General Lagos 2086, Campus 

Miraflores, Valdivia, Chile, mruiztag@uach.cl, Esther Dopazo 

The problem of importance weights analysis and determination from multiple 

source information is a critical issue in many fields such as machine learning, 

meta-search engines, multi-criteria decision making, etc. We focus on the 

problem of computing the importance weights and the corresponding rank ordering 

of a set of alternatives from information given by a group of experts 

into the form of pairwise comparison matrices. We present an approach based 

on lp distance-based aggregation functions and on the use of consensus-driven 

weights for quantifying the relative importance of the experts. 

4 - Integrated MADA Assessment Tool 

Grzegorz Ginda, Dept. of Operational Research in Management, 

Opole University of Technology, Faculty of Management and 

Production Engineering, ul. Warynskiego 4, 45-057, Opole, 

Poland, g.ginda@po.opole.pl, Miroslaw Dytczak 

Integrated tool for interdisciplinary assessment of decision alternatives in management 

and engineering is discussed in the paper. The tool makes use of several 

selected MADA approaches to obtain more diversified results. A common 

data structure is applied to make preparation of required data less expensive in 

terms of both time and work effort. Component methods make it possible to 

include and assess influence of both intangible and tangible aspects. They are 

also easily implementable. The tool addresses issues of input data consistency 

and group decision support too. The tool is unique with regard to ability of 

adaptation to particular needs. Sample analysis is included which shows its 

applicability and scale of potential application benefits. 

6 

� MA-13 


2.2.21 

Location and GIS 

Stream: Location Analysis 


Chair: Ioannis Giannikos, Business Administration, University of 

Patras, University of Patras, GR-26500, Patras, Greece, 

I.Giannikos@upatras.gr 

1 - Enhancing Location Optimization Modeling Capabilities 

through the use of GIS 

Alan Murray, Geographical Sciences and Urban Planning, 

Arizona State University, P.O. Box 875302, 85287-5302, Tempe, 

AZ, United States, atmurray@asu.edu 

The prevalence of widely available and accessible geographic information system 

(GIS) packages and associated geographic data has been important to all 

disciplines that study, analyze and evaluate spatial problems. In this paper we 

summarize major characteristics of GIS relevant to location modeling and spatial 

optimization. A number of widely relied upon optimization models are 

detailed. Particular attention is given to identifying implementation and application 

limitations, and how these can be overcome through integration with 

GIS. 

2 - The Geographical Information System ’Ptolemeos- 

Europe’ and the analysis of regional geo-economic dynamics 

of France 

John Karkazis, Business School, University of the Aegean, 

Chios, GR-82100, Chios, Greece, ikarkazis@aegean.gr 

This paper explores the regional geo-economic dynamics of France during the 

period 1985-2004 employing GIS "Ptolemeos-Europe’. In the beginning, the 

key notion of regional efficiency is introduced as well as other key regional 

analysis notions such as: regional discrimination cost and regional discrimination 

iso-curves. Based on the notion of regional efficiency the general geoeconomic 

gravity model is introduced and its outcome, the geo-economic gravity 

centers and other related strategic and dynamic notions such as the capital 

displacement factor and the velocity of geo-economic gravity centers are presented 

and discussed. The above are applied to France in order to explore the 

strategic geo-economic trends of the 22 administrative regions of it. 

3 - Improving the efficiency of WEEE collection systems 

using a web-based GIS application 

Simão Ribeiro, Production and Systems Department, University 

of Minho, Portugal, Universidade do Minho, Campus de Gualtar, 

4710-057, Braga, Portugal, secribeiro@gmail.com, Jorge 

Pereira, Joel Carvalho, José Oliveira, Manuel Figueiredo, José 

Telhada, Luis Dias 

This project focuses on the design of a web-based GIS application to support 

the planning and management of collecting waste of electrical and electronic 

equipments (WEEE) networks. It addresses the issues of gathering and managing 

information needed by network optimization modules being included in an 

integrated computerized application, and the issues of analyzing and mapping 

their outputs. Several GIS and database technologies will be used, and their 

applicability and utility will be discussed. In overall, it is expected that relevant 

economic and environmental benefits will be achieved. 

4 - Multiobjective Demand Covering Models based on GIS 

Ioannis Giannikos, Business Administration, University of 

Patras, University of Patras, GR-26500, Patras, Greece, 

I.Giannikos@upatras.gr, Georgios Alexandris 

In this paper we discuss a number of maximal demand covering models where 

the customers as well as the servers may be geographic objects rather than 

single points. Through the use of Geographic Information Systems (GIS), we 

consider different notions of coverage and develop a series of multiobjective 

programming models that take into account the following objectives: (a) maximization 

of total coverage, (b) maximization of minimum coverage and (c) 

minimization of distance to servers of uncovered demand objects. These models 

take into account the geography of each demand area in question and adjust 

the location of the servers accordingly.

� MA-14 


2.2.15 

Inventories in Supply Chains 

Stream: Supply Chain Planning 


Chair: Horst Tempelmeier, Supply Chain Management and 

Production, University of Cologne, Albertus-Magnus-Platz, D-50923, 

Cologne, Germany, tempelmeier@wiso.uni-koeln.de 

1 - A discrete time multi-level inventory system with a 

make-to-order supplier 

Horst Tempelmeier, Supply Chain Management and Production, 

University of Cologne, Albertus-Magnus-Platz, D-50923, 

Cologne, Germany, tempelmeier@wiso.uni-koeln.de 

We study a supply network comprising a factory following a make-to-order 

strategy, a warehouse using a reorder point-reorder quantity policy and distribution 

centers using base-stock policies in discrete time. The factory is modeled 

as a discrete time G/G/1 queueing system. The system is decomposed into three 

layers that are linked through random waiting times. An overall optimization 

model including as decision variables the processing time in the factory and 

the parameters of the inventory policies applied in the distribution system is 

formulated and solved. 

2 - Two-Capacitated-Supplier Two-Stage Periodic-Review 

Supply Chain Problem Investigation 

Kai Luo, Operations Management & Information Technology, 

HEC Paris, 1, rue de la Libération, 78351, Jouy en Josas, Paris, 

France, kai.luo@mailhec.net, Laoucine Kerbache, Ramesh 

Bollapragada 

In this paper, we investigate a two-product (high-end, low-end), one retailer / 

two suppliers problem with random demand and periodic review. The problem 

is decomposed into sub-models and solved sequentially. For simple cases, 

closed form expressions are provided for the optimal solution. We show that, 

under certain conditions, the retailer should place the high-end product in the 

secondary inventory. For complex cases, we propose a heuristic to solve the 

problem and provide managerial insights. 

3 - Heuristics for multi-item, two-echelon inventory system 

with aggregate mean wait time constraint. 

Arjun Subramaniam, Applied Materials, Mountain View, 

California, United States, arjun.subramaniam@gmail.com, 

Deepak Bhatia 

We consider a multi-item, two-echelon spare parts inventory system, with one 

central warehouse and multiple local warehouses. We present close-to-optimal, 

scalable heuristics to minimize total cost with each local warehouse subject to 

an aggregate mean wait time constraint. All locations operate under a continuous 

review system with base stock policies. We test effectiveness by comparing 

with the lower bound and demonstrate better performance compared to results 

from recently published works. 

4 - Price and Perception - Understanding the Consumer 

Side of Recovered Products 

Jonathan Linton, School of Management, University of Ottawa, 

39 Sachs Forest Place, K2G 6V2, Ottawa, Ontario, Canada, 

linton@telfer.uottawa.ca, Leila Hamzaoui 

We seek to address the gap in understanding consumer willingness to pay for 

products that are comprised of recovered materials and parts. Consequently, a 

survey of 320 respondents was conducted to determine the willingness to pay 

for different types of products containing recovered materials and components. 

A series of related hypotheses are provided and tested. In addition to considering 

issues personal attitudes to the environment and perceived risk, we consider 

the effects of branding and product characteristics. While the work is empirical 

in nature it is critical to supply chain planning as there is limited research and 

understanding of the consumer side of close-looped supply chains. 


� MA-15 


2.2.12 

Location-routing problems 

Stream: Vehicle Routing 


Chair: Christian Prins, Laboratoire LOSI, Université de Technologie 

de Troyes, BP 2060, 10010, Troyes Cedex, France, prins@utt.fr 

1 - Centralized Distribution System of Infusion Solutions 

on a Network of Health Care Units 

William Guerrero, Industrial Engineering, Universidad de los 

Andes, Av Cll 147 No 17-81 ap 502, 472, bogotá, bogotá D.C., 

Colombia, industrialito_@hotmail.com, Nubia Velasco, Ciro 

Alberto Amaya, Christelle Gueret, Thomas Yeung 

A methodology to improve inventory control and distribution policies in hospitals 

is presented. The strategy is to centralize the management of medicines 

into a single depot to reduce costs. It is aimed to find optimal inventory control 

policies for one-warehouse n-retailer distribution system based on a Markov 

Chain model. Results are evaluated on a real hospital. An MIP model to decide 

the location of the central depot and distribution routes to the Care units is also 

proposed. The objective is the minimization of the costs of the project and the 

inventory-on-hand value of the system. 

2 - Solution methods for the periodic location-routing 

problem 

Caroline Prodhon, University of Technology of Troyes, 12 rue 

Marie Curie, 10000, Troyes, France, caroline.prodhon@utt.fr 

The well-known Vehicle Routing Problem (VRP) has been deeply studied over 

the last decades. Nowadays, generalizations are developed toward tactical or 

strategic decision levels but not both. The tactical extension or Periodic VRP 

(PVRP) plans a set of trips over a multiperiod horizon. The strategic extension 

or Location-Routing Problem (LRP) is motivated by interdependent depot location 

and routing decisions. The goal here is to present the very recent methods 

that solve the Periodic LRP, a combination of the PVRP and LRP into an even 

more realistic problem covering all decision levels. 

3 - A cutting plane approach for the single truck and trailer 

routing problem with satellite depots (STTRPSD) 

Juan G. Villegas, LOSI, Universite de Technologie de 

Troyes/Universidad de los Andes, 12, rue Marie Curie, BP 2060, 

10010, Troyes, France, juan_gmo_vr@yahoo.com, Jose M. 

Belenguer, Enrique Benavent, Antonio Martinez Sykora, 

Christian Prins, Caroline Prodhon 

In the STTRPSD a truck with a removable trailer based at a main depot serves 

the demand of a set of customers reachable only by truck. Thus, before serving 

the customers in truck routes, it is necessary to detach the trailer at appropriate 

parking places and to transfer products between the truck and the trailer. We 

present a two index formulation of the STTRPSD and valid inequalities that 

are used within a cutting plane method to produce lower bounds, and to solve 

the problem with branch and cut. The results are compared with upper bounds 

found by GRASP/VND and ILS on 32 random instances 

4 - A hybrid GRASP x Path Relinking for the Two-Echelon 

Location Routing Problem 

Viet Phuong Nguyen, LOSI - Université de Technology de 

Troyes, Troyes, France, viet_phuong.nguyen@utt.fr, Christian 

Prins, Caroline Prodhon 

This paper presents a hybrid between GRASP and Path Relinking to solve the 

Two-Echelon Location Routing Problem (LRP-2E). The GRASP reinforced by 

a Learning Process uses three constructive heuristics to generate the initial solutions. 

The Path-relinking adds a memory mechanism by combining intensification 

strategy and post-optimization. Our method uses local searches structured 

by a Variable Neighbourhood Descent (VND). Computational results confirm 

the efficiency of this approach on two sets of LRP-2E instances. Furthermore 

it is competitive with the best meta-heuristic published for the LRP. 

7


� MA-16 


2.2.14 

Rolling stock and Re-scheduling 

Stream: Public Transport 


Chair: Markus Reuther, Optimization, Zuse-Institut Berlin, 

Takustrasse 7, 14195, Berlin, Germany, reuther@zib.de 

1 - Rolling Stock Rotation Planning for Intercity Rail Traffic 

Markus Reuther, Optimization, Zuse-Institut Berlin, Takustrasse 

7, 14195, Berlin, Germany, reuther@zib.de 

We consider one of the basic operational planning problems in public rail transport, 

the construction of a rolling stock schedule. The problem deals with the 

optimization of feasible rotations for individual railcars and, simultaneously, 

the composition of train sets from these railcars. In addition, we have to integrate 

serveral maintenance and regularity aspects. Modeling and computational 

preliminiary results for instances of our industrial partner, DB Fernverkehr AG, 

which operates approximately 1.300 trains in Europe per day, are presented. 

2 - Railway Rolling Stock Rescheduling with Rerouting of 

Passengers 

Gabor Maroti, Department of Decision and Information 

Sciences, Rotterdam School of Management, Erasmus 

University Rotterdam, Burg Oudlaan 50, 3062 PA Rotterdam, 

The Netherland, 3062 PA, Rotterdam, Netherlands, 

gmaroti@rsm.nl, Lars Nielsen, Leo Kroon 

In this presentation we describe disruption management processes for a passenger 

railway system. In a disrupted situation, the timetable, the rolling stock 

circulation, and the crew duties must be rescheduled. We focus on rescheduling 

the rolling stock circulation. In case of a disruption, the passengers may be 

willing to take a detour route around the disrupted area. Then the rolling stock 

circulation must be rescheduled in such a way that additional seating capacity 

is provided along the detour route. In this presentation we describe an iterative 

procedure that reroutes the passengers, and that modifies the rolling stock 

circulation accordingly. Computational results based on real-life instances of 

Netherlands Railways have shown that this procedure may substantially reduce 

their delays. 

3 - Rapid Transit Networks: Time Table and Rolling Stock 

Ángel Marín, Matemática Aplicada y Estadística, Universidad 

Politécnica de Madrid, E.T.S.Ingenieros Aeronáuticos, Plaza 

Cardenal Cisneros, 3, 28040, Madrid, Madrid, Spain, 

angel.marin@upm.es, Luis Cadarso 

In rapid transit networks, the daily operations management process includes 

two major tasks: 1. Train services Timetable (TT). 2. Rolling Stock (RS) 

assignment to the TT. The tasks are interdependent but are often solved sequentially 

due to restrictions on computational time and the intractability of an 

integrated approach. In our modeling approach we consider the integration of 

TT and RS. Some computational experiments will be presented. 

4 - Assignment of services in bus lines under congestion 

Esteve Codina, Statistics and Operational Research, UPC, Edifici 

C5, Desp 216 Campus Nord, 08034, Barcelona, Spain, 

esteve.codina@upc.edu, Ángel Marín, Francisco Lopez 

A model is presented for dimensioning the number of services in bus lines operating 

under congestioned situations, which may arise in case of disruption of 

a Rapid Transit Network. The model takes into account bus capacity limitations 

and fleet availability as well as the dwell times of buses at stations. Also, 

an analysis of the waiting time of passengers at bus stops is made with special 

emphasis on this factor on the model results. The model is formulated under a 

system-optimum point of view and a heuristic algorithm approach is developed 

for larger size networks. 

8 

� MA-17 


1.3.14 

Collaborative Planning I 

Stream: Transportation Planning 


Chair: Herbert Kopfer, Department of Business Studies & 

Economics, Chair of Logistics, University of Bremen, 

Wilhelm-Herbst-Strasse 5, 28359, Bremen, Germany, 

kopfer@uni-bremen.de 

Chair: Melanie Bloos, Chair of Logistics, Bremen University, 

Wilhelm Herbst Str.5, 28359 , Bremen, Germany, 

bloos@uni-bremen.de 

1 - Transportation Operations Planning and Cost Allocation 

in a Cooperative Scenario 

Andrea Nagel, Dept. of Information Systems, FernUniversität - 

University of Hagen, 58084, Hagen, Germany, 

andrea.nagel@fernuni-hagen.de, Giselher Pankratz, Hermann 

Gehring 

Cooperative scenarios in transportation planning usually have to cope with the 

task of solving an optimization problem, as well as finding a fair allocation of 

the costs among the partners. We identify and characterize these problems for a 

real-life cooperation of four producers in the food and beverages industry, who 

decided to coordinate their distribution activities by inter-organisational transportation 

planning. Furthermore, we present a solution method that has been 

implemented, integrating a GRASP heuristic with the Shapley value approach. 

Finally, we show computational results. 

2 - Allocating Cost of Service to Customers in Inventory 

Routing 

Okan Ozener, Industrial Engineering, Ozyegin University, 

Kusbakisi Cad No:2, Altunizade Uskudar, 34662, Istanbul, 

Turkey, orsan.ozener@ozyegin.edu.tr, Ozlem Ergun, Martin 

Savelsbergh 

Vendor managed inventory replenishment is a collaboration between a supplier 

and its customers where the supplier is responsible for managing the customers’ 

inventory levels. In our VMI setting, the supplier exploits synergies between 

customers, e.g., their locations, usage rates, and storage capacities, to reduce 

distribution costs. Due to the intricate interactions between customers, calculating 

a fair cost-to-serve for each customer is a daunting task. However, 

cost-to-serve information is useful when marketing to new customers, or when 

revisiting routing and delivery quantity decisions. We design mechanisms for 

this cost allocation problem and determine their characteristics both analytically 

and computationally. 

3 - Collaborative vehicle routing in a multi-depot environment 

Julia Rieck, Department for Operations Research, Clausthal 

University of Technology, Julius-Albert-Str. 2, 38678, 

Clausthal-Zellerfeld, Germany, julia.rieck@tu-clausthal.de 

Fierce competition urges carriers to cooperate. Particularly, medium-sized carriers 

only achieve the adequate area coverage by splitting transportation requests 

into multiple tasks (pick-up, line haul, delivery) that can be handled separately 

by different carriers. Hence, a carrier has to perform the delivery and 

pick-up services around the depot while minimizing the transportation costs. In 

order to improve the resulting set of single-depot solutions, we present a new 

collaborative method that tries to find a reassignment of tasks to carriers which 

decreases the overall transportation costs. 

4 - The evaluation of pickup and delivery requests in cases 

of asymmetric information 

Melanie Bloos, Chair of Logistics, Bremen University, Wilhelm 

Herbst Str.5, 28359 , Bremen, Germany, bloos@uni-bremen.de, 

Herbert Kopfer 

Collaborative transport planning aims at creating the most efficient allocation 

of requests to carriers for a groupage system. However, due to the nature of 

this system, only limited relevant information on the carriers’ current planning 

is available system-wide. Our research focuses on evaluation criteria that create 

an efficient solution despite restricted information on the carriers’ situation and 

we present initial results on the performance of evaluation criteria for individual 

requests.

� MA-18 


1.3.15 

New Achievements in Stochastic Models 

and Optimization 

Stream: Stochastic Modeling and Simulation 


Chair: Erik Kropat, Department of Computer Science, Universität der 

Bundeswehr München, Werner-Heisenberg-Weg 39, 85577, 

Neubiberg, Germany, erik.kropat@unibw.de 

1 - Inflation Derivatives: HJM Framework and Market Models 

Kwai-sun Leung, Systems Engineering and Engineering 

Management, The Chinese University of Hong Kong, Shatin, 

New Territories„ Hong Kong, N.A., Hong Kong, Hong Kong, 

ksleung@se.cuhk.edu.hk, Lixin Wu 

In this paper, we establish a Heath-Jarrow-Morton (HJM) type framework that 

governs the co-evolution of the term structure of both nominal and inflation 

rates. Pricing of inflation derivatives under this framework can be carried out 

similarly to that of nominal interest-rate derivatives under the classic HJM 

model. Based on the HJM framework, we further develop a market model 

with simple forward inflation rates using displaced diffusion processes, which 

results in closed-form pricing for inflation caplets and inflation swaptions. The 

smile model can also be developed based on the market model. 

2 - On some antagonistic game related to majority voting 

Michael Khachay, Ural Branch of RAS, Institute of Mathematics 

and Mechanics, S.Kovalevskoy, 16, 620990, Ekaterinburg, 

Russian Federation, mkhachay@imm.uran.ru 

Simple majority voting is a classical approach to aggregation of individual decisions 

suggested by a committee of experts. In this paper, stability of such a 

collective decision, s.t. exclusion of some fixed number of experts, is investigated. 

Let some given list L of decisions be accepted by some committee of q 

equivalent experts, and let some number k

MA-20 EURO 24 - Lisbon 2010 

� MA-20 


1.3.33A 

Cutting and Packing 1 

Stream: Cutting and Packing 


Chair: Jose Fernando Oliveira, Faculty of Engineering / INESC 

Porto, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465, 

Porto, Portugal, jfo@fe.up.pt 

1 - Radical-free phi-functions for 2D objects and their applications 

Tatiana Romanova, Department of Mathematical Modeling and 

Optimal Design, Institute for Mechanical Engineering Problems 

of the National Academy of Sciences of Ukraine, 2/10 Pozharsky 

St., 61046, Kharkov, Ukraine, sherom@kharkov.ua, Nikolai 

Chernov, Yuri Stoyan, A. Pankratov 

Phi-functions are used to solve packing, cutting, and covering problems. Our 

purpose is to construct phi-functions by using simple mathematical formulas 

without radicals. First we introduce a special class of basic objects and prove 

that any 2D object whose frontier is formed by circular arcs and line segments 

may be represented as a union of basic objects. Then we derive a complete 

class of radical-free phi-functions for all pairs of basic objects. Lastly we show 

how to form phi-functions for more general objects. A software package is 

developed based on these results. Some applications are given. 

2 - A constructive algorithm for leather nesting in the automotive 

industry 

Pedro Brás, Universidade do Minho, 4710-057, Braga, Portugal, 

bras.pedro@gmail.com, Cláudio Alves, J. M. Valério de 

Carvalho, Telmo Pinto 

We address the leather nesting problem in the context of an automotive company. 

In this 2-dimensional problem, irregular shapes (car seats components) 

have to be cut from a natural leather hide with holes, defects and quality zones. 

We propose a solution algorithm based on a constructive procedure. We discuss 

the different aspects of this procedure and explain the strategic options on 

which it is based. We also briefly describe the no-fit polygon method used to 

guarantee valid placements in the leather hides. Computational results on real 

instances are presented. 

3 - A Hybrid Meta-heuristic Approach for Non-standard 

Packing Problems with Additional Conditions 

Giorgio Fasano, Space Infrastructures & Transportation, Thales 

Alenia Space Italia, Str. Antica di Collegno 253, 10146, Turin, 

Italy, giorgio.fasano@thalesaleniaspace.com 

This work focuses on the orthogonal packing of tetris-like items within a nonrectangular 

domain (with forbidden zones), in the presence of additional conditions, 

such as balancing. The overall problem is formulated in terms of mixed 

integer programming. Since non trivial cases give rise to very large-scale instances, 

a hybrid meta-heuristic approach has been adopted to solve recursively 

the problem. It is based on the concept of abstract configuration, deriving from 

the relative position of items. An extension considers the 2D case of polygons 

from a global optimization point of view. 

4 - Dual feasible functions for vector packing problems 

Jürgen Rietz, Departamento de Produção e Sistemas, Centro de 

Investigação Algoritmi, Universidade do Minho, Campus de 

Gualtar, 4715-082, Braga, Portugal, juergen_rietz@gmx.de, 

Cláudio Alves, J. M. Valério de Carvalho, François Clautiaux 

Dual-feasible functions (DFFs) were successfully used to obtain fast valid 

lower bounds for the one-dimensional cutting stock problem. To accelerate 

the calculations, only maximal, especially extremal functions should be used. 

This approach works for the vector packing problem too, if the domain of the 

DFFs is replaced by a more-dimensional unit cube. We state necessary and sufficient 

conditions for such functions to be maximal respectively extremal and 

present some non-trivial examples. New DFFs for the problem are discussed, 

and computational results are reported. 

10 

� MA-21 


6.2.47 

OR in Practice I 

Stream: Software for OR/MS 


Chair: Ana Moura, Economics, Management and Industrial 

Engineering, University of Aveiro, Campus Universitário de 

Santiago, 3810-193 , Aveiro, Portugal, ana.moura@ua.pt 

1 - Optimizing Fire Station Locations for Istanbul 

Metropolitan Municipality 

Emel Aktas, Industrial Engineering Department, Istanbul 

Technical University, ITU Isletme Fakultesi, Macka, 34367, 

Istanbul, Turkey, aktasem@itu.edu.tr, Ozay Ozaydin, Sule Onsel, 

Burcin Bozkaya, Fusun Ulengin 

Istanbul is a densely populated city with 2000+ years of cultural heritage. We 

provide a max-cover type mathematical model with limited budget to help Municipality 

authorities determine new fire station locations in addition to existing 

ones. We solve this model to optimality using GAMS and increase existing fire 

coverage from 56% to 86%. We also consider unlimited budget, overall relocation, 

and increased emphasis on city’s cultural treasures. We use GIS to 

process all geographical input data, calculate network distances and coverage 

ratios, and visualize various solutions of our model. 

2 - Design of a spatially explicit model to optimize the selection 

of reforestation projects and the sizing of a detention 

dam to reduce peak runoff in a small watershed 

Jochen Breschan, Department of Environmental Sciences, ETH 

Zurich, CHN K73.1, Universitaetstr. 22, 8092, Zurich, 

Switzerland, breschan@env.ethz.ch, Hans Rudorf Heinimann, 

Richard Church 

We address the problem of reducing potential flood hazards within a small 

Alpine watershed, using combinations of 1) spatially explicit reforestation 

projects and 2) a detention dam at the outlet in order to reduce peak storm 

runoff. The objective is to optimize the location of reforestation projects and 

the sizing of the detention dam to cost-effectively meet targets of peak runoff 

reduction. We show how to cast this problem as a Mixed Integer-Linear Programming 

optimization model that is solved using CPLEX. Finally, we present 

an application of this model to the Vogelbach watershed (CH). 

3 - A two-stage packing procedure for a Portuguese trading 

company 

Ana Moura, Economics, Management and Industrial 


Santiago, 3810-193 , Aveiro, Portugal, ana.moura@ua.pt, 

Andreas Bortfeldt 

This work reports on the development of a prototypical decision support system, 

called Packing and Routing Optimizer (PRO), which is devised to solve 

several packing and routing problems for a Portuguese company. The daily distribution 

process is analyzed and three decision problems regarding automated 

decision support are determined. The performance of the solution approaches 

is evaluated by computational tests based on actual company data. The test results 

show that the system is able to help improving the daily decisions and to 

strengthen the flexibility in negotiations with customers. 

� MA-22 


3.1.10 

Teaching OR/MS 

Stream: Teaching OR/MS 


Chair: Susana Colaco, Departamento de Ciências Matemáticas e 

Naturais, Escola Suerior de Educação- Instituto politécnico de 

Santarém, Compexo Andaluz, Apartado 131, 2000, Santarem, 

Portugal, susana.colaco@ese.ipsantarem.pt

1 - Generalized implementation of Program Evaluation Review 

Technique and Critical Path Method using EXCEL 

Spreadsheet. 

Shruthi S Kumar, Telecommunication, PES Institute of 

Technology, 100 Ft Ring Road BSK III Stage, 560085, 

Bangalore, Karnataka, India, shruthulisha@gmail.com, 

Guruprasad Nagaraj 

This paper describes implementation of the traditional PERT/CPM algorithm 

for finding the critical path in a project network using a spreadsheet. The problem 

is of importance due to recent shift of attention to using the spreadsheet 

environment as a vehicle for delivering MS/OR techniques to end-users. The 

purpose of this paper is intended to explore the need for systematic planning in 

teaching the network model in Operations Research to the under graduate students. 

We intend to create this spreadsheet mainly as a challenge to see whether 

the algorithm could be implemented in Excel. 

2 - On using WinQSB software to solve Linear Programming 

problems on 11th grade High School classrooms 

Ana Paula Teixeira, Mathematics, UTAD / CIO, Portugal, 

ateixeir@utad.pt, Helena Monteiro 

In this work we will describe an experimental study involving one hundred 

and eight Portuguese students of the 11th grade, who attended to the classes 

of Mathematics A. The aim of this study is to investigate if the students are 

able to use WinQSB software to solve Linear Programming, LP, problems. We 

also intend to identify not only the opinion of the students about the resolution 

of LP problems with this software but also their main difficulties during this 

experience. Moreover, this study will help us to understand if it is possible to 

use WinQSB as a tool when teaching LP on High School. 

3 - Some experiences on the design and implementation 

of Operations Research courses in Web-based environments 

Angel A. Juan, Computer Science, Open University of Catalonia, 

Rambla Poblenou, 156, 08018, Barcelona, Spain, 

ajuanp@gmail.com, Teresa Oliveira, Maria Martinez, Amilcar 

Oliveira, Javier Faulin, Sven Trenholm 

Web-based technologies have driven the growth of new learning opportunities 

in the Operations Research (OR) arena. We analyze how these technologies 

facilitate the shifting to an educational paradigm which considers students as 

active and central actors in their learning process. Use of computer software 

and collaborative learning as methodological policies to increase students’ motivation 

for OR are also discussed, and some real experiences regarding the 

design and implementation of OR courses in Web-based environments are described. 

4 - Modeling operations research problems with middle 

school students 

Susana Colaco, Departamento de Ciências Matemáticas e 

Naturais, Escola Suerior de Educação- Instituto politécnico de 

Santarém, Compexo Andaluz, Apartado 131, 2000, Santarem, 

Portugal, susana.colaco@ese.ipsantarem.pt, Margarida Pato, 

Cecilia Rebelo 

This talk describes a classroom modeling tasks for middle school students using 

classical operations research problems, such as linear programming in single 

and multi-objective versions, set covering, set packing and set partitioning. 

Some students’ productions to model real world problems are presented. The 

potentialities of using these contexts to develop communication skills, mathematical 

reasoning and identification of connections among mathematical ideas 

will be discussed. 

� MA-23 


6.2.49 

MOO: Algorithms for Multi-Objective 

Combinatorial Optimization I 

Stream: Multi-Objective Optimization 


Chair: Luis Paquete, Department of Informatics Engineering, 

University of Coimbra, Polo II, 3030-290, Coimbra, Portugal, 

paquete@dei.uc.pt 


1 - Finite representation of nondominated sets in multiobjective 

linear programming 

Matthias Ehrgott, Engineering Science, University of Auckland, 

Private Bag 92019, 1001, Auckland, New Zealand, 

m.ehrgott@auckland.ac.nz 

Can a continuous set of nondominated solutions of a multiobjective linear programme 

be represented by a finite subset? We prove that a related optimisation 

problem is NP-hard and show that some earlier methods may perform badly. 

We propose a new combined method which overcomes the limitations of these 

methods. We show that our method computes a set of evenly distributed nondominated 

points for which the coverage error and the uniformity level can be 

measured. Finally, we apply this method to an optimization problem in radiation 

therapy and present illustrative results for some clinical cases. 

2 - Large neighborhood search for solving the multiobjective 

multidimensional knapsack problem 

Thibaut Lust, Laboratory of Mathematics and Operatonial 

Research, University of Mons, 9, rue de Houdain, 7000, Mons, 

Belgium, thibaut.lust@umons.ac.be, Jacques Teghem 

We present a large neighborhood search (LNS) to solve the multiobjective multidimensional 

knapsack problem. The LNS is integrated into the two-phase 

Pareto local search method (Lust and Teghem, 2007). Different ratios are used 

to identify the items that interfere in the generation of the neighbors. The neighborhood 

is solved with an exact method or with a heuristic method, depending 

on the size of the neighborhood. We show that we obtain results never reached 

before (about 90% of the non-dominated points of a 250 items instance can be 

generated in less than 45 seconds). 

3 - A polynomial time algorithm for a cardinality constrained 

multicriteria knapsack problem 

Florian Seipp, Mathematics, University of Kaiserslautern, 

Paul-Ehrlich-Str. 14 -459, 67663, Kaiserslautern, Germany, 

seipp@mathematik.uni-kl.de, Luis Paquete, Stefan Ruzika 

This talk is concerned with the cardinality constrained multicriteria knapsack 

problem. In this combinatorial optimization problem two binary weight functions 

and a real valued profit function on the items are given. The task is to 

choose k out of the n given items with the aim of minimizing the weights and 

maximizing the profit. Whereas the general multicriteria knapsack problem is 

known to be NP-hard, we propose an exact algorithm with polynomial running 

time for our problem. This algorithm computes all nondominated points by 

efficiently exploring the weight space. 

4 - Three Algorithms for Finding Mines in a Line 

Luis Paquete, Department of Informatics Engineering, University 

of Coimbra, Polo II, 3030-290, Coimbra, Portugal, 

paquete@dei.uc.pt, Jochen Gorski, Mathias Jaschob, Kathrin 

Klamroth 

We introduce three algorithms for the problem of finding mines in a line. In 

this problem we are given a line partitioned into small segments, the time taken 

to either search in or travel through each segment, as well as a score value assigned 

to each segment. The goal is to choose the segments to visit such that the 

sum of the corresponding scores is maximized and the total travel and search 

time is minimized. The two algorithms are based on dynamic programming 

approaches for the multi-criteria knapsack problem. The third algorithm solves 

a bi-criteria shortest path problem formulation. 

� MA-24 


6.2.50 

Bioinformatics I 

Stream: Computational Biology, Bioinformatics and 

Medicine 


Chair: Jacek Blazewicz, Instytut Informatyki, Politechnika 

Poznanska, ul.Piotrowo 2, 60-965, Poznan, Poland, 

jblazewicz@cs.put.poznan.pl 

Chair: Piotr Formanowicz, Institute of Computing Science, Poznan 

University of Technology, Piotrowo 2, 60-965, Poznan, Poland, 

piotr@cs.put.poznan.pl 

11


1 - A new approach for measuring and visualizing quality 

of 3D protein models 

Piotr Lukasiak, Institute of Computing Science, Poznan 

University of Technology, ul.Piotrowo 2, 60-965, Poznan, 

Poland, Piotr.Lukasiak@cs.put.poznan.pl, Krzysztof Fidelis, 

Jacek Blazewicz 

Nowadays, there are a lot of computational methods in bioinformatics that can 

build protein 3D models based on the sequence, however even if real 3D protein 

structure is known it is hard to clearly evaluate which method is the most 

efficient. Accurate assessment of quality of protein models can be a crucial 

point for evaluation of currently available 3D protein prediction methods. In 

our approach one based on the idea of local descriptors similarity building appropriate 

measure to distinguish between "good’ and "bad’ models followed by 

new ways of visualization of models quality. 

2 - Prediction of protein-protein interaction network 

Xiang-Sun Zhang, Academy of Mathematics and Systems 

Science, Zhongguancun East Road 55, 100190, Beijing, China, 

zxs@amt.ac.cn 

Protein-protein interaction network (PPIN) plays an indispensable role in systems 

biology research. In this research we predict the unknown part of the PPIN 

by not only collecting all predicted PPIs based on the Domain-Domain Interaction 

(DDI) information but also satisfying the characteristics of the PPIN as a 

complex network, such as a small-world network, a network without rich club. 

Parsimony principle is used to find a spanning DDI structure through solving 

an integer linear programming. 

3 - Minimal Information for Automated Protein Function 

Prediction 

Daniel Faria, Informatics, Faculty of Sciences, University of 

Lisbon, Campo Grande, Edifício C6, 1749-016, Lisboa, 

Portugal, dfaria@xldb.di.fc.ul.pt 

The BLAST algorithm has become the de facto standard for protein function 

prediction from sequence. However, its applicability is limited and its 

widespread use promotes the propagation of errors. Several machine learning 

approaches have been tested as alternatives to BLAST, but the main focus has 

been on accuracy rather than speed. However, these approaches can’t expect 

to rival BLAST unless they are extremely efficient. In this work we assess the 

discriminatory power of several simple representations of protein sequences by 

testing them with simple machine learning approaches. Our goal is to obtain 

very efficient classifiers that can rival BLAST in terms of both accuracy and 

speed. 

4 - A guide through the labyrinth of gene prioritization 

tools 

Francisco Bonachela Capdevila, KU Leuven, campus Kortrijk, 

B-8500, Kortrijk, Belgium, 

Francisco.BonachelaCapdevila@kuleuven-kortrijk.be, 

Léon-Charles Tranchevent, Daniela Nitsch, Bart De Moor, 

Patrick De Causmaecker, Yves Moreau 

Finding the most promising genes among large lists of candidate genes has 

been defined as the gene prioritization problem. In the last decade, several different 

computational approaches have been developed to tackle this challenging 

task. We review 18 computational solutions for human gene prioritization that 

are freely accessible as web tools and illustrate their differences. In addition 

we have developed a website containing detailed information about these and 

other tools, which is regularly updated. 

Web: http://www.esat.kuleuven.be/gpp 

� MA-25 


6.2.48 

Financial Mathematics and OR 1 

Stream: Financial Mathematics and OR 


Chair: Koray Simsek, Faculty of Management, Sabanci University, 

Orhanli, Tuzla, 34956, Istanbul, Turkey, ksimsek@sabanciuniv.edu 

1 - Robust portfolio construction for downside risk control 

12 

Woo Chang Kim, Industrial and Systems Engineering, KAIST, 

#3107, E2-2, KAIST, 335 Gwahangno, Yuseong-gu, 305-701, 

Daejeon, Korea, Republic Of, wkim@kaist.ac.kr 

In this paper, we propose an alternative way to construct a robust portfolio of financial 

product. We utilize the market information under the stressful times and 

estimate the parameters directly from the data. The empirical analysis indicates 

that the performance could be improved much compared to other approaches, 

while our model requires less parameters to estimated. 

2 - Stochastic Diffusion Processes and High Frequency 

Sampling Data 

Ahmet Can Inci, College of Business - Finance, Bryant 

University, 1150 Douglas Pike, 02917, Smithfield, RI, United 

States, ainci@bryant.edu 

Studies propose that higher frequency data provide a more accurate separation 

of continuous and jump components of a dynamic process. Stochastic processes 

may not be fully utilized for low frequency monthly/quarterly data since 

processes such as jumps may be smoothed out. In this study, the empirical 

performance of a sophisticated multi-country multi-state quadratic stochastic 

model is explored in the context of sampling frequency for the same sample 

period. Differences in the coefficient estimates of the parameters and the empirical 

performance of the model are documented. 

3 - Mathematical Programming Approaches for Generating 

p-Efficient Points 

Nilay Noyan, Manufacturing Systems/Industrial Engineering, 

Sabanci University, Orhanli, Tuzla, 34956, Istanbul, Turkey, 

nnoyan@sabanciuniv.edu, Miguel Lejeune 

The concept of p-efficiency is widely used to develop efficient methods to 

solve probabilistically constrained problems, such as the cash-flow matching 

problem. Those methods require the generation of p-efficient points (pLEPs). 

We consider a random vector characterized by a set of scenarios and generate 

pLEPs by solving a mixed integer programming problem. We propose to 

solve this challenging problem by a new mathematical programming framework, 

which involves solving a series of outer approximations. We present 

numerical results showing the computational efficiency of the proposed framework. 

4 - Performance Enhancements for Defined Benefit Pension 

Plans 

Koray Simsek, Faculty of Management, Sabanci University, 

Orhanli, Tuzla, 34956, Istanbul, Turkey, 

ksimsek@sabanciuniv.edu, John Mulvey 

Over the next 30-50 years, traditional corporate and government pension plans 

will encounter severe problems in many countries. Contributing factors include: 

demographic trends, low savings rates, and inefficient investment strategies. 

This paper takes up the last point, showing that a forward-looking asset 

liability management model can improve performance across many reward and 

risk measures. We approximate a multi-stage stochastic program via a set of 

robust policy rules. Empirically, we show that a duration enhancing overlay 

strategy improves performance during economic contractions. 

� MA-26 


3.1.11 

Mathematical Programming Approaches for 

Classification Problems 

Stream: Machine Learning and Its Applications 


Chair: Gurkan Ozturk, Industrial Engineering, Anadolu University, 

AU-MMF-Industrial Engineering, Iki Eylul Campus, 26480, 

Eskisehir, Turkey, gurkan.o@anadolu.edu.tr 

1 - A New Contribution to Mean Shift Outlier Model with 

Continuous Optimization 

Pakize Taylan, Mathematics, Dicle University, 21280, 

Diyarbakır, Diyarbakir, Turkey, ptaylan@dicle.edu.tr, Fatma 

Yerlikaya Ozkurt, Gerhard-Wilhelm Weber 

The outlier detection problem is one of the important problems in statistics 

since observation of outliers negatively affects the estimation and the inference. 

There are several outlier detection methods. One of these methods is 

given by Mean Shift Outlier model. In our study, Mean Shift Outlier model 

is constructed with Tikhonov Regularization. Here, our new model is solved 

by continuous optimization techniques, in particular, via the elegant framework 

of conic quadratic programming which becomes an important complementary 

technology and alternative to the outlier detection methods.

2 - A Mixed-Integer Programming Approach to Multi-Class 

Data Classification Problems 

Fadime Uney-Yuksektepe, Industrial Engineering, Istanbul 

Kultur University, E5 Karayolu Londra Asfalti Uzeri, Atakoy 

Kampusu, 34156, Istanbul, Turkey, f.yuksektepe@iku.edu.tr, 

Metin Turkay 

In this study, a novel MILP based hyper-box enclosure approach is presented 

for multi-class data classification problems. In order to deal with large data 

sets, a three-stage mathematical programming based approach is developed for 

training part. The efficiency of the method is tested by the simple distance 

based testing algorithm. The computational results on the UCI Repository data 

sets show that the simplicity and accuracy of the proposed method provides 

scientific insight into the multi-class data classification problems. 

3 - A heuristic classification algorithm for large scale problems 

Mehmet Tahir Ciftci, Logistics, Eczacibasi Vitra, 11300, Bilecik, 

Bozuyuk, Turkey, tahir.ciftci@eczacibasi.com.tr, Gurkan Ozturk 

In last years, polyhedral conic functions are used in developing successful classification 

algorithms. The performance of the classification algorithm highly 

depends on the parameters of function used. One of the these parameters is the 

vertex point of the graph of polyhedral conic function under consideration. In 

this study, several heuristic approaches based on different center point determination 

strategies are proposed. These approaches are then used in algorithms 

for solving large scale clasification problems. Clasification performances are 

reported for the literature test problems. 

4 - A novel mathematical programming approach for classification 

based on linear and conic functions 

Gurkan Ozturk, Industrial Engineering, Anadolu University, 


Eskisehir, Turkey, gurkan.o@anadolu.edu.tr, Refail Kasimbeyli 

Mathematical programming approaches for classification problems are generally 

based on linear functions. Conic functions are defined by expanding linear 

functions with different norms. In this study we use a two-step approach. In 

the first step, linear and conic functions are obtained by solving linear programming 

problems. Then in the second step, an integer programming model 

is constructed and solved. Eventual separating function is defined as a pointwise 

minimum of selected functions. The performance of the proposed method 

is shown on test problems. 

� MA-27 


8.2.06 

Financial Optimization 1 

Stream: Financial Optimization 


Chair: Cesarone Francesco, La Sapienza Universita’ di Roma, 

Facolta’ di Economia, 00161, Roma, 

francesco.cesarone@uniroma1.it 

1 - Log-Robust Portfolio Management 

Aurelie Thiele, Industrial and Systems Engineering, Lehigh 

University, 200 West Packer Ave Room 329, 18015, Bethlehem, 

PA, United States, aurelie.thiele@lehigh.edu, Ban Kawas 

We present a robust optimization approach to portfolio management under 

uncertainty that (i) builds upon the well-established Lognormal model for 

stock prices while addressing its limitations, and (ii) incorporates the imperfect 

knowledge on the true distribution of the continuously compounded rates 

of return, i.e., the increments of the logarithm of the stock prices, in an intuitive 

manner. We derive theoretical insights into the worst-case uncertainty and 

the optimal allocation. We also consider extensions to short sales and discuss 

risk-return trade-offs. 

2 - Pension Fund ALM with multiple asset classes 

Katharina Schwaiger, Economics and Finance, Brunel 

University, Kingston Lane, Brunel University, UB8 3PH, 

Uxbridge, Middlesex, United Kingdom, 


katharina.schwaiger@brunel.ac.uk, Gautam Mitra, Nicola 

Spagnolo 

Asset and Liability Management models for pension funds have been recently 

recast as Liability Driven Investment models. The industry has interpreted this 

as different allocation strategies: i.e. 100% fixed income, fixed proportions 

of equities and fixed income plus overlay strategies. We propose asset allocation 

strategies of two-stage stochastic programming and integrated chanceconstrained 

programming models and compare their performance to static fixed 

mix strategies. The stochastic programming models trade-off initial injected 

cash and PV deviations between assets and liabilities. 

3 - Enhanced indexation using SSD 

Diana Roman, Mathematics, Brunel University, UB8 3PH, 

Uxbridge, United Kingdom, diana.roman@brunel.ac.uk, Gautam 

Mitra, Csaba I. Fabian 

We analyse the effectiveness of SSD models in the context of enhanced indexation. 

Enhanced indexation models are strategies for asset allocation that seek 

to outperform an index. Most of the index tracking / outperforming models aim 

to tackle the computational difficulty posed by limiting the number of stocks 

(which leads to MIP or QMIPs). We show that the SSD-based models naturally 

select a small number of stocks, which makes unnecessary the imposition 

of cardinality constraints. The effectiveness in over-performing the index is 

shown using 3 datasets drawn FTSE100, Nikkei 225 and SP500. 

4 - A new portfolio selection approach: Models and Algorithms 

Cesarone Francesco, La Sapienza Universita’ di Roma, Facolta’ 

di Economia, 00161, Roma, francesco.cesarone@uniroma1.it, 

Andrea Scozzari, Fabio Tardella 

In this paper we present a comparison study between different portfolio models. 

We focus on the Markowitz, the Mean Absolute Deviation, and the CVaR 

models, with the introduction of quantity and cardinality constraints. In particular, 

we provide a new solution method for the limited asset Markowitz model 

based on a reformulation as a Standard Quadratic Program and on some recent 

theoretical results. We report optimal solutions of some unsolved test problems 

from Beasley’s OR Library. Finally, we present a new model based on CVaR 

minimization with constraints on the correlation among stocks. 

� MA-28 



Stochastic Integer Programming 

Stream: Stochastic Programming 1 


Chair: Gloria Perez, Applied Mathematics and Statistics and 

Operational Research, Universidad del País Vasco, Science and 

Technology Faculty, Barrio Sarriena s/n, 48940, Leioa, Spain, Spain, 

gloria.perez@ehu.es 

1 - On BFC-MSMIP strategies for Twin Node Family branching 

selection and bounding for multistage stochastic 

mixed integer programming 

Gloria Perez, Applied Mathematics and Statistics and 

Operational Research, Universidad del País Vasco, Science and 

Technology Faculty, Barrio Sarriena s/n, 48940, Leioa, Spain, 

Spain, gloria.perez@ehu.es, Laureano Fernando Escudero, María 

Araceli Garín, María Merino 

In the Branch-and-Fix Coordination (BFC-MSMIP) algorithm for solving multistage 

stochastic mixed integer programming problems, we find it crucial to 

decide the stages where the nonanticipativity constraints are explicitly considered 

in the model. We present a scheme for obtaining strong bounds and 

branching strategies for the Twin Node Families to increase the efficiency of 

BFC-MSMIP. Also we explain the computational description using the free optimization 

software COIN-OR, COmputational INfrastructure for Operations 

Research and it is reported some computational experience. 

2 - FRC: A heuristic extension of the Branch-and-Fix Coordination 

approach for solving very large multistage 

mixed 0-1 stochastic problems 

Celeste Pizarro Romero, Dpto. de Estadística e Investigación 

Operativa, Universidad Rey Juan Carlos, Escuela de CC. 

Experimentales y Tecnología, 28933, Móstoles, Spain, 

13


celeste.pizarro@urjc.es, Antonio Alonso-Ayuso, Laureano 

Fernando Escudero, Pablo Olaso 

We present a Branch-and-Fix Coordination (BFC) framework for solving largescale 

multistage mixed 0-1 stochastic problems. A mixed 0-1 model for each 

scenario cluster is considered plus the non-anticipativity constraints that equate 

the 0-1 and continuous so-called common variables from the same group of 

scenarios in each stage. We propose the heuristic extension of BFC, so-called 

Fix-and-Relax Coordination (FRC) algorithm, to exploit the characteristics of 

the non-anticipativity constraints of the stochastic model for solving very largescale 

instances. Computational results are reported. 

3 - Subgradient scheme, Volume algorithm, Progressive 

Hedging algorithm and Dynamic Constrained Cutting 

plane scheme in Langragean Decomposition Models 

María Araceli Garín, Applied Economy III, UPV/EHU, 

Lehendakari Aguirre 83, 48015, Bilbao, Bizkaia, Spain, 

mariaaraceli.garin@ehu.es, Laureano Fernando Escudero, Gloria 

Perez, Aitziber Unzueta 

We represent the two stage stochastic mixed 0-1 problem by a splitting variable 

representation of the DEM. In order to satisfy the non-anticipativity constraints, 

we consider the Lagrangean decomposition and compare the solution 

given by different schemes based on dual Lagrangean problems, where the Lagrange 

multipliers are updated by using the subgradient method, the volume 

algorithm, the progressive hedging algorithm and the dynamic constrained cutting 

plane scheme. Lagrangean decomposition is proposed for satisfying the 

non-anticipativity constraints. Computational results are reported. 

4 - Multistage Stochastic Programming Problem: Decomposition, 

Stability and Empirical Estimates 

Vlasta Ka˘nková, Econometrics, Institute of Information Theory 

and Automation of ASCR, Pod Vodárenskou vĕzí 4, Praha 8, CZ 

18208, Prague, Czech Republic, Czech Republic, 

kankova@utia.cas.cz 

A multistage stochastic programming problem can be mostly introduced as a 

system of parametric one stage optimization problems. Consequently, results 

(achieved for one stage problems) can be employed to investigate stability and 

empirical estimates of the multistage problems. The aim of the contribution is 

to investigate assumptions under which this approach is suitable and the convergence 

rate of empirical estimates is acceptable. The introduced assumptions 

cover heavy tailed distributions and conditions guaranteeing nonempty constraints 

set of decomposed problems. 

� MA-30 


8.2.13 

MCDA II: Multicriteria ranking and 

classification vs portfolio decision analysis 

(Panel) 

Stream: MCDA II: Axiomatic Basis, Meaningfulness, and 

other Issues 

Panel session 

Chair: Alec Morton, Management/ Operational Research, London 

School of Economics, Houghton St, London, wc2a2ae, London, 

England, United Kingdom, a.morton@lse.ac.uk 

1 - Multicriteria ranking and classification vs portfolio decision 

analysis 

Alec Morton, Management/ Operational Research, London 


England, United Kingdom, a.morton@lse.ac.uk, Christian 

Stummer, Valerie Belton, Ahti Salo, José Rui Figueira, Theodor 

Stewart 

Three very similar problems are as follows: rank n multiattributed items; classify 

the items as "yes", "no" and "maybe"; and select a number of these items 

such that the total cost of the selected items is less then some budget. The 

former two sorts of problems have been studied under the heading of ranking 

and classification, and the last sort of problem have studied under the heading 

portfolio decision analysis. Are these problems as similar as they appear, and 

if so, why do we think about them differently? 

14 

� MA-31 


8.2.15 

Societal Complexity and Climate 

Stream: Methodology of Societal Complexity 





1 - Sustainability assessment for solar plant and wind 

power projects in the con co island, quang tri province, 

vietnam 

Le Trinh Hai, Human Ecology, Vrije Universiteit Brussel, 103 

Laarbeeklaan, 1090, Brussels, Viet Nam, letrinhhai@gmail.com, 

Pham Hoang Hai, Nguyen Ngoc Khanh, Jan Kuilman, Dinh Thi 

Phuong Loan, Luc Hens 

Background: Clean Development Chemical Programme mentioned in Article 

12 of the Kyoto Protocol to the United Nations Framework Convention on Climate 

Change is a hot issue, in particular in developing countries (e.g. Vietnam) 

to limit or reduce their greenhouse gas emissions. Solar plant and wind power 

are two important aspects in this programme. 

Aim: To assess sustainability of the planning projects on solar plant and wind 

power on the island of Con Co Quang Tri, Vietnam as a case study. 

Methods: The analytic hierarchy process and multi-criteria/indicator assessment 

are applied to evaluate the planning projects. 

Results: Scenario I provides the sustainability scores of composite sustainable 

development index (ICSD: 0.509) for the solar plant and of for the wind plant 

(ICSD: 0.490). In Scenario II both power projects score high (the solar plant 

ICSD: 0.86 and the wind power ICSD: 0.838). 

Conclusions: The multi-criteria/indicator assessment allows to evaluate the two 

planning projects on the Con Co island. Based on the results, the projects are 

not only help to reduce the environmental pollution in particular in Scenario II 

and also to improve the local life, especially in Con Co as a poor island and important 

district of the province of Quang Tri. In addition, it will possibly help 

the decision-makers are able to understand the values of the energy systems as 

part of Clean Development Mechanism. 

2 - Online web based solution for environmental sustainable 

development 

Unal Akyuz, Software Engineering, Bahcesehir University, 

Bahcesehir University, Faculty of Engineering, 34353, ˙Istanbul, 

Turkey, unal.akyuz@stu.bahcesehir.edu.tr, Onur Yazici, Okan 

Kilic, Gizem Akturk, Erkan Mert, Boran Uzunlu, Buse Ozturk, 

Sureyya Ozogur-Akyuz 

The aim of this multidisciplinary research is to help reducing the green house 

effect by constituting an online web solution with a mathematical programming 

approach which calculates Co2 emission and evaluates the necessary number 

of trees to recover Co2 emission, and further calculates the area of the land 

needed with respect to the region and the type of the tree by the help of Integer 

Linear Programming (ILP). 

3 - Eco-dimension of Tourist Service Quality 

Majda Bastic, Faculty of Economics and Business, University of 

Maribor, Razlagova 14, 2000, Maribor, Slovenia, 

majda.bastic@uni-mb.si 

Tourism has not only contributed to increase of service export but also to 

climate changes. Climate changes and their impact on a quality of our life 

emerged a new increasing segment of tourists who demonstrate a particular 

sensitivity to the environment. An eco-component has not been included in 

the existing scales such as SERVQUAL, SERVPERF or ECOSERV. The objective 

of this paper was to investigate the service quality expectations of the 

ecotourists in order to develop eco-scale refers to environmentally responsible 

behaviour of hotels’ management. 

4 - Compram Method for Handling Societal Problems - 

Case Study in a Brazilian Research and Development 

(R&D) Program for the electric sector 

André Bacellar, Research and Development and Efficiency 

Superintendence (SPE), Brazilian Electricity Regulatory Agency 

(ANEEL), SQN 403 Bl.M apto.107, 70835130, Brasília, DF,

Brazil, ambacellar@yahoo.com.br, Aurelio Calheiros de Melo 

Junior 

The Brazilian Electricity Regulatory Agency — ANEEL establishes guidelines 

and instructions that regulate investments made by companies in R&D Program 

for the Brazilian Electric Energy Sector. The main problem in ANEELs point 

of view is to reduce the utilities R&D financial risk as they develop research 

projects and at the same time reduce bureaucratic procedures, especially to the 

expenditures control. More control means less risk, but with more transaction 

costs. Lots of others variables increase the complexity. 

� MA-32 


8.2.17 

Transportation planning in forest products 

industry 

Stream: OR in Agriculture and Forest Management 


Chair: Manfred Gronalt, Institute of Production and Logistics, 

University of Natural Resources and Applied Life Sciences, 

Feistmantelstr. 4, A-1180, Vienna, Austria, 

Manfred.Gronalt@boku.ac.at 

1 - Combining logs and biomass transportation planning 

in a Portuguese integrated pulp and paper company 

Alexandra Marques, Instituto Superior de Agronomia, 

Universidade Técnica de Lisboa, Tapada da Ajuda, 1349-017, 

Lisboa, Portugal, alexmarques@isa.utl.pt, Jose Borges 

The high economical impact of transportion over total wood production cost 

fosters the development of innovative optimization approaches. In this article 

we propose an heuristic approach for combined logs and biomass transportation 

scheduling in an vertically integrated pulp and paper company. It is a particular 

case of the Log-truck scheduling problem aiming to find minimum cost 

routes for transporting all product assortments from the forest to transformation 

centers or terminals, according to the wood delivery plan defined beforehand. 

Innovative time windows constraints are considered. 

2 - OR in support of forest products management and supply 

Alexandra Marques, Instituto Superior de Agronomia, 

Universidade Técnica de Lisboa, Tapada da Ajuda, 1349-017, 

Lisboa, Portugal, alexmarques@isa.utl.pt, Mikael Rönnqvist, 

Sophie D’Amours, Jose Borges 

Optimization models and methods are extensively used for wood procurement 

planning, encompassing forest planning, harvesting, transportation, storing and 

plant supply. They foresee the optimal sustainable use of forest resources and 

available equipments, thus ranging from strategic planning to equipments allocation 

and operations daily scheduling. Their role in forest decision making is 

perceived, as they integrate individual supply chain agents’ goals. This paper 

discusses models and solution approaches applied for wood procurement. The 

data flows among the supply chain are also addressed. 

3 - The impact of using foldable containers in wood transport 

Manfred Gronalt, Institute of Production and Logistics, 


Feistmantelstr. 4, A-1180, Vienna, Austria, 

Manfred.Gronalt@boku.ac.at, Patrick Hirsch, Jan Zazgornik 

In this study the impact of two different transport modes on vehicle routing is 

reported for cases in forest wood supply chains. The daily routing of vehicles 

with traditional round log trucks is compared to a more flexible concept of using 

foldable containers as unified loading equipment. Test instances are solved by a 

modified nearest neighbor insertion heuristic and improved by a metaheuristic 

called dynamic Tabu Search with Alternating Strategy, where neighborhoods 

are changed dynamically. The efficiency of both modes is measured in terms 

of duration and costs, respectively. 

4 - Design of a biomass supply chain — an integrated approach 

Tiago Gomes, Universidade do Minho, Portugal, 

tiago.gomes@dps.uminho.pt, Filipe Alvelos, Sameiro Carvalho 


Successful biomass supply chain management requires many decisions relating 

to the supply of raw material, such as, the harvest of raw material, wood chipping, 

facility location, transportation and inventory decisions. The definition of 

an integrated approach to deal with such a diversity of problems aims to raise 

the supply chain surplus. In this research work, an integer programming model 

to support tactical and operational decisions in a biomass supply chain is proposed 

with the objective of minimizing total costs and satisfies the customers 

demand. 

� MA-33 


8.2.19 

Energy Modeling 

Stream: Energy, Environment and Climate [c] 


Chair: Reinhard Madlener, Faculty of Business and Economics / 

E.ON Energy Research Center, RWTH Aachen University, 

Mathieustrasse 6, 52074, Aachen, Germany, 

rmadlener@eonerc.rwth-aachen.de 

1 - A Multi-Sector Energy-Economy Model for Turkey 

Bora Kat, The Scientific and Technological Research Council of 

Turkey, Atatürk Bulvarı No:221, Kavaklıdere, 06100, Ankara, 

Turkey, borakat@gmail.com, Caglar Güven, Ebru Voyvoda 

In this study, an energy-economy model within the optimization framework is 

proposed. The model consists of a detailed representation of the energy activities 

and disaggregates the rest of the economy into five main sectors unlike 

the conventional energy-economy optimization models in which the rest of the 

economy is represented in an aggregated fashion. Besides the ability of representing 

the energy-economy interactions, the model is extended as it can assess 

the impacts of each scenario on the environmental variables, specifically on the 

GHG emissions. 

2 - Sustainable Development Goals and the Use of Modern 

Portfolio Theory for Optimizing the Power Generation 

Mix at the Company Level 

Barbara Glensk, Faculty of Business and Economics, E.ON 

Energy Research Center, RWTH Aachen University, Institute for 

Future Energy Consumer Needs and Behavior, Mathieustr.6, 

52074, Aachen, North Rhine-Westphalia, Germany, 

bglensk@eonerc.rwth-aachen.de, Reinhard Madlener 

Most countries and energy providers have plans to change their power generation 

mix, aiming at the development and pursuance of sustainable energy 

strategies, which typically imply the reduction of CO2 emissions and an increased 

share of power from renewable energy sources. These goals ought to 

be guided through optimal capacity investment planning based on risk-return 

considerations. In this paper, by using variants of modern portfolio theory, we 

investigate the impact of new investments on the efficiency of the power generation 

mix of E.ON, one of the largest energy companies in Europe. 

3 - A coupled geo-spatial energy - air quality assessment 

model for the region of Luxembourg 

Laurent Drouet, CRTE, CRP Henri Tudor, Technoport 

Schlassgoart, 66, rue de Luxembourg, 4221, Esch-sur-Alzette, 

Luxembourg, ldrouet@gmail.com, Daniel Zachary, Ulrich 

Leopold, Lara Aleluia Reis 

We present a general method to couple an energy model and an air-quality 

model. The energy model (MARKAL-like) aims to minimize the energy cost 

at a given level of emissions. The air quality model simulates the chemical 

reactions to produce ozone for worst-case episodes. We build a decomposition 

problem to find the optimal energy system with a constraint on the ozone 

concentrations. An oracle-based optimization method is implemented to solve 

the coupled model using the cutting plane algorithm (Proximal-ACCPM). We 

report on preliminary results of a prototype version developed for Luxembourg. 

4 - Energy visions, resource allocation and multi-criteria 

assessment for a stakeholder discourse 

Evelina Trutnevyte, Institute for Environmental Decisions (IED), 

Natural and Social Science Interface (NSSI), ETH Zurich, 

Universitätstrasse 16, CHN J 70.1, 8092, Zurich, Switzerland, 

15


evelina.trutnevyte@env.ethz.ch, Michael Stauffacher, Alexander 

Scheidegger, Roland W. Scholz 

Decisions for the future energy system are influenced by discourses among 

stakeholders. These discourses must be informed by analytical expertise on 

the allocation of scarce energy resources and the assessment of implications. 

We present a methodology, which links (i) visions (overall preferences) of 

stakeholders, (ii) representative sets of resource allocation options for the visions, 

and (iii) multi-criteria assessment to appraise implications. Such analysis 

improves support for strategic discourses among multiple stakeholders. The 

methodology is applied on a case of a Swiss community. 

� MA-34 


8.2.23 

Solution Approaches for Lot-sizing 

Problems I 

Stream: Lot-sizing and Scheduling, Economic Order 

Quantity 


Chair: Marco Caserta, Economics, University of Hamburg, 

Von-Melle-Park, 5, 20146, Hamburg, Germany, 

marco.caserta@uni-hamburg.de 

1 - Genetic Algorithm Based Approaches for the Capacitated 

Lot Sizing Problem with Set-up Carryover and 

Backordering 

Hacer Guner Goren, Department of Industrial Engineering, 

Dokuz Eylul University, Buca Campus, Buca, Izmir, Turkey, 

hacer.guner@deu.edu.tr, Semra Tunali, Raf Jans 

This production planning problem deals with multiple products produced on 

a single machine. A setup is assumed to be carried over from one period to 

the next and the partial sequencing of the first and last product is incorporated. 

When the demand of a period can not be satisfied, it is backordered. In this 

study, a heuristic hybrid approach combining Genetic Algorithms (GAs) and 

Fix-and-Optimize heuristic is proposed to solve capacitated lot sizing problem 

with set-up carryover and backordering. The performance of the proposed approach 

is compared with pure GAs by generating various problem instances. 

2 - An exact solution approach for the discrete lot-sizing 

and scheduling problem with identical parallel resources 

Céline Gicquel, Laboratoire Genie Industriel, Ecole Centrale 

Paris, Grande Voie des VIgnes, 92290, Chatenay-Malabry, 

France, celine.gicquel@ecp.fr, Laurence Wolsey, Michel 

Minoux 

We consider the discrete lot-sizing and scheduling problem with identical parallel 

resources. We propose to solve this optimization problem using a tight 

MILP formulation and a commercial solver. We first study two alternative formulations 

of the problem: one involving machine-specific binary variables and 

one using system-wide integer variables. We then derive a new family of valid 

inequalities and use it to devise a Cut & Branch algorithm. Our computational 

experiments show that the proposed approach is effective at solving large instances 

of the problem within a reasonable computation time. 

3 - Lower bounds for the capacitated lot sizing problem 

with setup time 

Silvio Araujo, Departamento de Ciências da Computação e 

Estatística-DCCE, Universidade Estadual Paulista-UNESP, R. 

Cristovao Colombo, 2265 - Jd Nazareth, 15054-000, São José do 

Rio Preto, São Paulo, Brazil, saraujo@ibilce.unesp.br, Zeger 

Degraeve, Raf Jans 

We study the Capacitated Lot Sizing Problem with Setup Times (CLST) and 

present some theoretical results regarding the quality of the lower bounds obtained 

by simultaneous per item and per period decompositions of the problem. 

Moreover, we develop and compare, by presenting some computational results, 

different algorithms based on column generation and Lagrange relaxation for 

finding lower bounds for the problem. 

16 

4 - Capacitated Lot Sizing with Setup Times: Decompositions 

of the Simple Plant Location Formulation 

Bert De Reyck, Management Science & Operations, London 

Business School, Regent’s Park, NW1 4SA, London, United 

Kingdom, bdereyck@london.edu, Zeger Degraeve, Ioannis 

Fragkos 

Implementing Dantzig—Wolfe Decomposition (DW) and Lagrange Relaxation 

(LR) in extended formulations may be computationally challenging due to the 

inherent degeneracy these formulations often exhibit. We apply DW/LR to the 

Simple Plant Location formulation of the capacitated lot sizing problem with 

setup times. Its per—period decomposition gives rise to a series of subproblems 

that do not have the integrality property. Thus, we obtain an improved 

lower bound. We discuss the subproblem solution and show that cutting off 

some alternative optimal solutions improves computational efficiency. 

� MA-35 


6.2.46 

Teaching Soft OR and PSMs 

Stream: Soft OR and Problem Structuring Methods 


Chair: Leroy White, Management, University of Bristol, Social 

Science, 8 Woodland RD, BS8 1TN, Bristol, United Kingdom, 

leroy.white@bris.ac.uk 

Chair: John Mingers, Kent Business School, Kent University, CT2 

7PE, Canterbury, Kent, United Kingdom, j.mingers@kent.ac.uk 

1 - Soft Systems Methodology (SSM): a reconfiguration 

Ion Georgiou, IMQ, Fundação Getulio Vargas, Rua Itapeva 474 

(9 andar), Bela Vista, 01332000, Sao Paulo, Sao Paulo, Brazil, 

iongeorgiou@gmail.com, Maria Fernanda Mendes 

Teaching Soft Systems Methodology (SSM) is a non-trivial exercise due to 

a number of reasons. There is no textbook approach currently available, although 

illustrative case studies of varying pedagogical utility do exist. The 

methodology introduces elaborate conceptual novelty and technical language, 

although some of its tools are arguably but formalizations of decision making 

processes that might otherwise receive scant attention. The methodology 

emerged through an intricate historical and philosophical background, although 

one that weaved theoretical developments with perceived practical necessities. 

The literature presents the methodology as available in two distinct modes, although 

only one has received the most scholarly attention. And, perhaps most 

challenging of all, its inherent flexibility renders it difficult to categorize: it can 

be taught as an exploratory approach, a formal decision making approach, an 

organizational learning vehicle, and it may even be used in part, in whole, iteratively 

or not, and even in conjunction with other methods. Indeed, the very 

question of whether SSM is a methodology or a method is debatable and will 

influence the pedagogic approach taken. Finally, it is an approach designed to 

help decision makers tread through the complexity and uncertainty generated 

by systems of interrelated problems. Where a usual recourse in such situations 

is to computerized decision support systems, none are available to accompany a 

process led by SSM. A formal pedagogical approach to SSM will be presented. 

The aim is to provide teachers and trainers with the confidence to communicate 

the relevance of the approach in a transferable manner that students/clients can 

then use in their careers. The focus will be on the so-called Mode 1 of SSM, 

this being the one most heavily discussed in the literature. However, based 

on recent research, a complete reconfiguration of the structure of this Mode 

will be presented. As such, one novelty lies in how SSM may be understood 

through a new configuration, one that emphasizes decision making, but also 

one that emphasizes the interrelated logic between all the tools and concepts 

used in the methodology. In addition, a useful multimethodological input will 

be provided with the use of cognitive mapping in one particular part of SSM. In 

sum, the paper concentrates on teaching SSM as a formal decision making approach 

which simultaneously enables understanding of the entire methodology 

and its related parts. From such a foundation, the methodology’s facilitative 

potential for exploratory and organizational learning purposes, as well as its 

multimethodological flexibility, can be appreciated. 

2 - Teaching supporting strategy: a multimethodological 

perspective using ’soft’ and ’hard’ tools 

Frances O’Brien, Warwick Business School, University of 

Warwick, Gibbet Hill Road, CV4 7AL, Coventry, United

Kingdom, Frances.O-Brien@wbs.ac.uk, Robert Dyson, Martin 

Kunc 

The development of a course promoting the use of tools to support a strategy 

process is described. Tools, from a number of disciplines, including OR/MS, 

cover both the soft & hard spectrum & can be used either individually or in 

combination. 

The content of the course will be presented along with how delivery differs 

across student groups. A noticeable feature of the course has been the development 

of multimethodological approaches; examples will be presented to 

illustrate some of the soft tools taught. The paper ends with reflections on the 

current challenges faced by course designers. 

3 - Teaching Soft OR: Role Play Simulation and Social 

Learning 

Leroy White, Management, University of Bristol, Social Science, 

8 Woodland RD, BS8 1TN, Bristol, United Kingdom, 


This paper describes the use of role play simulation (RPS) in teaching soft OR 

methods. RPS provides an opportunity for active learning. The paper will describe 

the use of a simulation designed for teaching soft OR and will provide 

reflections to show the efficacy of simulations as active learning techniques. 

The reflections will also provide an outline of how the knowledge of Soft OR 

builds on itself over time and that people learn from one another, via observation, 

imitation, and modelling. The process of social learning (learning fromm 

the learning of others) will be described. 

4 - Can Problem Structuring Methods be taught? 

Ashley Carreras, Leicester Business School, De Montfort 

University, Marketing, The Gateway, LE19BH, Leicester, United 

Kingdom, acarreras@dmu.ac.uk, Parmjit Kaur 

This paper examines the idea that because the techniques included under the 

umbrella heading of Problems Structuring Methods might best be considered 

as crafts, then the most effective way for a student to understand and be able 

to apply such techniques is for them to experience these techniques in a real 

workshop as a participant. It will draw upon empirical data gathered from a 

series of student membership workshops carried out on behalf of the OR society 

at three UK universities, and the qualitative data generated in these Causal 

Mapping workshops. 

� MA-36 


3.1.05 

Fuzzy expert systems 

Stream: Fuzzy Systems, Neural Networks & Artificial 

Intelligence 


Chair: Heinrich Rommelfanger, Economics and Business 

Administration, Goethe University, Niebergallweg 16, 65824, 

Schwalbach a. Ts., Hessen, Germany, 

Rommelfanger@wiwi.uni-frankfurt.de 

1 - Employer Awards: Fuzzy Ratings and Rankings 

Jessica Naundorf, Department of Economics & Management, 

Otto-von-Guericke-University Magdeburg, Chair of Strategic 

Management and Organisation, P.O. Box 41 20, 39 016, 

Magdeburg, Germany, jessica.naundorf@ovgu.de, Thomas 

Spengler 

The participation in employer awards has evolved into a major topic so as to 

be the candidate’s first choice in increasing "War for Talents’. Given this managerial 

interest, the present article discusses the aims of employer awards from 

the employer’s viewpoint in addition to the instruments, which need to be used 

within the rating and ranking process regarding the "100 Best Companies to 

Work For’. Due to vagueness, the current ratings and rankings are not applicable 

to mitigate the valuations of experts. Hence, fuzzy logic seems particularly 

suitable to model this decision-making process. The corresponding procedures 

are based on fuzzy sets as well as on a fuzzy rule-based expert system. 

2 - Dynamics of Diffuse Information Processing: The Example 

of Stock Price Movements 

Andreas Uphaus, Faculty of Economics and Management, 

Otto-von-Guericke University Magdeburg, Universitaetsplatz 2, 


Chair of Empirical Economics Research, 39106, Magdeburg, 

Sachsen-Anhalt, Germany, auphaus@online.de, Bodo Vogt 

We explore the idea that stock prices are not known exactly. Traders normally 

know a range of possible prices which we call range of reasonable alternatives. 

In former studies we have shown how to model the process for one point in 

time. This procedure is closely related to fuzzy intervals. Now we analyze sequences 

of stock prices with respect to the question how diffuse information is 

dealt with over several points in time. We present a model which is based on an 

iterated application of the Numerical Response Process over time. The model 

predicts price movements for short time intervals well. 

3 - Operational Risk Assessment: A Fuzzy Logic Approach 

Sebastian Hain, Economic and Business Administration, Goethe 

University Frankfurt, Seckbacher Landstr. 30, 60389, Frankfurt 

am Main, Hessen, Germany, shain@stud.uni-frankfurt.de, 

Heinrich Rommelfanger 

This paper presents an integrated approach to evaluating operational risk based 

on a hierarchical system of risk factors. Employing a fuzzy logic expert system 

both quantitative and qualitative data can be aggregated to the total operational 

risk. The procedure is explained by the subsystem IT security. For selected risk 

categories the expert rule maps and the fuzzy inference process are described 

in detail. A numerical example illustrates the course of the fuzzy expert system 

using the new developed software tool Visual Fuzzy. 

4 - Optimization of Fuzzy Models Obtained From Typicality 

and Membership Partitions 

Rui Jorge Almeida, Department of Econometrics, Erasmus 

University Rotterdam, P.O. Box 1738, 3000 DR, Rotterdam, 

Netherlands, rjalmeida@ese.eur.nl, Uzay Kaymak 

We study how to simplify, combine and optimize the fuzzy rule-based model 

derived using either Fuzzy Possibilistic C-Means or Possibilistic Fuzzy C- 

Means algorithms, where multiple rules per cluster are extracted from both 

the membership partition matrix and the typicality matrix. We make use of 

a three-step approach to obtain fuzzy models. An initial rule-based model is 

obtained by product space clustering. Then we perform simplification of the 

fuzzy sets in the rule base and rule reduction. Finally the parameters of the 

model are optimized using genetic algorithms or a neuro-fuzzy system. 

� MA-37 


3.1.09 

OR Applications in the Health Field 

Stream: OR for Development and Developing Countries 


Chair: Habib Chabchoub, LOGIQ, Institut Supérieur de Gestion 

Industrielle, 3018, Sfax, Sfax, Tunisia, 

Habib.chabchoub@fsegs.rnu.tn 

Chair: Youssef Masmoudi, Sfax University, Hight School of 

Commerce of Sfax, BP 954, 3018, Sfax, Sfax, Tunisia, 

youssef_m_tn@yahoo.fr 

1 - Predicting cancer risk group of countries based on 

oecd health data 2005 - 2007 

Idil Erte, Medical Informatics, METU, Orta do˘gu Teknik 

Üniversitesi, Rektörlük 6.kat, 06531, Ankara, Turkey, 

idilerte@gmail.com, Elif Cakir, Pinar Koseoglu, özlem özkan 

The study aims to use classification and clustering algorithms for different 

health related problems and compare the results of these. This study includes 

discovering knowledge from OECD Health Data 2007 and 2005 which includes 

data related to many attributes about thirty countries all over the world such as 

life expectancy, tobacco usage, cancer and aids.We tried to generate a model for 

predicting the cancer risk levels of countries by using some attributes of training 

data set.The main knowledge we want to discover is "cancer risk group of 

the countries according to OECD Health Data 2007" 

2 - Breast cancer prognosis based on Bayesian Networks 

Arij Mkaouar, ENIS, 3018, Sfax, Tunisia, 

arij.mkaouar@yahoo.fr, Hekma Louati, Ahmed Rebai 

17


The aim of this study is to use a database of real examples of 84 patients to 

model a system of prognosis of breast cancer. We have adapted for the resolution 

of this issue the Bayesian networks. Our network is formed of a set of 

medical measures corresponding to the nodes of the network. The arcs describe 

the relations between these measures as conditional probabilities. We used the 

implicit approach for the estimation of parameters. When a new case arrives, it 

is inserted into the network and is propagated to all nodes and arcs to obtain a 

final prediction of good quality. 

3 - Emergency Department Performance Measures: Multi 

Class Queuing Networks 

Jihen Jlassi, LOGIQ, 3018, Sfax, Tunisia, 

jihene2000tn@yahoo.fr 

Emergency department becomes a useful way to the access to hospital and it 

is a subject of study for many researchers. The research developed in this paper 

aims to improve the performance of the emergency department (ED) of 

Sfax Hospital by analytical method. So, a network of queues with multiple 

customer types is proposed. Different indicators of performance are used. 

4 - The menu planning problem: a formal study and a practical 

study 

Amine Lamine, LOGIQ, Institut Supérieur de Gestion 

Industrielle de Sfax, 3018, Sfax, Tunisia, 

amine.lamine@yahoo.fr, Mahdi Khemakhem, Habib Chabchoub 

The healthy choice of food can help people to reduce the risk of chronic diseases 

like heart disease, diabetes, etc. We study in this work the menu planning 

problem MPP which determines the appropriate composition of meal’s menus. 

Theoretically, we present this problem as a new NP-hard optimization problem 

called the MD-GUBMKP which is an extension of variants of the knapsack 

problem KP. We propose a formulation for the MD-GUBMKP and its direct 

application to the MPP. Next, we show how we can, formally and practically, 

transform the MD-GUBMKP to the others variants of KP and inversely. 

� MA-38 


6.2.44 

Experimental Economics and Game Theory 

Stream: Experimental Economics and Game Theory 


Chair: Ulrike Leopold-Wildburger, Statistics and Operations 

Research, Karl-Franzens-University, Universitätsstraße 15/E3, 8010, 

Graz, Austria, ulrike.leopold@uni-graz.at 




1 - Simulation Results of Nash-Cournot Equilibrium 

Amnon Gonen, Management of Technology, Holon Institute of 

Technology - HIT, 52,Golomb St., 58102, Holon, Israel, 

agonen1@gmail.com 

In this study, duopoly market equilibrium is tested by a group of players using 

a special business simulator that was adopted specifically for this task. The 

results showed that the players’ decisions were mainly influenced by the difference 

in quantities and price, rather than the actual prices and quantities, as 

was expected. Moreover, if the prices and quantities of both firms were quite 

similar, they believed they had achieved a state of equilibrium, although they 

were sometimes quite far from it. 

2 - Analysis of Serbian development indicators and calculation 

of the competitiveness index 

Aleksandra Marcikic, University of Novi Sad, Faculty of 

Economics Subotica, Mihaila Pupina 32, 24000, Subotica, 

Serbia, amarcikic@ef.uns.ac.rs, Boris Radovanov 

Economics changes in most of transition countries, such as Serbia, often lead to 

inequality in income distribution, high inflation rate and reduction of real gross 

domestic product. The paper presents actual position of Serbia by quantitative 

analysis of basic development indicators, which is followed by comparative 

analysis of development factors in surrounding countries. Furthermore, factor 

analysis is used to create the index of competitiveness, as an integral indicator 

that represents influences of ten variables, and finally the position of Serbia is 

compared with other European countries. 

18 

3 - Global division problems with preferences 

Miguel A. Hinojosa, Universidad Pablo de Olavide, Cta. Utrera 

s/n, 41013, Seville, Spain, mahinram@upo.es, Amparo Mármol, 

Francisca Sanchez 

We address a multi-dimensional extension of classical division problems to situations 

in which several goods have to be divided simultaneously among a set 

of agents, and both references points for the division of each good, and preferences 

of the agents on the shares they obtain of the various goods, have to be 

taken into account. 

We analyse the implications of the agents exhibiting additive preferences and 

also address the case in which the preferences of the agents are maximin and 

leximin. 

4 - A Simulation Study of an Iterated Prisoners’ Dilemma 

Experiment 

Ulrike Leopold-Wildburger, Statistics and Operations Research, 

Karl-Franzens-University, Universitätsstraße 15/E3, 8010, Graz, 

Austria, ulrike.leopold@uni-graz.at, Thomas Burkhardt 

We study the development of cooperation in a repeated prisoner’s dilemma experiment 

with unknown length and unknown continuation probability. Players 

are rematched with a new team twice. We found that players use prior experience 

to anticipate the end of a matching. There is a dramatic restart effect. 

Stability of mutual cooperation crucially depends on the initial move of both 

players. In the simulation study we raise the question whether all the effects 

can be shown by simulated situations. 

� MA-39 


6.2.45 

Optimal Control in Economics and 

Economic Demography 

Stream: Optimal Control 


Chair: Raouf Boucekkine, Université catholique de Louvain, B-1348, 

Louvain-La-Neuve, Belgium, Raouf.Boucekkine@uclouvain.be 

1 - Forest Management, Carbon Sequestration and Biodiversity 

in the Presence of Climate Change 

Renan Goetz, Department of Economics, University of Girona, 

Campus Montilivi, s/n, 17071, Girona, Spain, 

renan.goetz@udg.edu, Natali Hritonenko, Rubén Javier Mur 

Torrentó, Yuri Yatsenko, Àngels Xabadia 

This paper determines the optimal forest management for timber and carbon 

sequestration in the presence of climatic changes. It does not only focus on 

mitigation costs for carbon sequestration but also includes the cost and benefits 

of adaption of the management regime to climate change. Furthermore, the paper 

compares the flow method and the rental value approach for the valuation 

of carbon sequestration. It shows that they are only equivalent under very specific 

conditions. Finally, the paper determines the optimal forest management 

regime if biodiversity is also considered as an objective. 

2 - Optimal Population Problem and endogenous Growth: 

Re-assessing the Parfits repugnant Conclusion 

Giorgio Fabbri, Department of Economic Studies, University of 

Naples "Parthenope", Via Medina 40, 80133, Napoli, Italy, 

giorgio.fabbri@uniparthenope.it, Raouf Boucekkine, David de la 

Croix 

Recently some authors have tried to evaluate Parfit’s repugnant conclusion 

within a neoclassical growth model (see Arrow et al., 2009). Here we consider 

an AK model with a broad concept of capital. A planner maximizes a social 

welfare function depending on consumption per capita and population size. In 

this framework, we characterize optimal population dynamics and optimal consumption 

per capita paths. Parfits repugnant conclusion is shown to arise under 

some precise conditions on the preference and technological parameters of the 

model. 

3 - Spatial dynamics and convergence: The spatial AK 

model 

Carmen Camacho, Economics Department, Université catholique 

de Louvain, Place Montesquieu 3, 1348, Louvain la Neuve,

Belgium, carmen.camacho@uclouvain.be, Raouf Boucekkine, 

Giorgio Fabbri 

We study the optimal dynamics of an AK economy where population is uniformly 

distributed along the unit circle. Locations only differ in initial capital 

endowments. Despite constant returns to capital, we prove that transition dynamics 

will set in. In particular, we prove that the spatio-temporal dynamics, 

induced by the willingness of the planner to give the same (detrended) consumption 

over space and time, lead to convergence in the level of capital across 

locations in the long-run. 

4 - Happiness due to Consumption and its Increases, 

Wealth and Status 

Andreas Novak, Business Administration, Bruennerstrasse 72, 

A-1210, Vienna, andreas.novak@univie.ac.at, Franz Wirl 

This paper departs from the standard open-economy Ramsey model and introduces 

additional concerns for wealth, status and Easterlin’s hypothesis that 

consumption changes, in particular increases, are important and not only the 

level. These extensions induce first of all interior steady states, multiple steady 

states and limit cycles. Surprisingly, introducing status conferred by private 

wealth or conspicuous consumption has no effect despite the involved externalities 

as long as the social influence associated with the status externalities 

remains moderate. 

� MA-40 


6.2.52 

Wireless and sensor networks 

Stream: Network Optimization 


Chair: Bernard Fortz, Département d’Informatique, Université Libre 

de Bruxelles, CP 210/01, Bld du Triomphe, 1050, Bruxelles, 

Belgium, bfortz@euro-online.org 

1 - Stochastic Second Order Cone Programming in Mobile 

ad hoc networks 

Francesca Maggioni, Dept. of Mathematics, Statistics, Computer 

Science and Applications, University of Bergamo, Via dei 

Caniana n. 2, 24127, Bergamo, Italy, Italy, 

francesca.maggioni@unibg.it, Florian Potra, Marida Bertocchi, 

Elisabetta Allevi 

We propose a two-stage stochastic second-order cone formulation of a stochastic 

location-aided routing (SLAR) model for mobile ad-hoc network. The aim 

is to provide a sender node S, with an algorithm for optimally determining a 

region that contain a destination node D. The movements of D are given by random 

ellipsoid scenarios. A stochastic second order cone model allows to solve 

problems wich a much larger number of scenarios (20250) than is possible with 

semidefinite models (500). Sensitivity analysis to reveal the stochasticity is included. 

2 - Exact models for the k-connected wireless survivable 

network problem 

Christina Burt, Mathematics and Statistics, University of 

Melbourne, 139 Barry St, 3053, Carlton, Victoria, Australia, 

c.burt@ms.unimelb.edu.au, Yao-ban Chan 

Survivable wired networks have been extensively studied up to 2-connected. 

More recently, 1-connected wireless networks have been addressed, but kconnected 

networks remain an open problem. 

We prove that, in contrast to the wired case, the minimum transmission kconnected 

optimisation problem is NP-hard even for k = 1. 

We construct a multi-commodity flow model, making use of preprocessing and 

valid inequalities from the literature; and develop a new warm start heuristic to 

improve computation time. 

3 - Sensor Network Lifetime Maximization with Mobile Sink 

Routing 

M. Emre Keskin, Industrial Engineering, Bogazici University, 

Bogazici Universitesi Endustri Muhendisligi, Hisarustu / 

Besiktas, 34470, Istanbul, Turkey, m.emre.keskin@gmail.com, I. 

Kuban Altinel, Necati Aras, Cem Ersoy 


It is a known phenomenon of the WSNs that the sensors located near to the 

sink deplete their energies faster than the rest of the network. This causes the 

sink to be disconnected from the rest of the network while most of the sensors 

are still functional. To remedy this weakness, moving the sink toward different 

regions of the network area can be offered. In this study, we attempt to find an 

optimal path for the sink with and without optimal data flow decisions in order 

to maximize the network lifetime. We also take sink travel time and the data 

accumulated during that time into consideration. 

� MA-41 


3.1.06 

Recent Advances in Revenue Management 

Stream: Revenue Management 


Chair: Joern Meissner, Management Science, Lancaster University 

Management School, Room A48, LA14YX, Lancaster, United 

Kingdom, j.meissner@lancaster.ac.uk 

1 - Decomposition Techniques in Large-Scale Network 

Revenue Optimization 

Arne Karsten Strauss, Management Science, Lancaster 

University Management School, Lancaster, LA1 4YX, Lancaster, 

United Kingdom, a.strauss@lancaster.ac.uk 

Many service providers need to optimize the availability of products over a 

large network of resources. A heuristic solution approach is to somehow decompose 

the network optimization problem into a collection of small problems 

corresponding to the individual resources. However, most recent approaches 

are tested only on very small problems (


In many implemented network revenue management systems, a bid price control 

is being used. Yet it is still unknown how well a bid price control actually 

performs. We propose a simple and fast heuristic that iteratively improves on 

an initial guess for the bid price vector. We demonstrate that using these dynamic 

marginal capacity values directly as bid prices can lead to significant 

revenue loss as compared to using our heuristic. Finally, we investigate numerically 

how much revenue performance is lost due to the confinement of product 

combinations that can be represented by a bid price. 

� MA-42 


3.1.07 

Bilevel Programming 

Stream: Variational Inequalities, Complementarity 

Problems and Bilevel Programming 


Chair: Stephan Dempe, Mathematics and Computer Sciences, 

Technische Universitaet Freiberg, 09596, Freiberg, Germany, 

dempe@math.tu-freiberg.de 

1 - An improved exact method for the discrete (r, p)centroid 

problem 

Yury Kochetov, Information Technology Department, 

Novosibirsk State University, Pirogova str., 2, 630090, 

Novosibirsk, Russian Federation, jkochet@math.nsc.ru, 

Ekaterina Alekseeva, Nina Kochetova, Alexander Plyasunov 

In the discrete (r, p)-centroid problem two decision makers, a leader and a follower, 

compete to attract clients from a given market. The decision makers try 

to maximize their own profits. So, we deal with a noncooperative Stackelberg 

game. We present the game as a mixed integer linear program with an large 

number of constraints and variables. An exact iterative method based on the 

column generation technique is proposed. A feasibility problem is solved in 

each iteration by metaheuristics. Computational results for benchmarks from 

the library "Discrete Location Problems’ are discussed. 

2 - Interval uncertainty in the context of bilevel fractional 

programs 

Carmen Galé, Métodos Estadísticos, Universidad de Zaragoza, 

CPS, Edificio Torres Quevedo, María de Luna,3, 50015, 

Zaragoza, Spain, cgale@unizar.es, Herminia I. Calvete 

This work focuses on linear fractional bilevel programming problems in which 

the coefficients are only approximately known. In particular, the proposed approach 

considers that the coefficients of the problem are specified as intervals. 

We analyze methods that find the best optimum and the worst optimum, as well 

as the coefficient settings which achieve these two extremes. The range of the 

upper level objective function between the best and the worst optima gives an 

insight into the possible results and the risk involved in the decision process. 

3 - Resolution of bilevel problems using Inexact Restoration 

Ana Friedlander, Applied Mathematics, Institute of 

Mathematics, Statistics and Scientific Computing, Rua Sergio 

Buarque de Holanda 651, Cidade Universitária, 13083859, 

Campinas, São Paulo, Brazil, friedlan@ime.unicamp.br 

Inexact restoration methods are used in general nonlinear programming problems. 

Their appeal is greatly due to the freedom of choice of the algorithms 

that can be used to solve the subproblems defined in a typical iteration. Bilevel 

programming problems present characteristics that suggest situations when 

this approach could be promising. In this presentation we discuss this topic 

and present examples. Theory and practice will be commented. 

� MA-43 


8.2.02 

Algorithmic Decision Theory 1 

Stream: Algorithmic Decision Theory 


Chair: Ulrich Junker, ILOG, An IBM Company, 1681, route des 

Dolines, 06560, Valbonne, France, uli.junker@free.fr 

20 

1 - A calibration-based approach to combine experts prior 

information 

María Jesús Rufo Bazaga, Mathematics, University of 

Extremadura, Avda de la Universidad s/n, 10071, Cáceres, Spain, 

mrufo@unex.es, Carlos Javier Pérez Sánchez, Jacinto Martín 

This work proposes a Bayesian approach based on a log-linear pooling to aggregate 

experts’ information. It uses an expected Kullback-Leibler divergence 

to obtain the weights in the combined prior distribution. Specifically, the combined 

prior distribution that minimizes the expectation with respect to the corresponding 

predictive prior distribution is found. Therefore, a pooled prior 

distribution is achieved for which the expected calibration is the best one. This 

proposal is based on a criterion that can be used as calibration tool when subjective 

prior distributions are used. This approach can be used in several decision 

making contexts as it is shown with an illustrative application. 

2 - Scaling Invariance and a Characterization of Linear Objective 

Functions 

Sasa Pekec, Fuqua School of Business, Duke University, 1 

Towerview Road, 27708-0120, Durham, NC, United States, 

pekec@duke.edu 

A decision-maker who aims to select the "best" collection of alternatives from 

the finite set of available ones might be severely restricted in the design of the 

selection method. If the representation of valuations of available alternatives is 

available to linear scaling, such as the choice of the unit of measurement, the 

only sensible way to compare choices is to compare weighted sums of individual 

valuations corresponding to these choices. This scaling invariance also 

provides a characterization of linear 0-1 programming objective functions. The 

problem of finding an optimal subset of available data to be aggregated, allowing 

for use of different aggregation methods for different subsets of data, is also 

addressed. 

3 - Invanriants of Decision Problems and impacts onto 

methods 

Alexis Tsoukiàs, CNRS - LAMSADE, Université Paris 

Dauphine, 75775, Paris Cedex 16, France, 

tsoukias@lamsade.dauphine.fr, Alberto Colorni, Wassila 

Ouerdane 

In this presentation we continue the exploration of "what is a decision problem". 

We generalise the concepts of optimisation and constraint satisfaction 

and we then discuss the problem generated by the presence of multidimensional 

information (multiple stakeholders, multiple states of the nature, 

multiple criteria). In the presentation we explore the general problem where 

independence among such dimensions is not set. Among the possible uses of 

such an analysis we focus on the problem of constructing explanations and 

justifications for the end user of the model using argumentation theory. 

4 - Effects of hierarchical weighting in preference programming 

Jyri Mustajoki, Department of Automation Science and 

Engineering, Tampere University of Technology, P.O.Box 692, 

33101, Tampere, Finland, jyri.mustajoki@tut.fi 

Preference programming is an approach to incorporate imprecision in multiattribute 

value trees with interval judgments. We study how the hierarchical 

model structure affects the overall imprecision in the results, and in which level 

of the hierarchy it is most efficient to give more precise judgments. The aim 

is to find out good practices for carrying out the weighting process in practice. 

The simulation experiment provides theoretical results, which are discussed 

with respect to practical issues. The results suggest paying attention to increasing 

precision in the lower level judgments. 

� MA-44 


8.2.03 

Portfolio Decision Analysis I 

Stream: Portfolio Decision Analysis 


Chair: Juuso Liesiö, Systems Analysis Laboratory, Aalto University, 

P.O.Box 11100, 00076Aalto, Espoo, Finland, juuso.liesio@tkk.fi 

Chair: Ahti Salo, Systems Analysis Laboratory, Aalto University 

School of Science and Technology, P.O. Box 11100, Otakaari 1 M, 

00076, Aalto, Finland, ahti.salo@tkk.fi

1 - A PROMETHEE.based approach to portfolio selection 

problems 

Rudolf Vetschera, Dept. of Business Administration, University 

of Vienna, Bruenner Str. 72, A-1210, Vienna, Austria, 

rudolf.vetschera@univie.ac.at, Adiel Teixeira de Almeida 

Multi-attribute decision methods like PROMETHEE cannot be directly applied 

to portfolio selection problems because the require pairwise comparisons of alternatives 

and the number of potential portfolios is usually huge. We introduce 

a simplified approach to portfolio selection based on PROMETHEE net flows 

of individual items rather than portfolios. We develop an equivalence theorem 

for mathematical programming formulations of the problem and present results 

of a computational study comparing our approach to a PROMETHEE ranking 

of portfolios, which shows very promising results. 

2 - Optimal revision of uncertain estimates in project portfolio 

selection 

Eeva Vilkkumaa, Department of mathematics and systems 

analysis, Aalto University, School of science and technology, 

Systems analysis laboratory, Aalto University, P.O.Box 11100, 

00076 Aalt, Espoo, Finland, eeva.vilkkumaa@tkk.fi, Juuso 

Liesiö, Ahti Salo 

Organizations typically select projects based on uncertain point estimates about 

the projects’ future values. In this paper, we analyze how the explicit consideration 

of uncertainties in these estimates helps improve decision quality. Specifically, 

we develop a Bayesian adjustment approach where the sum of revised 

estimates for the selected projects will, on average, coincide with that of the 

projects’ actual values. Decisions based on adjusted estimates are also shown 

to yield a higher expected portfolio value and a greater number of correctly 

selected projects. 

3 - Measurable Multiattribute Value Functions for Project 

Portfolio Selection and Resource Allocation 

Juuso Liesiö, Systems Analysis Laboratory, Aalto University, 

P.O.Box 11100, 00076Aalto, Espoo, Finland, juuso.liesio@tkk.fi 

The linear-additive value model (i.e. portfolio value is the sum of additive 

multi-attribute project values) is widely employed to support multiobjective 

project portfolio selection. Motivated by empirical evidence, we relax the assumption 

of mutually preference independent projects underlying the additivelinear 

model and derive corresponding multilinear value models. We show that 

the elicitation of a multilinear value function can be decomposed into scoring 

each project with regard to each attribute and develop optimization models to 

maximize portfolio value subject to resources constraints. 

4 - Comparison of two approaches for evaluating a portfolio 

of R&D projects with a budget 

Anabela Costa, Quantitative Methods, ISCTE - Instituto 

Universitário de Lisboa, Av. das Forças Armadas, 1649-026 , 

Lisboa, Portugal, anabela.costa@iscte.pt, José Paixão 

The Real Options approach has proved to be a suitable methodology for the 

financial evaluation of R&D projects since it captures the value of flexibility in 

R&D projects. Based on a dynamic programming model, presented in the literature, 

we present two approaches to valuing a portfolio of R&D projects with a 

budget: one approach makes an extension of the model above mentioned, and 

the other one, estimates the value of the projects by simulation. Computational 

experiments are discussed. 

� MA-45 


8.2.12 

Dynamic Programming I 

Stream: Dynamic Programming 


Chair: Lidija Zadnik Stirn, Biotechnical Faculty, University of 

Ljubljana, Vecna pot 83, 1000, Ljubljana, Slovenia, Slovenia, 

lidija.zadnik@bf.uni-lj.si 

Chair: Moshe Sniedovich, Dept. of Mathematics and Statistics, 

University of Melbourne, Parkville, 3010, Melbourne, Victoria, 

Australia, m.sniedovich@ms.unimelb.edu.au 


1 - Operational optimization as a part of a bi-level optimization 

problem 

Mikko Linnala, Department of Physics and Mathematics, 

University of Eastern Finland, Yliopistonranta 1 F, 70210, 

Kuopio, Finland, mikko.linnala@uef.fi 

A bi-level optimization problem covering design and operational levels is defined. 

The problem is solved using dynamic multiobjective optimization, i.e. 

values of state and control variables change over a time horizon, and competing 

criteria are optimized simultaneously. Moreover, interaction between the 

design and operational levels is analyzed. Benefits of bi-level dynamic multiobjective 

optimization are illustrated by giving an operational optimization 

example of a paper mill in which process design is optimized simultaneously 

in order to maintain efficiency of the mill. 

2 - Dynamic Programming Approach for Arrivals Scheduling 

at Heathrow Airport 

Stanislava Armstrong, School of Computer Science & IT, The 

University of nottingham, Jubilee campus, Wollaton rd, NG8 

1BB, Nottingham, Nottinghamshire, United Kingdom, 

saw@cs.nott.ac.uk, Jason Atkin, Edmund Burke 

Busy airports experience periods when demand exceeds runway capacity, 

which leads to an accumulation of aircraft circling in nearby stacks. This could 

cause delays that can adversely affect ground resource utilisation and may have 

implications for connecting flights. An accurate estimation of landing times 

could contribute to better resource allocation and time management. The arrival 

stacks can limit the possible landing sequences. Their effects upon the 

sequences and on the delays will be shown and potential benefits from operating 

mode changes will be detailed. 

3 - Dynamic and group decision-making model for optimal 

management of natural resources 

Lidija Zadnik Stirn, Biotechnical Faculty, University of 



Sustainable management of natural resources requires not only satisfying specified 

sustainability criteria, but also the perception of these criteria by the various 

segments of society, and involves a long-term perspective. We generated 

a dynamic, multi-criteria model involving several decision-makers to address 

these issues. Specifically, our model is based on a network which presents a 

Bellman multi-stage iterative decision process, and a DEA and weighted geometric 

mean method modified ANP for deriving a group priority at each stage 

of the process. An application of the model is presented. 

4 - Lot-sizing model with dynamic safety inventories 

Jinhua Zheng, Maritime Technology and Logistics, Tokyo 

University of Marine Science and Technology, Tokyo University 

of Marine Science and Technology, 

4-5-7,Konan,Minato-ku,Tokyo,Japan, 108-8477, Tokyo, Japan, 

jinhua@logopt.com, Mikio Kubo 

In this paper we introduce an extended model of the classical lot-sizing model 

in which demand parameters are stochastic. One characteristic of our model 

is to determine lot-sizes and safety stocks simultaneously. We develop a new 

dynamic programming algorithm for the single item uncapacitated model, and 

prove an extended planning horizon theorem that accelerates the dynamic programming 

algorithm. Numerical experiments suggest the efficacy of the proposed 

algorithm and the extended planning horizon theorem. 

� MA-46 


8.2.14 

OR Challenges Related to the Recent and 

Future Disasters I 

Stream: OR for Madeira (and related challenges) 





21


1 - A model for high-speed railway planning optimization 

considering risk analysis and management 

Ana Costa, Civil Engineering, University of Coimbra, Rua Luís 

Reis Santos, Polo II da UC, 3030 - 788, Coimbra, Coimbra, 

Portugal, alcosta@dec.uc.pt, Maria Cunha, Paulo Coelho 

Planning of High-Speed Railways must consider multiple and uncertain future 

conditions under which to perform (e.g. floodings, earthquakes). Options 

in corridors and technical solutions to adopt exist and different solutions may 

yield different overall performance. This paper discusses the main issues to 

consider in developing a systematic tool to support the decision process with 

risk analysis and management through scenario descriptions of uncertainty. A 

methodology for alignment optimization using Simulated Annealing is proposed 

and an overview of obtainable results is presented. 

2 - Infrastructures in Natural Disasters 

Huey-Kuo Chen, Civil Engineering, National Central University, 

No. 300, Jung-Da Road, 320, Chung-Li City, Taoyuan, Taiwan, 

ncutone@ncu.edu.tw 

When a natural disaster occurs, the transportation infrastructures in the affected 

area are often seriously damaged. The fatalities and property losses will get 

worsened if the damaged transportation infrastructures cannot be repaired in a 

reasonabe time. We address the task of repairing damaged infrastructure as the 

real-time vehicle routing problem with time windows that is repetitively solved 

in a time rolling frame. The results show that the proposed solution metaheuristic 

is comparable to Wang et al. (2008) but the idea of our heuristic is more to 

our intuition. 

3 - Emergency Transportation Decisions for the Post- 

Disaster Istanbul Earthquake Case 

Emin Ahmet Tanrioven, Industrial Engineering, Graduate School 

of Science and Engineering, Koc University, Rumeli Feneri Yolu, 

34450, Sariyer, Istanbul, Turkey, etanrioven@ku.edu.tr, Sibel 

Salman, Esma Gel 

This study develops a decision support tool for dispatching ambulances and 

transporting casualties to hospitals or temporary emergency units in a masscasualty 

incident. Specifically, post-earthquake case for six crowded districts 

in Istanbul is studied. The ambulance dispatching problem is inherently large 

scale in the context of disaster response, with a surge in transportation needs 

that changes over time in the post-disaster time frame. Dispatching strategies 

are studied while considering injury types of patients and fairness among six 

districts in a dynamic and stochastic environment. 

4 - Emergency medical systems: a stochastic generalized 

assignment case study 

Susana Baptista, CMA, FCT - UNL, Campus de Caparica, 

2829-516 , Caparica, Portugal, sbb@fct.unl.pt 

The assignment problem in an emergency ambulance system can be formulated 

as a stochastic on-line generalized assignment problem: an ambulance 

must be assigned to a service request as soon as its location is known, service 

time and assignment costs are stochastic, and workshift final time should not 

be exceeded. In order to evaluate solutions found by alternative assignment 

rules of ambulaces belonging to the National Institute for Medical Emergency 

and serving the city of Lisbon, off-line solutions are obtained using a scenario 

analysis approach and applying a stochastic recourse model. 

22 

� MA-47 


8.2.16 

Various Aspects of Sustainable Living in 

Developing Countries 

Stream: Sustainable Living: Cognitive, Social, 

Economical, Ecological and World View 


Chair: Pedamallu Chandra Sekhar, Department of Medical Oncology, 

Dana-Farber Cancer Institute, Boston, 02115, Ipswich, MA, United 

States, pcs.murali@gmail.com 

Chair: Ali Gökmen, Department of Chemistry, Middle East Technical 

University, 06531, Ankara, Turkey, agokmen@metu.edu.tr 

Chair: Inci Gokmen, Chemistry, Middle East Technical University, 

METU Department of Chemistry, 06531, Ankara, Turkey, 

igokmen@metu.edu.tr 




Chair: Canan Pehlivan, Industrial Engineering, Hacettepe University, 

Beytepe Campus, 06800, Ankara, Turkey, cananp@hacettepe.edu.tr 

1 - Maximum entropy in land use change - a methodology 

for an urban sustainability model 

Daniel Zachary, CRTE, CRP Henri Tudor, 66, rue de 

Luxembourg, L-4421, Esch sur Alzette, Luxembourg, 

dan.zachary@tudor.lu 

This paper explores the use of informational entropy as a guide to land use 

selection for urban and regional planning. The Maximum Entropy Allocation 

Land Use model (METALU) is constructed around the paradigm that cost can 

be represented in terms of available energy and land use choice as populated 

states. These analogies have allowed for the formulation of a planning model 

to be describe in terms of Boltzmann statistics and state population properties 

have been exploited using the concept of maximum entropy. 

2 - Incentives of no-lose targets for developing countries 

Vicki Duscha, Fraunhofer Institute, Germany, 

v.duscha@isi.fraunhofer.de 

For an international environmental agreement to be effective industrialized as 

well as developing countries like China and India need to agree to significant 

emission reductions. One idea to integrate developing countries are no-lose targets. 

Three factors directly affect the emission reduction incentive of a no-lose 

target: the target setting, the price for emission allowances and the costs for 

emission reductions. The price for emission allowances and the costs for emission 

reductions are highly influenced by future technology development which 

is, however, extremely uncertain. 

3 - A location-routing problem for the municipal solid 

waste management system 

Canan Pehlivan, Industrial Engineering, Hacettepe University, 

Beytepe Campus, 06800, Ankara, Turkey, 

cananp@hacettepe.edu.tr, Sedef Meral, Can Ayanoglu 

This study deals with a municipal solid waste management system in which 

the strategic and tactical decisions are addressed simultaneously. We formulate 

this system as a location-routing problem with two facility layers. Mathematical 

models are presented, an iterative capacitated-k-medoids clustering-based 

heuristics proposed and a sequential clustering-based method is presented. We 

apply the clustering-based algorithm to a real life application. Our algorithm 

developed gives consistent results within acceptable running time.

� MA-48 


8.2.04 

Ill-posed Variational Problems I 

Stream: Ill-posed Variational Problems - Theory, 

Methods and Applications 





1 - Proximal Auxiliary Problem Principle with Logarithmic- 

Quadratic Functions for Solving Variational Inequalities 

Christina Jager, Mathematics, University of Trier, 

Universitätsring 15, 54286, Trier, Germany, 

christina.jager@uni-trier.de 

Variational inequalities with operators that are the sum of a maximal monotone, 

set-valued operator and a single-valued, continuous operator are considered. 

Our solution scheme extends the auxiliary problem principle by using 

a logarithmic-quadratic regularization function that provides an interior point 

effect. Operator approximations and an inexact solving of the auxiliary problems 

are allowed. Convergence is proved under mild assumptions and numerical 

examples with nonsmooth optimization problems are given where a bundle 

technique is used to approximate the set-valued part of the operator. 

2 - The exponent of Hölder calmness for level sets of polynomials 

Jan Heerda, Institut für Mathematik, Humboldt-Universität zu 

Berlin, Unter den Linden 6, 10099, Berlin, Germany, 

janjh@math.hu-berlin.de 

Hölder calmness is a tool to describe stability of contraint set mappings. Hoffman 

error bounds yield proper calmness for systems of affine functions. Using 

the Hörmander-Lojasiewicz inequality Luo/Luo and Luo/Pang showed Hölder 

calmness for systems of polynomials and even analytic functions. But this 

result is based on the Tarski-Seidenberg principle, so one only has existance 

of some exponent for Hölder calmness. Here we show that for one quadratic 

polynomial the exponent is one-half. Also some further relations between polynomial 

degree and exponent of Hölder calmness will be given. 

3 - On the cutting plane property and the Bregman Proximal 

Point Algorithm 

Nils Langenberg, FB IV - Mathematik, Universität Trier, 

Universitätsring, 54286, Trier, Germany, 

langenberg@uni-trier.de 

The well-known Bregman PPA for variational inequalities is studied. Under 

the assumption of the cutting plane property (CPP, e.g. paramonotonicity) a 

constrained ill-posed problem can efficiently be solved by means of well-posed 

unconstrained subproblems. Motivated by the fact that the CPP fails to hold 

in saddle problems and Nash games, we deal with the method’s behaviour for 

problems without the CPP. 

We show that if the generated sequence is convergent, then its limit is a solution 

and give a large class of (nonlinearly constrained) feasible sets for which the 

sequence indeed converges. 

EURO 24 - Lisbon 2010 MB-02 

Monday, 10:40-12h00 

� MB-01 


Aula Magna 




Chair: Gerhard J. Woeginger, Department of Mathematics and 

Computer Science, Eindhoven University of Technology, 5600 MB, 

Eindhoven, Netherlands, gwoegi@win.tue.nl 

1 - The combinatorics of Social Choice 

Noga Alon, Tel Aviv University, Israel, nogaa@post.tau.ac.il 

The early work of Condorcet in the 18th century, and that of Arrow and others 

in the 20th century, revealed the complex and interesting mathematical problems 

that arise in the theory of Social Choice, showing that the simple process 

of voting leads to strikingly counter-intuitive paradoxes. I will describe some of 

these, focusing on several recent intriguing examples whose analysis combine 

combinatorial and probabilistic ideas with techniques from Discrete Optimization 

and the theory of the VC dimension of range spaces. 

� MB-02 


3.2.14 

Advanced Combinatorial Optimization 1 



Chair: Paolo Toth, DEIS, University of Bologna, Viale Risorgimento 

2, 40136, Bologna, Italy, paolo.toth@unibo.it 

1 - On the Geometry of Lift-and-Project 

Egon Balas, GSIA, Carnegie Mellon University, Pittsburgh, PA, 

United States, eb17@andew.cmu.edu 

We give geometric interpretations to several aspects of the lift-and-project procedure 

executed on the LP simplex tableau. They are meant to deepen our understanding 

of the procedure and to explain how and why the procedure works. 

2 - A New Lagrangean Based Branch and Bound Algorithm 

for the 0-1 Knapsack Problem 

Cid de Souza, Institute of Computing, University of Campinas, 

Cidade Universitária, Barão Geraldo, Caixa Postal 6176, 

13083-970, Campínas, São Paulo, Brazil, cid@ic.unicamp.br, 

Alexandre Cunha, Laura Bahiense, Abílio Lucena 

This paper describes a new Branch and Bound algorithm for the 0-1 Knapsack 

Problem (KP). The algorithm is based on the use of a Lagrangean Relax-and- 

Cut procedure that allows exponentially many Fractional Gomory Cuts and 

Extended Cover Inequalities to be candidates to Lagrangean dualization. In 

doing so, the upper bounds thus obtained are stronger than the standard Linear 

Programming relaxation bound for KP. The algorithm is aimed at solving 

instances with coefficients as large as 1015, a class of KP instance for which 

existing solution algorithms might not be directly applicable. 

3 - Local search inequalities 

Paolo Serafini, Dept. of Mathematics and Computer Science, 

University of Udine, Via delle Scienze 206, 33100, Udine, Italy, 

serafini@dimi.uniud.it, Giuseppe Lancia, Franca Rinaldi 

We describe a general method for deriving new inequalities for integer programming 

formulations of combinatorial optimization problems. The inequalities, 

motivated by local search algorithms, are valid for all optimal solutions 

but not necessarily for all feasible solutions. These local search inequalities can 

help in either pruning the search tree at some nodes or in improving the bound 

of the LP relaxations. In particular we apply this idea to the TSP and to the 

max-cut problems. 

23

MB-03 EURO 24 - Lisbon 2010 

4 - An Exact Method for the Multi-Trip Vehicle Routing 

Problem 

Aristide Mingozzi, Department of Mathematics, University of 

Bologna, C.d.L. Scienze dell’Informazione, Via Sacchi, 3, 

47023, Cesena, FC, Italy, mingozzi@csr.unibo.it, Roberto 

Roberti, Paolo Toth 

The Multi-Trip Vehicle Routing Problem (MTVRP) is an extension of the Capacitated 

Vehicle Routing Problem where vehicles are allowed to perform more 

routes during the working period. 

To our knowledge, only heuristic algorithms but none exact methods have been 

proposed for the MTVRP so far. We present an exact method based on two different 

set partitioning formulations of the problem and a sequence of bounding 

procedures. 

The computational results show that benchmark instances with up to 120 customers 

can be consistently solved to optimality within acceptable computing 

times. 

� MB-03 


3.2.15 

VRP 



Chair: Sandra Ulrich Ngueveu, LOSI (Laboratory of Industrial 

Systems Optimization), University of Technology of Troyes, 12, rue 

Marie Curie, 10010, Troyes, France, sandra_ulrich.ngueveu@utt.fr 

1 - A Genetic Algorithm for Capacitated Vehicle Routing 

Problem 

Lai-Soon Lee, Laboratory of Applied and Computational 

Statistics, Universiti Putra Malaysia, Institute for Mathematical 

Research, Universiti Putra Malaysia, 43400, UPM Serdang, 

Selangor, Malaysia, lee@math.upm.edu.my, Habibeh Nazif, 

Hsin-Vonn Seow 

A genetic algorithm is proposed for capacitated vehicle routing problem. The 

proposed algorithm uses an optimised crossover operator designed by a complete 

undirected bipartite graph to find an optimal set of delivery routes satisfying 

the requirements and giving minimal total cost. The algorithm is tested with 

benchmark instances and compared it with other known heuristics in the literature. 

Computational results showed that the proposed algorithm is competitive 

in terms of the quality of the solutions found. 

2 - A novel decoding algorithm to solve Vehicle Routing 

Problem by Particle Swarm Optimization 

Babak Farhang Moghaddam, Industrial Engineering, Islamic 

Azad University -parand branch, no.131- mokhberi st.- satari st.- 

Tehran - Iran, 1447174646, Tehran, Tehran, Iran, Islamic 

Republic Of, farhang@iust.ac.ir, Seyed jafar Sadjadi, Elham 

Farhash 

In this paper, A Particle Swarm Optimization (PSO) approach is applied to 

solve the CVRP. Since PSO solves the problems with continues variables and 

VRP is a mix-integer problem, a decoding method is needed to apply PSO for 

VRP. In this way, designing an effective decoding algorithm can empower the 

PSO. So we develop a novel decoding method for interpreting PSO solutions 

to VRP. Also, some heuristic methods are used as local search to have better 

solution. This fact is demonstrated by comparing PSO with the optimal solution. 

Furthermore, our PSO results are compared with some other methods in 

the literature. 

3 - A GRASP ELS Metaheuristic for the Vehicle Routing 

Problem with Conflicts 

Khaoula Hamdi, ROSAS, UTT, 12 rue Marie Curie, 10010, 

Troyes, France, khaoula.hamdi@utt.fr, Labadi Nacima, Alice 

Yalaoui 

The Vehicle Routing Problem (VRP) is a classical problem for which several 

exact and approximation methods were proposed. In real life applications, such 

as in Hazardous Material transportation, transported items may be incompatible. 

This variant of routing problems represents a new problem for which 

only few algorithms have been developed. In a previous work we introduced 

the mathematical model and proposed some heuristics and an Iterated Local 

Search to resolve the Vehicle Routing Problem with Conflicts (VRPC). In this 

study a hybrid GRASP-ELS is designed to improve the previous results. 

24 

4 - A new GRASPxELS approach to solve Vehicle Routing 

Problems with Backhauls 

Raksmey Phan, LIMOS, Université Blaise Pascal, Complexe 

scientifique des Cézeaux, 63173, Aubière, France, 

phan@poste.isima.fr, Christophe Duhamel, Philippe Lacomme 

The Vehicle Routing Problem with Backhaul is a classical extension of the 

VRP where two types of customers are considered: linehauls and backhauls. 

On each route, linehauls have to be served before the backhauls. Several efficient 

ways to solve this problem have been proposed recently: tabu search, LNS 

and MACS. We propose a hybrid metaheuristic scheme called GRASPxELS. 

Two search spaces are used alternatively: the giant trip space and the VRP tour 

space. The numerical results show that GRASPxELS outperforms both three 

methods in the same experimental context. 

� MB-04 


3.2.13 

Scheduling with metaheuristics (2) 



Chair: André Rossi, Lab-STICC - UMR 3192, Université de 

Bretagne-Sud, Centre de Recherche, BP 92116, 56321, Lorient, 

France, andre.rossi@univ-ubs.fr 

Chair: Marcin Siepak, Institute of Informatics, Wroclaw Univesity of 

Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland, 

marcin.siepak@pwr.wroc.pl 

1 - A scatter search method for uncertain P||Cmax problem 

with interval processing times 

Marcin Siepak, Institute of Informatics, Wroclaw Univesity of 

Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, 

Poland, marcin.siepak@pwr.wroc.pl, Jerzy Jozefczyk 

It is assumed that processing times for P||Cmax problem are not known a priori, 

but they belong to intervals of known bounds. The worst-case absolute regret 

is used to evaluate resulting uncertain problem. The deterministic counterpart 

is NP-hard. Moreover, no approximate algorithm is possible to develop for 

the uncertain case. Therefore, an heuristic solution algorithm, based on scatter 

search, is applied. It is compared with another heuristic algorithm based 

on LPT and middle values of uncertain processing times as well as with B&B 

based exact algorithm which is proposed for small instances. 

2 - A Multi Period Model for Multi-Stage Multi-Product Production 

Planning with Random Yield and Compressible 

Processing Time 

A. Soudi, IE, IUST, Narmak, 1684613114, Tehran, Tehran, Iran, 

Islamic Republic Of, asie_soudi@ind.iust.ac.ir, M.b. 

Aryanezhad, M. Karimi-Nasab 

This paper discuses about imperfect production planning in the case of process 

compressibility for each product type at each production stage in every planning 

period. Each product type should pass a series of sequential production 

stages and due to process imperfectness, after each production stage a 100% 

inspection is carried out to screen defective items from non-defectives. As the 

rework is not allowed at all, defective items are scrapped. On the other hand, in 

the presence of random yield, production plans are unreliable rather than perfect 

yield. Consequently, we encounter with more numbers of shortage times. 

The production manager aims at minimizing total costs of setups, production, 

finished items inventory holding, work-in-process inventory holding, shortages 

(in the term of backorders) for a set of product types over a finite number of 

planning periods under given constraints. The problem is formulated and analyzed 

via mathematical modeling and the model is solved on a set of random 

test problems by particle swarm optimization (PSO). Computational experiences 

report that the PSO and the model together lead to better solutions in a 

reasonable CPU time than other previous approaches. 

3 - Lot Sizing Problem with Production Compressibility for 

Deteriorating Items 

A.h. Ebrahimian, Industrial Engineering, Islamic Azad 

University, no. 61, Mir-Emad st., Beheshti ave„ 1587853539, 

Tehran, Tehran, Iran, Islamic Republic Of,

ebrahimian_amirh@yahoo.com, M.b. Aryanezhad, Seyed 

Mohammad Ghoreyshi, M. Karimi-Nasab, E. Noorollahi 

Real life production planning is involved in considering many complicating 

factors. In this paper a new mathematical analysis is investigated for determining 

optimal production-inventory policies in the case of dealing with multiple 

deteriorating products with dynamic demands. The deterioration of each product 

type occurs at a randomly distributed rate. Production rates for each product 

type are decision variables, such that process compressibility is assumed to be 

existed. In other words, there is a time/ cost trade-off in producing each product 

type, where mathematical function has been used to show the relation between 

time and cost of production. A new particle swarm optimization (PSO) algorithm 

is developed based on the problem’s structure for obtaining near-optimal 

solution of the problem by considering the proved theorems. The proposed approach 

is run on a set of random test problem. The results are then compared 

with Lingo 8.0 for small size instances. Computational experiences show that 

both the model and the PSO algorithm have excellent performance even in the 

worst cases. This is illustrated via finding the best tuning relations for different 

parts of the PSO algorithm by using design of experiments’ techniques. 

4 - A Meta-heuristic Approach for the Design and Scheduling 

of Multipurpose Batch Plants - An exploratory study 

Nelson Chibeles-Martins, Dep. de Matemática, FCT, FCT-UNL, 

Quinta da Torre, 2829-516, CAPARICA, Portugal, 

npm@fct.unl.pt, Tânia Pinto_Varela, Ana Paula Barbósa-Póvoa, 

Augusto Novais 

Multipurpose industrial facilities involve the production of a range of products 

through different recipes, hence optimal facility design must also include production 

scheduling aspects. Research in this area has been focused on the use 

of MILP/MINLP models that might become intractable when applied to real 

problems. A contribution is presented, based on a Simulated Annealing approach, 

to overcome this drawback. Examples are solved and the performance 

of SA and exact algorithms compared, with sensitivity analysis on SA parameters 

undertaken. In all cases, design and scheduling are determined. 

� MB-05 


3.2.16 

Genetic algorithms 



Chair: Alexander Plyasunov, Information Technology Department, 

Novosibirsk State University, Pirogova str. 2, 630090, Novosibirsk, 

Russian Federation, apljas@math.nsc.ru 

Chair: Hwai-En Tseng, Department of Industrial Engineering and 

Management, National Chin-Yi University of Technology, 35, 

Lane215, Section 1, Chung-Shan Road, 411, Taiping City, Taichung 

County, Taiwan, hwai_en@seed.net.tw 

1 - Grouping genetic algorithms for green product modular 

design 

Hwai-En Tseng, Department of Industrial Engineering and 

Management, National Chin-Yi University of Technology, 35, 

Lane215, Section 1, Chung-Shan Road, 411, Taiping City, 

Taichung County, Taiwan, hwai_en@seed.net.tw 

Recently, people all over the world paid much attention to green and environmental 

design. Under the consideration of economy, the green recycle of 

a product depends upon the profit/cost relationship of the discarded product. 

This study attempted to build up appropriate algorithms for product modular 

design in three stages. At Stage 1, the liaison intensity was used to quantify the 

connection relations among parts. At Stage 2, the product to be recycled was 

modularized by the grouping genetic algorithms. At Stage 3, the disassembly 

cost and recycle profit of the modules were evaluated. 

2 - Reliability Optimization of Series-k-out-of-M Systems 

Using a Genetic Algorithm 

Ceki Franko, Mathematics, Izmir University of Economics, 1488 

sokak No:1 Daire 12 Alsancak/˙Izmir, 35520, ˙Izmir, Turkey, 

ceki.franko@ieu.edu.tr, Cemal Murat Özkut, Cihangir Kan 


Series-parallel systems are commonly used in optimal design problems. This 

paper presents a generalization of the series-parallel system. Since it is diffi 

cult to obtain the exact solution for series-k-out-of-M system, a heuristic 

method is developed for solving a multi-objective model in which system reliability 

and costs are considered. A mathematical model is provided by transforming 

the multi-objective model into single-objective model. Also, a modifi 

ed genetic algorithm is proposed in order to solve this model. Finally, the numerical 

example is provided and the result is illustrated. 

3 - Genetic local search algorithm for the mill pricing problem 

Alexander Plyasunov, Information Technology Department, 

Novosibirsk State University, Pirogova str. 2, 630090, 

Novosibirsk, Russian Federation, apljas@math.nsc.ru 

In the mill pricing problem we are given two finite sets of facilities and customers. 

Each customer has a budget, a demand and patronizes the facility 

providing the lowest price and travel cost. The objective is to maximaze the 

overall profit for the facilities. We present the problem as a Stackelberg-type 

leader-follower game and show that it is NP-hard in the strong sense. Genetic 

local search algorithm is developed for finding near optimal solutions. Computational 

results are discussed. 

4 - A Preference Based Genetic Algorithm For Bi-Objective 

Capacitated Facility Location Problem With Partial Coverage 

Berk Orbay, Industrial Engineering, Middle East Technical 

University, Turkey, berkorbay@gmail.com, Esra Karasakal 

In this study, we address the problem of locating minimum number of capacitated 

facilities in order to maximize total coverage of the demand points. We 

assume that a facility fully covers the demand points that are located within 

its critical distance and after that distance coverage level of the facility decays 

linearly. We develop a preference based interactive multi-objective genetic algorithm 

to solve the problem. We test the performance of the algorithm on 

randomly generated problems of different sizes. 

� MB-06 


8.2.30 

DEA Methodology II 



Chair: Yannis Smirlis, University of Piraeus, 80 Karaoli - Dimitriou 

Str, 18534, Piraeus, Greece, smirlis@unipi.gr 

1 - Piecewise linear virtual inputs/outputs in Interval DEA 

Yannis Smirlis, University of Piraeus, 80 Karaoli - Dimitriou Str, 

18534, Piraeus, Greece, smirlis@unipi.gr, Dimitris Despotis 

Recent publications relax the DEA assumption of single weights for inputs/outputs 

and deal with cases that virtual inputs/outputs exhibits diminishing/increasing 

marginal functions between specified threshold values. This 

concept has been studied only for crisp data sets. In this work we extend this 

concept of diminishing/increasing returns to Interval DEA. We model nonlinear 

value functions in a piece-wise linear fashion and we transform the original 

data set to an augmented data set, compliant with the Interval DEA methodology. 

Numerical and empirical data sets illustrate this approach. 

2 - Ranking DMUs by Using Stochastic MAJ Model 

Mohammad Hassan Behzadi, Statistics, Science and Research 

Branch, Islamic Azad University, Tehran, Iran, Tehran, Iran, 

Islamic Republic Of, behzadi@srbiau.ac.ir, Farhad Hosseinzadeh 

Lotfi, Mahnaz Mirbolooki 

Using traditional DEA models, many decision making units (DMUs) are classified 

as efficient. These models works with deterministic data. Considering 

stochastic data in DEA is one of the important methods to deal with imprecise 

data. Therefore, in this research studies we provide an additive model for 

ranking efficient DMUs with stochastic data condition. To solve the stochastic 

model, a deterministic equivalent is obtained. Although the deterministic 

equivalent is non-linear, it can be converted to a quadratic program. Using 

numerical example, we will demonstrate how to use the result. 

25


3 - A DEA based approach for solving the multiple objective 

shortest path problem 

Alireza Davoodi, Mathematics, Islamic Azad University, 

Neyshabur Branch, ———, Neyshabur, Khorasan, Iran, Islamic 

Republic Of, alirzd@yahoo.com 

Finding the shortest path in a network is one of the important and interesting 

subjects in network flow problems. When each arc has just one type of cost, 

there exist some simple methods to find the shortest path. But if there are more 

than one type of cost(vector of cost), the non-dominated path plays the role 

of the best path. In this case a Multiple Objective problem is created to find 

the non-dominated path. In this paper a DEA based approach is introduced to 

find the non-dominated path(s) in a multiple cost network. This method can 

determine all efficient paths and the best one. 

4 - Fractional regression models for second stage DEA efficiency 

analyses 

Esmeralda Ramalho, Economics, Universidade de Evora, Largo 

dos Colegiais, 2, 7000-803 , EVORA, Portugal, ela@uevora.pt, 

Joaquim Ramalho, Pedro Henriques 

Data envelopment analysis (DEA) is commonly used to measure the relative efficiency 

of decision-making units. Often, in a second stage, a regression model 

is estimated to relate DEA efficiency scores to exogenous factors. In this paper, 

we argue that the traditional linear or tobit approaches to second-stage DEA 

analysis do not constitute a reasonable data-generating process for DEA scores. 

Under the assumption that DEA scores can be treated as descriptive measures 

of the relative performance of units in the sample, we show that using fractional 

regression models are the most natural way of modeling bounded, proportional 

response variables such as DEA scores. We also propose generalizations of 

these models and, given that DEA scores take frequently the value of unity, 

examine the use of two-part models in this framework. Several tests suitable 

for assessing the specification of each alternative model are also discussed. 

� MB-07 


8.2.47 

New Achievement in Mathematical 

Programming 






Chair: Domingos Cardoso, Departamento de Matematica, 

Universidade de Aveiro, Campus Universitario de Santiago, 

3810-193, Aveiro, Portugal, dcardoso@ua.pt 

1 - Algorithmic strategies for the recognition of graphs 

with convex quadratic stability number 

Maria F Pacheco, Instituto Politécnico de Bragança - ESTiG, 

Quinta de Santa Apolónia, Gab. 112, 5301-857, Bragança, 

Portugal, pacheco@ipb.pt, Domingos Cardoso, Carlos J. Luz 

A major difficulty in the recognition of graphs with convex quadratic stability 

number is the existence of adverse subgraph(s) (subgraph such that the smallest 

eigenvalue of its adjacency matrix doesn’t change when a (neighbourhood of) 

any vertex is deleted). It is a challenge to find adverse graphs without convex 

quadratic stability number. We present the main results about graphs with convex 

quadratic stability number and conclusions about the existence of adverse 

subgraphs belonging to this family in certain classes of graphs. 

2 - A new hybrid cryptosystem based on the satisfiability 

problem 

Sadek Bouroubi, Faculty of Mathematics, Dept of Operations 

research, USTHB University, Laboratory LAID3, BP32 Bab 

Ezzouar 16111, 16111, Algiers, Algeria, bouroubis@yahoo.fr 

With the development of the mathematical methods which ensure safe electronic 

communication, more sophisticated techniques emerged which allow to 

attack codes on increasingly powerful computers. Modern cryptosystems are 

based on number theory, commutative group theory, or algebraic geometry. 

This paper presents a simple hybrid cryptosystem whose security is based upon 

the satisfiability problem well known NP-complete. Its execution is finally illustrated 

by an example to understand better how it processes. 

26 

3 - A two-phase heuristic for the K Clusters with Fixed Cardinality 

Problem 

Lídia Lourenço, Dept. de Matemática, Centro de Investigação 

Operacional, FCT - Universidade Nova de Lisboa, Monte de 

Caparica, 2829-516 , Caparica, Portugal, lll@fct.unl.pt, 

Margarida Pato, Graça Gonçalves 

Given an undirected graph, the K Clusters with Fixed Cardinality Problem consists 

of finding K subsets of nodes with fixed cardinality maximising the total 

similarity of nodes in the same cluster. We propose a two-phase heuristic procedure 

designed to obtain a feasible solution for this NP-hard problem. First 

a greedy rule is used to build a solution which is improved in the next phase 

by an exchange-based heuristic. A computational experiment is designed to 

evaluated the performance of this heuristic. 

4 - A combinatorial approach to assess the separability of 

clusters 

M. Joao Martins, Dep. Matematica, Inst. Superior Agronomia, 

Tapada da Ajuda, 1349-017, Lisboa, Portugal, 

mjmartins@isa.utl.pt, J. Orestes Cerdeira, Pedro C. Silva 

Given a set of entities, to what extent a particular subset is separated from the 

other entities? This is a common question that arises in different relevant areas. 

We propose to assess the separability of a set of entities X based on the following 

notion of intrusion associated to some aggregation criterion. An entity 

y is a k-intruder if every (k+1)-partition where y is singleton is not an optimal 

(k+1)-partition of X U y. We study the optimization problems resulting from 

different aggregation criteria, and develop this approach further to evaluate the 

overall separability of a partition. 

� MB-08 


6.1.36 

Scheduling at Container Terminals 



Chair: Florian Jaehn, Business Administration, Management 

Information Science, Hoelderlinstrasse 3, 57068, Siegen, Germany, 

florian.jaehn@uni-siegen.de 

1 - Positioning freight trains in rail container terminals 

Malte Fliedner, Wirtschaftswissenschaftliche Fakult, 

Friedrich-Schiller-Universitaet Jena, Carl-Zeiss-Strasse 3, 07743, 

Jena, Thueringen, Germany, malte.fliedner@uni-jena.de, Nils 

Boysen 

Modern rail container terminals are an essential node of intermodal transportation 

networks. In such a terminal, several gantry cranes transship containers 

simultaneously from incoming trains to their designated destinations on the 

yard. In this work, we investigate the decision problem of positioning freight 

trains on the tracks of a terminal in order to balance workload among gantry 

cranes. We propose mathematical formulations of the core decision problem, 

provide complexity results and develop heuristic and exact algorithms to solve 

the problem and several extensions. 

2 - Parking Freight Trans in Rail-Rail Transshipment Yards: 

The Train Location Problem 

Nils Boysen, Lehrstuhl für ABWL/ Operations Management, 

Friedrich-Schiller-Universität Jena, Carl-Zeiß-Str. 3, 07743, 

Jena, Germany, nils.boysen@uni-jena.de, Malte Fliedner 

In modern rail-rail transshipment yards huge gantry cranes spanning all railway 

tracks allow for an ecient consolidation of containers among freight trains. An 

important decision problem during daily operations is the train location problem, 

which assigns each train to a railway track (vertical position) and decides 

on each train’s parking position on the track (horizontal position), so that the 

train processing time is minimized. For this problem different solution procedures 

are described and tested in a comprehensive computational study. 

3 - Truck Scheduling Problem in Intermodal Container 

Transportation 

Jenny Nossack, Managment Information Sciences, University of 

Siegen, Hölderlinstraße 3, 57068, Siegen, North

Rhine-Westphalia, Germany, jenny.nossack@uni-siegen.de, 

Erwin Pesch 

We address a truck scheduling problem that arises in intermodal container 

transportation, where containers need to be transported between customer 

places (shipper or receiver) and container terminals (rail or maritime) and vice 

versa. The transportation requests are handled by a trucking company which 

operates several depots and a fleet of homogeneous trucks that must be assigned 

and scheduled to minimize the total truck operating time under hard time window 

constraints imposed by the customers and terminals. Empty containers are 

considered as transportation resource and are provided by the trucking company 

for transportation. The truck scheduling problem at hand is formulated as 

Full Truckload Pickup and Delivery Problem and is solved by a 2-stage heuristic 

solution approach. Computational results for randomly generated instances 

are presented. 

4 - New bounds and algorithms for the Transshipment Yard 

Scheduling Problem 

Florian Jaehn, Business Administration, Management 

Information Science, Hoelderlinstrasse 3, 57068, Siegen, 

Germany, florian.jaehn@uni-siegen.de, Nils Boysen, Erwin 

Pesch 

In modern rail-rail transshipment yards huge gantry cranes transship containers 

between different freight trains, so that hub-and-spoke railway systems are enabled. 

In this context, we consider the transshipment yard scheduling problem 

(TYSP) in which trains have to be assigned to bundles, which jointly enter and 

leave the yard. Although feasible solutions can easily be obtained, the problem 

is NP-hard if certain, realistic objectives are chosen. We present several heuristics 

as well as one exact algorithm for solving the problem. The presentation 

concludes with some computational results. 

� MB-09 


6.2.53 

Optimization of Transport Problems on 

Networks I 



Chair: Simone Göttlich, Mathematics, TU Kaiserslautern, Post Box 

3049, 67653, Kaiserslautern, Germany, 

goettlich@mathematik.uni-kl.de 

1 - A measure theoretic approach to continuous and discrete 

flows over time 

Ebrahim Nasrabadi, Mathematics, Technische Universität Berlin, 

Institut für Mathematik, Sekr. MA 5-2, Straße des 17. Juni 136, 

10623, Berlin, Germany, nasrabadi@math.tu-berlin.de, Ronald 

Koch, Martin Skutella 

Network flows over time form a fascinating area of research. They model the 

temporal dynamics of network flow problems occurring in a wide variety of applications. 

Research in this area has been pursued in two different and mainly 

independent directions with respect to time modeling: discrete and continuous 

time models. 

In this talk we deploy measure theory in order to introduce a general model 

of network flows over time combining both discrete and continuous aspects 

in only one model. We focus on the maximum flow problem and extend the 

famous MaxFlow-MinCut Theorem to this setting. 

2 - An Adaptive Model Switching and Discretization Algorithm 

for Gas Flow on Networks 

Pia Domschke, Mathematics, TU Darmstadt, Dolivostr. 15, 

64293, Darmstadt, Germany, 

domschke@mathematik.tu-darmstadt.de, Oliver Kolb, Jens Lang 

The equations describing the transport of gas in pipeline networks are based on 

the Euler equations. Compressor stations and valves increase the complexity. 

We present a hierarchy of models that describe the flow of gas qualitatively 

different. Using adjoint techniques, we may specify model and discretization 

error estimators. We present a strategy that adaptively applies the different 

models in different regions of the network in order to reduce the complexity of 

the problem while maintaining the accuracy of the solution. This approach is 

suitable to be used within an optimization framework. 

EURO 24 - Lisbon 2010 MB-10 

3 - Approximate Parametric Dynamic Programming in Inventory 

Management 

Marco Laumanns, IBM Research Zurich, 8803, Rueschlikon, 

Switzerland, mlm@zurich.ibm.com, Stefan Woerner, Apostolos 

Fertis, Rico Zenklusen 

We present an approximation method for inventory management under multiple 

suppliers and stochastic lead times. The method is based on parametric dynamic 

programming using convex piecewise linear approximations of the differential 

cost function in approximate relative value iteration steps and works 

for a range of stochastic control problems with linear dynamics and convex 

piecewise linear immediate costs. The obtained differential cost approximation 

not only defines a policy that performs well in practice, it also yields valid 

lower bounds that are useful to judge the quality of heuristics. 

4 - Solution techniques for time-continuous network problems 

Simone Göttlich, Mathematics, TU Kaiserslautern, Post Box 



Many phenomena appearing in economics can be described by continuous 

models consisting of ordinary and partial differential equations. Typical application 

areas include supply chain management, scheduling problems and 

network flow problems in general. The focus is on the mathematical modelling 

as well as on techniques for simulation and optimization purposes. In fact, in 

various cases those models can be related to mixed-integer programming models. 

To ensure feasibility and to reduce the computational effort of large-scale 

instances, there is evidently need for suitable algorithms. 

� MB-10 


6.2.56 

Graphs and Networks II 



Chair: Jean Fonlupt, Maths, Université de Paris 6, 175 Rue du 

Chevaleret, F 75013, Paris, France, jean.fonlupt@math.jussieu.fr 

1 - The sandwich line graph 

Denis Cornaz, LAMSADE, Universite Paris-Dauphine, Pce Mal 

Lattre de Tasigny, 75016, Paris, France, 

cornaz@lamsade.dauphine.fr 

A sandwich function is one that computes a bound which dual both for the 

chromatic number and for the clique number of any given graph G. In an orientated 

graph D, two arcs form a simplicial pair if they share the same tail vertex 

and if the two distinct head vertices are linked by an arc. The sandwich line 

graph S(D) of an orientated graph D is the line graph L(G) of the underlaying 

undirected graph G minus the edges of L(G) corresponding to simplicial pairs 

of arcs of D. If D is any acyclic orientation of a graph G, both following sums 

are equal (as they are both equal to the number of vertices of G): The sum of 

the clique covering number of G and of the stability number of S(D); The sum 

of the stability number of G and of the clique covering number of S(D). From 

these relations one can use any sandwich function in S(D) in order to produce 

another sandwich function for G. Numerical experiments show that the quality 

of (polynomial) Lovasz Theta (sandwich) function is significantly improved for 

the chromatic number. 

2 - Computing solutions of the Paintshop-Necklace problem 

Frédéric Meunier, LVMT, Ecole Nationale des Ponts et 

Chaussées, 6-8 avenue Blaise Pascal, Cité Descartes, 77455, 

Marne-la-Vallée, France, frederic.meunier@enpc.fr 

How assign colors to occurences of cars in a car factroy ? How divide fairly 

a necklace between thives who have stolen it ? These two questions are two 

facets of a same combinatorial problem, which has for now attracted attention 

from a theoretical point of view. The main purpose of our work is to make a 

step in a more operational direction by discussing practical ways to compute 

solutions for instances of various sizes. Approaches, such as linear programming, 

valid inequalities, greedy algorithms, path-following methods, will be 

discussed. 

27


3 - Non-stable pairs in the Stable Marriage problem and 

multi-index configurations 

Pavlos Eirinakis, Management Science & Technology, Athens 

University of Economics & Business, 76 Patision Str., 10434, 

ATHENS, Greece, peir@aueb.gr, Dimitris Magos, Yiannis 

Mourtos, Panayiotis Miliotis 

In a Stable Marriage (SM) line-graph, some of the nodes corresponding to manwoman 

pairs that do not participate in any stable marriage cannot be removed, 

since the set of solutions would then get larger. We provide an algorithm that 

utilizes the rotation poset graph (a graph underlying the set of SM solutions) to 

identify all such nodes, thus deriving the minimal marriage graph. This allows 

us to determine the dimension and the facets of the SM polytope. Further, we 

explore the possibility of applying our work on a multi-index Supply Chain 

Network configuration and present some special cases. 

� MB-11 


8.2.38 

Technical Communication 



Chair: Stig C Holmberg, Information Technology and Media, Mid 

Sweden University, Mid Sweden University, ITM - Q351, 83125, 

Östersund, Sweden, shbg@ieee.org 

Chair: Lena-Maria Öberg, ITM, Mid Sweden University, 831 25, 

Östersund, Sweden, lena-maria.oberg@miun.se 

1 - Viable learning organizations — continuous change in 

a changeable world 

Lena-Maria Öberg, ITM, Mid Sweden University, Akademigatan 

1, 83125, Östersund, Sweden, lena-maria.oberg@miun.se, 

Christina Nyström, Hanna Olsson 

Organizations dealing with technical communication must be well prepared 

due to changes in the surroundings. Changes put new demands on competence, 

technology in use, routines as well as the existing doctrine. In this paper, we 

present a model which supports an organization’s continuously development of 

competence, doctrine, technology and routines. The know-how in the organization 

is gathered in an organizational memory (OM) — the Business Intelligence 

System. This BIS must be continuously updated in order to support identified 

business potentials. 

2 - XML-thinking for Technical Communication 

Stig C Holmberg, Information Technology and Media, Mid 


Östersund, Sweden, shbg@ieee.org, Lena-Maria Öberg 

All technical devices require adequate technical information (TI) in order to 

optimally serving us. The production and management of the necessary TI, 

however, often tend to be very cumbersome. Here XML-technologies (XML, 

XSLT, CSS and so on) are seen as a panacea. The problem, however, is that 

TC companies often oversee the need for an initial, careful and systemic information 

modelling and planning. Without such a planning the expectations of 

introducing XML often fall short. Hence, we here introduce XML-thinking as 

a set of guidelines for applying the XML-technologies to their full force. 

3 - Queueing Modeling and Analysis for a Web-Server System 

with Proxy Servers 

Yoshitaka Takahashi, Faculty of Commerce, Waseda University, 

1-Chome 6-1, Nishi-Waseda, 169-8050, Shinjuku, Tokyo, Japan, 

yoshitak@waseda.jp, Yoshiaki Shikata, Andreas Frey 

It is an important queueing issue to evaluate the delay in a web-server system 

with proxy servers handling real-time services. However, there exists almost 

no literature on the queueing models and analyses for the web-server system. 

The goal of this paper is to provide a queueing model and analysis. Using the 

diffusion approximation, we derive a mean-delay approximate formula. Our 

formula is validated by simulation results and is shown to be consistent with 

the previously-obtained exact result for a special case (single proxy-server and 

Poisson-arrival model). 

28 

4 - Churn Analysis with Bayesian Belief Network in 

Telecommunication Industry 

Y. Ilker Topcu, Industrial Engineering, Istanbul Technical 

University, Istanbul Teknik Universitesi, Isletme Fakultesi, 

Macka, 34367, Istanbul, Turkey, ilker.topcu@itu.edu.tr, Pınar 

Aykanat 

In this study, a Bayesian Belief Network (BBN) model is applied to find out 

the most important factors that have effects on customer churn in telecommunication 

industry. Since BBN depends on the conditional probabilities between 

variables, it is used to show the casual relations between variables. Average 

minutes of calls, average billing amount, the frequency of calls to people from 

different providers, and tariff type are found as the most important variables 

that explain churn. Finally, a number of scenarios are discussed and promotions 

are suggested to save the customers. 

� MB-12 


8.2.39 

AHP 02 


Processes 


Chair: Erdem Aksakal, Industrial Engineering, Engineering Faculty, 

Gazi University Engineering Faculty, Maltepe, 06570, Ankara, 

Turkey, eaksakal@gazi.edu.tr 

1 - Evaluations of Vessel Registration Alternatives after 

the Direct Shipping across Taiwan Strait Applying 

Fuzzy AHP 

Cheng-Chi Chung, Shipping and Transportation Management, 

National Taiwan Ocean University, No.2, Beining Rd., 

Jhongjheng District, Keelung City 202, Taiwan (R.O.C.), 202, 

Keelung, Taiwan, jackie@mail.ntou.edu.tw, Shih-Hao Yang 

Confronted by the competition of open registry nations and regions, the flagging 

out of vessels is getting more serious in the world; therefore the traditional 

maritime countries are all aggressively making relative shipping policies and 

arrangements in order to increase the tonnage of national fleet. Apart from economic 

considerations, the political reasons of the restriction of direct shipping 

across Taiwan Strait and forbidding the national fleet of the opposite side come 

into each own ports in the past have resulted in vessels’ flagging out among 

Taiwanese shipping corporations. After signing the shipping agreement across 

Taiwan Strait on 15 December 2008, the only qualification of "the vessel, which 

is registered in Taiwan, China, or Hong Kong, owned by above regions’ shipping 

corporations’ could sail across Taiwan Strait directly. Therefore, shipping 

corporations in Taiwan who operate cross Strait business have to decide vessel’s 

registration in Taiwan or in Hong Kong, or registering in China through joint 

venture. The change of operational environment together with the implication 

of tonnage tax system and the loose of employing foreign seafarers’ restrictions 

in Taiwan, all influential factors, will impact shipping corporations in Taiwan 

to evaluate the alternatives of vessel registration among Taiwan, Hong Kong, or 

China. This study will construct an evaluation framework of vessel’s registration 

by using Fuzzy AHP, and find out the optimal alternative among Taiwan, 

Hong Kong, or China. The results of this paper can serve as a good strategic 

reference on vessel registrations to the shipping corporations in Taiwan. 

2 - A hybrid approach to personnel selection problem: 

AHP and Weighted Fuzzy Axiomatic Design 

Erdem Aksakal, Industrial Engineering, Engineering Faculty, 

Gazi University Engineering Faculty, Maltepe, 06570, Ankara, 

Turkey, eaksakal@gazi.edu.tr, Metin Dagdeviren, Ergun Eraslan 

The aim of this study is to describe a hybrid model for supporting the personnel 

selection process in service business. Selecting the right personnel is like 

having an ahead step for the future in the name of the organization. On the 

other hand selecting the wrong or improper personnel affects the organizations 

in a negative way. Selecting a new personnel under these conditions make the 

condition difficult and because of its composition, the personnel selection process 

creates its own alternatives. Selecting the proper personnel among those 

alternatives is a multi-criteria decision making problem. In this study the use of 

analytic hierarchy process and weighted fuzzy axiomatic design in personnel 

selection process investigated.

3 - A Hybrid Fuzzy Approach for a Point-Factor Job Evaluation 

System 

Ahmet Can Kutlu, Industrial Engineering, Istanbul Technical 

University, Istanbul Technical University Management Faculty 

Industrial Engineering Department, Macka Istanbul, 34367, 

Istanbul, Turkey, kutluahm@itu.edu.tr, Hülya Behret 

In this study, a fuzzy approach is developed for job evaluation which is defined 

as the methods and practices of ordering jobs or positions with respect 

to their value or worth to the organization. The most comprehensive method 

used in job evaluation is the Point-Factor Method where a set of compensable 

factors are identified as determining the worth of jobs. In this paper, factors are 

weighted by using Fuzzy Analytical Hierarchy Process (FAHP) method. After 

determining the factor weights, the evaluated jobs are grouped by using Fuzzy 

Rule Based System (FRBS) to be a base for a waging system. 

4 - Selecting the software that will be developed for the 

purchasing department using fuzzy AHP 

Nurgül Demirta¸s, Yildiz Technical University, 0090, Istanbul, 

nurguldemirtas@gmail.com, Özge Nalan Alp, Hayri Baraçlı 

Software applications are very important to follow operations in a company. 

These applications are influenced by many factors when they are being developed. 

This situation is result in using MCDM technique. MCDM techniques 

are the most effective techniques to achieve the most appropriate decision when 

there are lots of criteria. In this study, firstly lacks of the software application 

which is used in purchasing department will be determined and then the most 

effective software application platform will be selected to developed applications 

by Fuzzy AHP which is the one of MCDM techniques. 

� MB-13 


2.2.21 

Facility Location and Supply Chain 

Management 



Chair: Isabel Correia, New University of Lisbon, Lisbon, Portugal, 

isc@fct.unl.pt 

1 - Robust models for dynamic multilevel facility location 

Marina Gebhard, Chair of Business Administration and 

Logistics, University Erlangen-Nuremberg, Lange Gasse 20, 

90403, Nürnberg, Germany, 

marina.gebhard@wiso.uni-erlangen.de, Jens Wollenweber 

We study a strategic capacity planning and facility location problem in supply 

chains operating under uncertainty. Our focus lies on the uncertainty of the customer 

demands which are modeled by alternative future scenarios. We present 

a new formulation for a robust dynamic multilevel capacitated facility location 

problem that minimizes the expectation of the relative regrets associated 

with the given scenarios. We compare our formulation to both a deterministic 

and an alpha-reliable mean-excess regret model formulation and analyze their 

differences in terms of robustness and effectiveness. 

2 - Optimal development of a national supply system in 

Germany and its geo-economic impact on the states 

and regions of it 

John Karkazis, Business School, University of the Aegean, 

Chios, GR-82100, Chios, Greece, ikarkazis@aegean.gr, Vassilis 

Angelis, Christina Chalimourda, Maria Mavri 

In this paper the concept of regional efficiency is introduced and the role of geoeconomic 

gravity centers in regional development is analyzed. In the above 

context, the paper focuses on the geo-economic ability of administrative units 

to act as supply centers covering regional demand for products and services 

with the minimum operational and transportation cost. This theoretical framework 

is applied to the case of Germany in order to evaluate the optimal location 

of supply centers in a single, double and triple national supply system on the 

road network of the country. 

3 - A Heuristic to Locate Distribution Centers in Large Supply 

Chains 

Jean-Sébastien Tancrez, Operations Management, EPFL, Ecole 

Polytechnique Fédérale de Lausanne, EPFL - TOM, Odyssea 


2.19, Station 5, 1015, Lausanne, Switzerland, 

jean-sebastien.tancrez@epfl.ch, Jean-Charles Lange, Pierre 

Semal 

Designing supply chain networks is crucial for companies but remains challenging. 

In this work, we study the location of distribution centers in a three 

layer network, with single commodity and single period. We include both 

transportation and inventory costs and allow direct flows from factories to customers. 

A continuous model is proposed, which decomposes into a simple 

problem (batch sizes) and a linear-like program (locations and flows). An iterative 

heuristic is then built on this, allowing to solve large realistic problems. 

The approach is exemplified on the case of a glass manufacturer. 

4 - A two-echelon supply chain network design problem 

with facility sizing decisions and variable operating 

costs 

Francisco Saldanha-da-Gama, CIO/DEIO, University of Lisbon, 

FCUL-DEIO, Bloco C6, Piso 4, 1749-016, Lisbon, Portugal, 

fsgama@fc.ul.pt, Isabel Correia, Teresa Melo 

A two-echelon supply chain network design problem is considered. In each 

echelon and at each potential site, a new facility can be set up for a group of 

product families. The size of the storage area occupied by a product family is 

to be selected from a discrete set of available sizes. Operating costs depend 

on the type of storage area installed in a given location and are charged to the 

total product quantity that is stored. A minimal cost network is to be determined. 

Several formulations are discussed and the results of a computational 

study based on randomly generated data are presented. 

� MB-14 


2.2.15 

Lotsizing and Supply Chain Planning 



Chair: Herbert Meyr, Chair of Production and Supply Chain 

Management, Technical University of Darmstadt, Hochschulstr. 1, 

64289, Darmstadt, Germany, Meyr@bwl.tu-darmstadt.de 

1 - Simultaneous lotsizing and scheduling for timedependent 

production speeds: dual reoptimization revisited 

Herbert Meyr, Chair of Production and Supply Chain 

Management, Technical University of Darmstadt, Hochschulstr. 

1, 64289, Darmstadt, Germany, Meyr@bwl.tu-darmstadt.de 

The GLSPPL determines lotsizes and production schedules for a given deterministic 

dynamic demand in a capacitated, 1-stage production system consisting 

of heterogeneous production lines. Meyr (2002) proposed a dual reoptimization 

heuristic for the GLSPPL, where generalized network flow problems 

are solved as embedded subproblems. In order to tackle a practical application 

requesting time-dependent productions speeds, it was necessary to adapt the 

principle of dual reoptimization to general LPs. The paper shows experiences 

with this adaptation and computational tests with different LP solvers. 

2 - Heuristics for the General Lotsizing and Scheduling 

Problem for Multiple production Stages 

Florian Seeanner, Rechts- und Wirtschaftswissenschaften, 

Fachgebiet Produktion & Supply Chain Management, 

Hochschulstr. 1, 64289, Darmstadt, Germany, 

seeanner@bwl.tu-darmstadt.de 

The General Lotsizing and Scheduling Problem for Multiple production Stages 

determines simultaneously lotsizes and setup sequences for a two or three-level 

continuous production system as can be found in the consumer goods industry. 

Unfortunately, this model can hardly be solved by standard MIP solvers, even 

for small problem instances. Therefore heuristic approaches have to be developed. 

As corresponding real-world-problems are quite large in data terms, 

scalability is a special requirement for those heuristics. In this talk scalable 

heuristics and preliminary computational results are shown. 

29


3 - Supply Chain Planning in Grocery Retail - An Operations 

Planning Framework 

Michael Sternbeck, Logistics, Catholic University of Eichstaett 

Ingolstadt, Auf der Schanz 49, 85049, Ingolstadt, 

michael.sternbeck@ku-eichstaett.de, Alexander Hübner, 

Heinrich Kuhn 

In our presentation we develop a consumer-back demand planning framework 

for grocery retailing to exemplify the interrelated planning issues. This integrates 

retail specifics, consumer interaction, and hierarchical and sequential 

decision aspects. We use two case examples to exemplify the interrelated planning 

issues. The first will demonstrate the intertwined impact of warehousing, 

distribution and instore logistics. The second will show hierarchical planning 

aspects for integrated assortment, shelf space and price management. 

4 - A hierarchical planning approach to physical and financial 

supply chain management 

Gerd J. Hahn, Chair of Production and Operations Management, 

Catholic University of Eichstaett-Ingolstadt, Auf der Schanz 49, 

85049, Ingolstadt, Germany, gerd.hahn@kuei.de, Heinrich Kuhn 

The physical and the financial perspective of the supply chain are inextricably 

interlinked. However, hierarchical approaches to supply chain planning 

mainly focus on material flows and omit financial flows as well as their impact 

on shareholder value creation. In this presentation, we introduce an integrated 

approach to hierarchical supply chain planning optimizing Economic 

Value Added (EVA) as a prevalent mid-term indicator of value-based financial 

performance. Robust optimization methods are utilized to manage operational 

supply chain risk due to the uncertainty of future events. 

� MB-15 


2.2.12 

Rich routing problems 



Chair: David Pisinger, DTU Management, Produktionstorvet 424, 

2800, Kgs. Lyngby, Denmark, pisinger@diku.dk 

1 - Reliability in Vehicle Routing Problem with Time Windows 

Duygu Tas, Industrial Engineering and Innovation Sciences, 

TU/e Eindhoven University of Technology, Technische 

Universiteit Eindhoven P.O. Box 513, 5600 MB, Eindhoven, 

Netherlands, d.tas@tue.nl, Nico Dellaert, Tom Van Woensel, Ton 

de Kok 

In the Vehicle Routing Problem with Time Windows (VRPTW), a time window 

can be considered as a service requirement set by the customer. Routing schedules 

need to meet these requirements to provide adequate delivery reliability. 

We consider the VRPTW with stochastic travel times with a known probability 

distribution and the reliability for all customers via an objective function which 

minimizes the weighted sum of the total expected delay calculated with respect 

to the TWs and the total distance. We will describe our model and present our 

results. 

2 - Dynamic routing in congested networks using real-time 

ITS information 

Ratna Babu Chinnam, Industrial & Manufacturing Engineering, 

Wayne State University, 4815 Fourth Street, 48202, Detroit, MI, 

United States, r_chinnam@wayne.edu, Ali Guner, Alper Murat 

Growing congestion on transportation networks is impacting the effectiveness 

of just-in-time (JIT) logistics through increased delays and variability. We 

propose compact yet effective models for characterization of congestion and 

offer stochastic dynamic models and methods for routing under real-time information. 

We illustrate applications and methods for single as well as milkrun 

deliveries subject to time windows. Results from a simulated network of 

Southeast-Michigan freeways using historical data from Michigan ITS are presented. 

3 - Driver-Pooling for Vehicle Routing Problems 

30 

Christian Doppstadt, IT-based Logistics, Goethe University 

Frankfurt, Grueneburgplatz 1, 60323, Frankfurt, Germany, 

doppstadt@wiwi.uni-frankfurt.de, Michael Schwind, Michael 

Schneider 

Most research done on vehicle routing problems (VRPs) considers drivers as a 

homogeneous set. This does not reflect the real world situation, where travel 

and service times are highly driver dependent e.g. due to learning effects. If 

inhomogeneous drivers are considered, a decision on which drivers to employ 

for a given problem instance is required. To our knowledge, we are the first 

to investigate this issue. We develop a method to determine the optimal set of 

drivers out of a driver pool. The results are used as input for an existing VRP 

heuristic supporting driver dependent times. 

4 - Order batching and routing with time windows in warehouse 

picking 

Katja Prnaver, Theoretical computing, Institute of Mathematics, 

Physics and Mechanics, Jadranska 19, 1000, Ljubljana, Slovenia, 

katja.prnaver@imfm.uni-lj.si, David Pisinger 

In warehouses, effective consolidating of order into batches may result in significant 

reduction of total warehouse costs. Past research has mainly been focused 

on reducing traveled time of pickers, with some recent analytical results 

for single- or 2-block warehouses. We will present an optimization algorithm 

for order batching with time windows, minimizing the number of pickers required 

and the total traveled distance. The algorithm works for block structured 

warehouses of arbitrary size. Experimental results, showing that 5-10% savings 

can be obtained, will be presented. 

� MB-16 


2.2.14 

Robust optimization in public transport 



Chair: Laura Galli, DEIS, University of Bologna, Viale 

Risorgimento, 2, 40136, Bologna, Italy, l.galli@unibo.it 

1 - Robust Aircraft Rotation Planning 

Ivan Dovica, Department of Applied Mathematics, Charles 

University, Malostranske namesti 25, 11800, Prague, Czech 

Republic, ivan@kam.mff.cuni.cz, Ralf Borndörfer, Ivo Nowak, 

Thomas Schickinger 

This talk addresses the robust aircraft rotation planning problem, for which we 

propose a novel column generation method. It is based on a stochastic model 

of the operation of an airline. The algorithmic core is a method to compute delay 

distributions on legs along individual aircraft rotations. To investigate the 

practical relevance of the method, a Monte-Carlo simulation including standard 

recovery actions is applied. 

2 - A Lagrangian Heuristic for Robust Train Timetabling 

Valentina Cacchiani, DEIS, University of Bologna, Viale 

Risorgimento 2, 40136, Bologna, Italy, 

valentina.cacchiani@unibo.it, Alberto Caprara, Matteo Fischetti 

Lagrangian heuristics approximately solve a relaxation of the problem at hand 

through an iterative optimization scheme and apply several times a heuristic 

driven by the Lagrangian dual information. We present a simple modification of 

this scheme so as to deal with robustness: we modify the problem formulation 

by introducing artificial parameters to control robustness and we dynamically 

change their weights so as to produce subproblems where robustness becomes 

more important. The approach is illustrated on the Train Timetabling Problem 

and tested on real-world instances. 

3 - Recoverable Robustness for Railway Rolling Stock 

Planning 

Laura Galli, DEIS, University of Bologna, Viale Risorgimento, 

2, 40136, Bologna, Italy, l.galli@unibo.it, Valentina Cacchiani, 

Alberto Caprara, Leo Kroon, Gabor Maroti, Paolo Toth 

In this paper we apply the notions of Recoverable Robustness and Price of 

Recoverability to railway rolling stock planning using a Benders decomposition 

approach. We evaluate the approach on real-life rolling stock planning 

problems of NS, the main operator of passenger trains in the Netherlands. We 

present a successful heuristic leading to robust solutions, whose value is very 

close to the continuous lower bound. Finally, we show the practical effectiveness 

of our method by proposing an evaluation framework and showing 

extensive computational results on a very large number of scenarios.

4 - Robust train platforming and routing 

Gabor Maroti, Department of Decision and Information 

Sciences, Rotterdam School of Management, Erasmus University 

Rotterdam, Burg Oudlaan 50, 3062 PA Rotterdam, The 

Netherland, 3062 PA, Rotterdam, Netherlands, gmaroti@rsm.nl, 

Alberto Caprara, Laura Galli, Leo Kroon, Paolo Toth 

Train Platforming and Routing is a problem of assigning to each train of a given 

timetable a platform and a conflict-free route through a railway node. The problem 

appears both in early planning stages and in real-time operations. In this 

work we focus on the robustness of the problem. 

We extend an existing platforming model to robustness considerations, and we 

design an optimization-based rescheduling algorithm. Further, we evaluate the 

algorithm in a simulation framework. The computational tests are based on 

real-life instances of the Italian train operator FS. 

� MB-17 


1.3.14 

Collaborative Planning II 



Chair: Karl-Heinz Küfer, Optimization, Fraunhofer ITWM, 

Fraunhofer-Platz 1, 67663, Kaiserslautern, Germany, 

kuefer@itwm.fhg.de 

Chair: Herbert Kopfer, Department of Business Studies & 

Economics, Chair of Logistics, University of Bremen, 



1 - Limit and Degree of Autonomy in Groupage Systems 

Herbert Kopfer, Department of Business Studies & Economics, 

Chair of Logistics, University of Bremen, 


kopfer@uni-bremen.de, Heiko Kopfer, Xin Wang 

Groupage systems enable collaborating forwarders to exchange transportation 

requests in order to harmonize their transportation plans. This contribution 

describes several approaches for collaboration in groupage systems. It assigns 

the described approaches to different stages of collaboration intensity. For each 

stage it analyses the resulting degrees of autonomy of the cooperating partners. 

The main distinguishing features of the presented approaches are illustrated 

and their influence on the degree of autonomy is discussed. 

2 - A Collaboration Platform for Freight Carriers 

Heiner Ackermann, Abteilung Optimierung, Fraunhofer ITWM, 


heiner.ackermann@itwm.fraunhofer.de, Hendrik Ewe, Herbert 

Kopfer, Karl-Heinz Küfer 

We present a collaboration platform for cooperating freight carriers that encourages 

complex exchanges of requests by means of an auction. The platform’s 

design goals are threefold: maximize the collaboration’s profit, keep participants’ 

autonomy regarding offering and bidding and giving maximal planning 

reliability concerning the exchange. We also discuss several decision support 

problems increasing the platform’s usability and present first results gathered 

using a prototypic implementation. 

3 - Profit Sharing Among Collaborating Freight Carriers 

Hendrik Ewe, Department of Optimization, Fraunhofer Institute 

for Industrial Mathematics (ITWM), Fraunhofer-Platz 1, 

D-67663 , Kaiserslautern, Germany, ewe@itwm.fhg.de, Heiner 

Ackermann, Herbert Kopfer, Karl-Heinz Küfer 

Sharing the jointly generated profit is a fundamental issue when it comes to collaboration 

among independent freight carriers. However standard profit sharing 

schemes such as the core or the Shapley value fail to be applicable due 

to several reasons. Based on a monetary flow network we characterize the set 

of feasible profit shares having a desirable locality property. We then discuss 

several specific sharing schemes derived from the network model. 


4 - A Combinatorial Auction Framework and Models for 

Carriers’ Collaboration in Less Than Truckload Transportation 

Chen Haoxun, 12, rue Marie Curie - BP 2060, 10010, TROYES, 

haoxun.chen@utt.fr, Bo Dai 

In collaborative logistics, multiple carriers or shippers may form an alliance 

to optimize their transportation operations by sharing vehicle capacities and 

delivery tasks for eliminating empty backhaul, raising vehicle utilization rate, 

increasing each actor’s profit. We propose a combinatorial auction framework 

and models for carriers’ collaboration in less than truckload transportation. It 

consists of 5 steps: Outsourcing Request Selection by each carrier, Price Setting 

by auctioneer, Bid Generation of each carrier, Winner Determination by 

auctioneer, Profit Allocation among carriers. 

� MB-18 


1.3.15 

Stochastic Modeling and Simulation I 



Chair: Alexander Grusho, Probability theory, Moscow State 

University, GSP-2, GSP-2, GSP-2, Moscow, grusho@yandex.ru 

1 - Fractal dimension cluster validation criteria 

Dvora Toledano-Kitai, Haifa University, 21982, Karmiel, Israel, 

dvora@braude.ac.il, Renata Avros, Zeev Barzily, Zeev 

(Vladimir) Volkovich 

We propose a new standpoint to the cluster validation problem based on a fractal 

dimension cluster quality model. Clusters dimension values in a partition 

are repeatedly assessed by means of simulated samples occurrences in groups 

such that the proximity of the result obtained is interpreted as partition goodness. 

Most concentrated at the origin empirical distribution of the calculated 

dimensions differences indicates an estimated number of clusters. Numerical 

experiments presented for different fractal dimensions exhibit capability of the 

offered method in artificial and real datasets. 

2 - On Goodness-Of-Fit Tests For Random Combinatorial 

Objects 

Alexander. Kolodzey, Institution of Russian Academy of 

Sciences Dorodnicyn Computing Centre of RAS, Shokalskogo 

proezg, 1/1 - 68, 127229, Moscow, Russian Federation, 

kolodzey@mtu-net.ru 

We consider the random combinatorial objects which possess the decomposition 

property of the individual components. The weight of the object equals 

to the sum of the components weights, and joint distribution of the number 

of components with a given weight can be represented as the joint conditional 

distribution of several independent Poisson random variables. A theorem about 

large deviations probabilities is proved, and Bahadur efficiency criteria of fit 

for random combinatorial objects are studied. 

3 - Learning Parameter Optimization of Stochastic Gradient 

Descent with Momentum for a Stochastic Quadratic 

Memmedaga Memmedli, Statistics, T.C. Anadolu University, 

T.C. Anadolu University, Faculty of Science, Department of 

Statistics, 26470 , Eskisehir, Turkey, 

mmammadov@anadolu.edu.tr, Engin Tas 

Stochastic gradient descent (SGD) with momentum is a competitive optimization 

method in particular classification problems with large and redundant data 

sets. Learning rate and momentum factor parameters should have to be carefully 

tuned. We propose to determine learning rate and momentum adaptively 

using the second-order information embedded in the Hessian. Convergence 

speed of stochastic gradient descent with adaptively tuned learning parameters 

(eSGD) compared with standard SGD on rosenbrock performance function 

with noisy measurements. Results show that eSGD outperforms standart SGD. 

4 - Identification of Local Distortions in Random Sequences 

Alexander Grusho, Probability theory, Moscow State University, 

GSP-2, GSP-2, GSP-2, Moscow, grusho@yandex.ru, Elena 

Timonina, Zeev (Vladimir) Volkovich 

31


We consider a sequence of consistent distributions P(0, n) with supports D(0, n) 

on the product Xn of a finite set X . The distortion is defined by functions f(m, 

n), mapping Xn in Xn which does not change last n-m coordinates, and define 

a consistent sequence of distributions. Here, the consistent sequence of tests 

for the hypothesis H(0, n): P(0, n) against H(1, n): P(1, n) exists iff there exists 

n > m-1, where intersection D(0, n) and D(1, n) is the empty set. That specifies 

the ability to find local distortions in a random sequence, which distribution is 

well defined. 

� MB-19 



Modelling the Human Decisions 

Stream: Dynamical Systems and Game Theory 


Chair: Helena Ferreira, Rua Dr. José Sampaio n o 861, 4810-275, 

Guimarães, helenaisafer@gmail.com 

Chair: Abdelrahim Mousa, Mathematics, University of Minho, 

Eshpinho, Eshpinho, PhD Student, Portugal, 

abedmousa2000@yahoo.com 

1 - Collaborative Dominance: When Doing Unto Others as 

You Would Have Them Do Unto You Is Rational 

Leandro Rego, Statistics, Universidade Federal de Pernambuco, 

Rua Muniz Tavares 25, apt. 902, Jaqueira, 52050-170, Recife, 

PE, Brazil, leandro.ufpe@gmail.com, Filipe Souza 

We discuss the intuition of mixed equilibria and analyze in what situations making 

the other players indifferent among their strategies is rational. We define 

the concept of stable collaborative dominance and, from it, we concluded that 

in 2x2 games, if there is a pair of stable collaborative dominant strategies, then 

the mixed equilibrium is not appropriate. Moreover, we show that there are 

situations in which the pair of collaborative dominant strategies is unstable, but 

players are able to cooperate via burning money agreements reaching a better 

situation than playing the mixed equilibrium. 

2 - Modeling Human Behavior 

Helena Ferreira, Rua Dr. José Sampaio n o 861, 4810-275, 

Guimarães, helenaisafer@gmail.com, Leandro Almeida, José 

Cruz, Alberto A. Pinto 

We apply Game Theory concepts to the Theory of Planned Behavior that studies 

the decision-making mechanisms of individuals. We propose the Bayesian- 

Nash Equilibria as one, of many, possible mechanisms of transforming human 

intentions in behavior. This process corresponds to the best strategic individual 

decision taking in account the collective response. We show that saturation, 

boredom and frustration can lead to splitted strategies, in opposition to no saturation 

that leads to a constant strategy. Furthermore, we study the role of 

leaders in individual/group behavior and decision-making. 

3 - Pinto’s individuals decision bussola 

Abdelrahim Mousa, Mathematics, University of Minho, 

Eshpinho, Eshpinho, PhD Student, Portugal, 

abedmousa2000@yahoo.com, Alberto A. Pinto, Marta Faias, 

Mohammad Mousa, Rasha Samarah, Gabriela Goes 

Here, we introduce the Yes-No decision model that is a simplified version of 

the general decision models. In this model, there are just two possible decisions 

(Yes or No) that individuals can take. We note that this model has several applications 

in psychology, education and economics. This model already exhibits 

all the high complexity of the split and no-split Nash equilibria that is common 

in decision models. We completely characterize the split and no-split equilibria, 

through a usual concept of dynamical systems, introducing Pinto’s thresholds 

in decision Nash equilibria. These thresholds describe hysteretic-like behavior, 

as in dynamics, that it is responsible by the occurrence of catastrophes consisting 

in abrupt changes of individuals and collective behavior. The way these 

thresholds evolve and interact with parameter change, called bifurcations in 

dynamical systems, it is completely characterized by Pinto’s bussola. These 

allow us to understand how small changes in psychological behavior or social 

behavior can create or annihilate possible individuals or collective behaviors to 

occur. 

4 - Delayed Effects of Fiscal Policy Measures 

32 

Orlando Gomes, Economics, ISCAL - Instituto Politécnico de 

Lisboa, Av. Miguel Bombarda, 20, 1069-035, Lisboa, Portugal, 

omgomes@iscal.ipl.pt, Ana Marques 

This presentation addresses the impact of fiscal policy changes over an economy 

resting on its long-term steady-state. The discussion, based upon a conventional 

intertemporal optimization setup, involves the consideration of a peculiar 

form of bounded rationality: it is assumed that only a small share of 

households is able to instantly recompute the optimal solution once the value 

of a tax rate is disturbed; all the other agents will then, gradually,follow the 

behavior of the first group (this can occur through contagion, social influence 

or social learning). As a result, the convergence towards the post-perturbation 

steady-state tends to follow a diffusion process and, consequently, policy measures 

will take time in affecting pervasively consumption-savings and laborleisure 

choices. 

� MB-20 


1.3.33A 




Chair: Andreas Bortfeldt, Dept. of Information Systems, University 

of Hagen, Profilstrasse 8, 58084 Hagen, Germany, 58084, Hagen, FR 

Germany, Germany, andreas.bortfeldt@fernuni-hagen.de 

1 - Ordering rules of the set of pieces in 2D- Open Orthogonal 

Dimensional Problem 

Garcia Perez, Matematicas Aplicadas, Universidad de Alcala de 

Henares, C/ Camilo Jose Cela n o 63-A, 2 o C, 28806, Alcala de 

Henares, Madrid, Spain, fernando.garciap@uah.es, Joaquin 

Aranda Almansa, Miguel Delgado Pineda 

We study the 2-Open Dimensional Orthogonal Problem. In this problem, it is 

usual to order the items by decreasing height or decreasing width to improve 

the best solution. We consider different ordering rules for the set of pieces 

and we analyze which one will probably be a good ordering depending on the 

objective, algorithm or set of pieces using 148 examples on the literature. 

2 - Rect-TOPOS: A constructive heuristic for the rectangle 

packing area minimization problem 

A. Miguel Gomes, Fauculty of Engineering / INESC Porto, 

University of Porto, Rua Dr. Roberto Frias s/n, 4200-465, Porto, 

Portugal, agomes@fe.up.pt, Marisa Oliveira, M a Eduarda Pinto 

Ferreira 

To solve the rectangle packing area minimization problem we use a variation 

of the TOPOS algorithm. In our adaptation, the layout is build by successively 

adding a new rectilinear piece to a partial solution while minimizing the enclosing 

rectangular area. Several criteria to choose the next piece and its orientation 

are proposed and compared. To evaluate and compare partial solutions different 

objective functions were used. Supported by Fundação para a Ciência e 

Tecnologia (FCT) project PTDC/GES/72244/2006. 

3 - A superior piecewise linearization approach for assortment 

problems 

Pei-Chun Wang, Graduate Institute of Industrial and Business 

Management, Taipei University of Technology, No. 1, Sec. 3, 

Chung Hsiao E. Road, Taipei 106, Taiwan, 106, Taipei, Taiwan, 

t5749005@ntut.edu.tw 

The aim of the assortment problem is to seek the best way of placing a given 

set of rectangular pieces within a large rectangle with minimal area. Such problems 

are often formulated as a quadratic mixed-integer program. This study reformulates 

the assortment model as a mixed-integer linear program for finding 

the global optimum within tolerable error based on a superior piecewise linearization 

approach. Numerical examples also demonstrate that the presented 

method uses much less CPU time than that in current methods for reaching the 

global optimum. 

4 - A two-stage approach for the rectangle packing area 

minimization problem 

Andreas Bortfeldt, Dept. of Information Systems, University of 

Hagen, Profilstrasse 8, 58084 Hagen, Germany, 58084, Hagen, 

FR Germany, Germany, andreas.bortfeldt@fernuni-hagen.de, 

Götz Goldacker

The rectangle packing area minimization problem (RPAMP) requires packing 

a set of rectangles without overlaps while minimizing the enclosing rectangular 

area. We assume that the small rectangles have fixed dimensions. A two-stage 

generic approach is presented that is able to integrate different 2D container 

loading procedures. Results are reported for well-known RPAMP benchmark 

instances. 

� MB-21 


6.2.47 

OR in Practice II 



Chair: Bernard Lamond, Operations & Decision Systems, Universite 

Laval, Pavillon Palasis-Prince, 2325, rue de la Terrasse #2634, G1V 

0A6, Quebec, Quebec, Canada, bernard.lamond@fsa.ulaval.ca 

1 - MES developing practice in tobacco manufacturing: A 

Case Study 

Zhong Yao, Dept. of Information System, Beihang University, 

37, Xueyuan,Haidian, 100191, Beijing, China, 

stiphen_zhong@126.com 

Many firms in China had implemented the ERP systems, including the tobacco 

manufacturing firms. However, these firms found that traditional ERP systems 

like SAP and Oracle can not satisfied data requirements in their routine production 

management. Therefore, they seek new methods to support their production 

management. Manufacturing executive systems (MES) can be considered 

one of the most effective systems and sometimes it is described as the ERP system 

in process manufacturing industry. This talk briefly introduces the MES 

building process in one of tobacco manufacturing firm of China. We present 

the general system developing methods and tools used in the MES development 

via MES practice in a tobacco manufacturing firm. We hopefully provide 

a reference method for similar manufacturing firms. 

2 - Comparing the Metaheuristics and Gradient Based Approaches 

in Multi-objective Urban Water Management 

Mahdi Zarghami, Faculty of Civil Engineering, University of 

Tabriz, 29 Bahman Blvd., 51664, Tabriz, zarghaami@gmail.com, 

Hassan Hajykazemian, Vahid Eghbaal 

In this paper, the efficiency of the evolutionary multi-objective optimization approaches 

are compared with respect to different gradient based methods. These 

methods are applied to an urban water problem in the large city of Tabriz, Iran. 

The evolutionary approaches include Genetic Algorithm, Particle Swarm Optimization 

and their new developed methods. According to the results, there are 

several trade-offs in selecting the most efficient methods and this paper will finally 

show the Pareto frontiers, which is needed to select the optimum solutions 

in uncertain conditions. 

3 - Dynamic control of a flexible machine with stochastic 

tool life 

Bernard Lamond, Operations & Decision Systems, Universite 

Laval, Pavillon Palasis-Prince, 2325, rue de la Terrasse #2634, 

G1V 0A6, Quebec, Quebec, Canada, 

bernard.lamond@fsa.ulaval.ca 

We present analytical results related to a tool management model for a flexible 

machine equipped with a tool magazine, variable cutting speed, and sensors 

to monitor tool wear, when tool life due to flank wear is stochastic. Under 

some regularity conditions, we derive the special structure of decision rules for 

adjusting the cutting speed as a function of remaining distance, each time a 

tool change occurs, in order to minimize the expected processing time (sum of 

cutting time and tool setup time). These are near-optimal and easy to compute. 

� MB-22 



Teaching with cases 



Chair: Peter Bell, Richard Ivey School of Business, University of 

Western Ontario, N6A 3K7, London, Ontario, Canada, pbell@ivey.ca 


1 - Workshop on Teaching OR using cases I 

Peter Bell, Richard Ivey School of Business, University of 

Western Ontario, N6A 3K7, London, Ontario, Canada, 

pbell@ivey.ca, Fredrik Odegaard, Mehmet Begen 

Workshop on Teaching OR using cases 

This workshop will provide practical help for instructors who wish to provide 

their OR students with real world problem solving experience using cases. We 

aim to provide help with such questions as: 

Where do I find OR cases? 

How can I use one or more cases in my course? 

What impact will the use of cases have on my students? 

What do I (the instructor) have to do to use a case successfully? 

2 - Workshop on Teaching OR using cases II 

Fredrik Odegaard, Richard Ivey School of Business, University 

of Western Ontario, 1151 Richmond Street North, N6A 3K7, 

London, Ontario, Canada, fodegaard@ivey.uwo.ca, Mehmet 

Begen, Peter Bell 




aim to provide help with such questions as: Where do I find OR cases? How 

can I use one or more cases in my course? What impact will the use of cases 

have on my students? What do I (the instructor) have to do to use a case successfully? 

3 - Workshop on Teaching OR using cases III 

Mehmet Begen, Richard Ivey School of Business, University of 

Western Ontario, 1151 Richmond St. N., Ivey, N6A3K7, 

London, ON, Canada, mbegen@ivey.uwo.ca, Peter Bell, Fredrik 

Odegaard 







4 - Workshop on Teaching OR using cases IV 

Antonia Carravilla, FEUP / INESC Porto, R. Dr. Roberto Frias, 

378, 4200-465, Porto, Portugal, mac@fe.up.pt 







� MB-23 


6.2.49 


Combinatorial Optimization II 



Chair: José Santos, Department of Mathematics, University of 

Coimbra, Department of Mathematics, FCTUC, Apartado 3008, 

3001-454 Coimbra, 3001-454, Coimbra, Portugal, zeluis@mat.uc.pt 

1 - Empirical study for the multi-objective shortest path 

problems using real data instances 

José Santos, Department of Mathematics, University of Coimbra, 

Department of Mathematics, FCTUC, Apartado 3008, 3001-454 

33


Coimbra, 3001-454, Coimbra, Portugal, zeluis@mat.uc.pt, José 

Paixão 

The real data instances have, in general, a specific structure turning the multiobjective 

shortest path problem (MSPP) harder to solve than when randomly 

generated instances are used. 

In this work, we study the number of Pareto optimal solutions on this kind of 

networks and compare it with the results obtained in randomly instances. Additionally, 

we compare the performance of the labelling and ranking algorithms 

to solve the MSPP when real data instances are considered. 

2 - Dynamic Location-Allocation Facility Modeling using a 

GIS-based Multiobjective Approach 

Lino Tralhão, INESC Coimbra, Rua Antero de Quental, 199, 

3000-141, COIMBRA, Portugal, linotralhao@gmail.com, João 

Coutinho-Rodrigues, José Paulo Almeida, Luís Alçada 

This work introduces a dynamic, multiobjective, mixed integer programming 

approach to deal with facility location-allocation problems. Aspects such as 

the planning in different time periods, the attractiveness of the facilities, and 

the population changes along the time, are addressed by a methodology encompassing 

economic, operational, and social objectives. Given the spatial nature 

of these problems, GIS technology is used; the comparison of the scenarios 

generated is performed in the objectives’ space along with adequate graphical 

representations. An application illustrates the approach. 

3 - Automated Schematic Maps: Point Positioning Enhancement 

Strategies 

João Mourinho, Instituto de Engenharia Mecânica e Gestão 

Industrial, Faculdade de Engenharia da Universidade do Porto, 

Campus da FEUP, Rua Dr. Roberto Frias, 400, 4200-465 Porto, 

4200-465, Porto, joaomourinho@gmail.com, Teresa Galvão 

Dias, João Cunha 

Schematic Maps are simplified maps used mainly for depicting transportation 

maps. The schematization process has a multiobjective combinatorial nature 

and involves the simplification of geographic information, such as stops and 

lines. We propose an algorithm for generating enhanced solutions for the initial 

stage of the schematization process, related to the alignment of transportation 

stops and lines to a grid. The algorithm merges knowledge from the fields 

of the human cognitive psychology, mathematics and computing. Preliminary 

results on real word instances are presented and discussed. 

4 - A Multi-objective Approach for High Quality Virtual 

backbones in Mobile Ad-Hoc Networks 

Pascal Bouvry, Univ. of Luxembourg, 5600, Luxembourg, 

Luxembourg, pascal.bouvry@uni.lu, Apivadee Piyatumrong, 

Frederic Guinand, Kittichai Lavangnananda 

Topology management for mobile ad-hoc networks is usually based on the notion 

of virtual backbones. The notion of quality of such backbones as a multiobjective 

problem addressing quality of the mobiles composing the backbone 

and of the connection links. Th approach using extensive simulation based on 

different mobility models and using backbones composed of spanning forests. 

We provide ways to address the global behavior of the algorithms by fine-tuning 

the greedy local decision policy and compare it to the solution quality that could 

be achieve using global knowledge and metaheuristics. 

� MB-24 


6.2.50 

Bioinformatics II 


Medicine 





Chair: Piotr Lukasiak, Institute of Computing Science, Poznan 

University of Technology, ul.Piotrowo 2, 60-965, Poznan, Poland, 

Piotr.Lukasiak@cs.put.poznan.pl 

1 - Insights into Protein-RNA complex modeling on example 

of Dicer-Like proteins. 

34 

Maciej Milostan, Institute of Computing Science, Poznan 

University of Technology, ul. Piotrowo 2, 60-965, Poznan, 

Poland, maciej.milostan@cs.put.poznan.pl, Piotr Lukasiak, 


Dicer is an enzyme responsible for processing double-stranded RNAs and plays 

a key role in an RNAi mechanism. Only one crystal structure of dicer is available 

at the moment and it is structure of eukaryotic Dicer from Giardia intestinalis. 

Dicer-Like proteins are also present in many higher eukaryotes including 

Arabidopsis thaliana. We would like to present results obtained by application 

of homology modeling, based on combination of multiple sequence and structural 

alignments, followed by Molecular-Dynamics (MD) simulations, Normal 

Mode Analysis (NMA) and docking experiments. 

2 - Generation of enriched gene sets from independent 

component analysis sources: an alternative for the 

identification of biologically relevant molecular signatures 

Adrien Six, CNRS UMR 7211 - "Integrative Immunology" I2D3 

team, University Pierre et Marie Curie, Bâtiment Cervi (Sce 

Biothérapies), 83, bd de l’Hopital, 75013, Paris, France, 

adrien.six@upmc.fr, Hang Phuong Pham, Nicolas Dérian, David 

Klatzmann 

Transcriptome analyses often yield no gene signatures, due to no or too many 

differentially expressed genes. However, the underlying data structure may 

bear such information at a finer level of organization. We propose an original 

combination of Independent Component Analysis followed by Gene Set 

Enrichment Analysis to identify state-specific signatures that could not be deduced 

from direct gene expression analysis. With this strategy, we identify 

signatures for discrete cell subsets in whole spleen datasets and characterize 

vaccination vector-specific signatures. 

3 - Assessment of protein structure models by systematic 

evaluation of local sequence-structure compatibility 

Krzysztof Fidelis, UC Davis, 11111, Davis, United States, 

kfidelis@ucdavis.edu, Piotr Lukasiak, Maciej Antczak, Wojciech 

Biniecki 

Sequence — structure compatibility in proteins has been a subject of investigation 

ever since Christian Anfinsen’s seminal work on the structure of ribonuclease 

(1954). Here we address the local structure version of the problem, with 

the immediate application to the evaluation of protein structure models. Model 

quality assessments (MQAs) are important as currently computational models 

outnumber experimentally derived structures by more than 100:1. We use the 

local descriptor of protein structure (LDPS) formalism and the corresponding 

shape libraries derived from known structures to assess the compatibility of any 

sequence with a known shape. This is undertaken with a learning approach, a 

support vector machine (SVM). Putative sequences and corresponding local 

structures are evaluated with several types of discriminatory functions (potentials), 

and SVM is trained to distinguish between these and native-like structures. 

We evaluate model quality by stringing local structure scores along the 

entire polypeptide chain of the model. This approach is particularly suited to 

assess the target sequence to template alignments in comparative modeling, a 

major source of modeling errors. 

4 - Combinatorial Optimization to Predict Protein Structure, 

Function and Evolutionary Relationships 

Susanne Pape, Mathematics, TU Darmstadt, Dolivostraße 15, 

Darmstadt, Germany, pape@mathematik.tu-darmstadt.de, 

Alexander Martin, Sebastian Pokutta 

During the last decades, continuing advances in molecular bioinformatics (e.g. 

the Human Genome Project) have led to increased information about protein 

sequences. Predicting structure and function of proteins from these sequences 

is one of the most difficult problems in molecular biology. Many approaches 

dealing with protein sequences and structures are based on combinatorial optimization. 

Here, we present some novel strategies for minimizing a free energy 

function or aligning multiple sequences, that are both important tools in the 

analysis of protein relationships and structures. 

� MB-25 


6.2.48 

Financial Modelling and Risk Management 



Chair: Omer Onalan, Faculty of Business Administration and 

Economics, Marmara University, Anaduluhisar, 34810, Istanbul, 

zittinonalan@gmail.com

1 - A multivariate approach to risk-return management 

within financial crisis environment 

Marie Theres Gregus, Faculty of Economics, University of Split, 

Matice hrvatske 31, Makarska 4, 21 000, Split, Dalmatia, 

Croatia, mariagregus@yahoo.com, Elza Jurun 

This paper is focused on multivariate risk-return management assuming timevarying 

estimation. The time-varying portfolio stocks are weighted by constraints 

on risk measure. Using assumption of bivariate Student‘s t-distribution, 

in multivariate GARCH(p,q) models, it becomes possible to forecast timevarying 

portfolio risk much more precisely. The complete procedure of analysis 

is established using real observed data by Zagreb Stock Exchange. It would be 

emphasize that the financial crisis environment makes the analysis challenging 

in methodological and practical sense. 

2 - Hedging uncertainty in software development — A real 

option modeling approach 

Emil Numminen, School of Management, Blekinge Institute of 

Technology, SE-371 79, KARLSKRONA, Sweden, 

emil.numminen@bth.se 

This paper shows how real option theory is useful to manage uncertain-ties in 

software development projects. The paper show how the option to defer, the 

option to scale up and the option to abandon create flexibility and add value 

while decreasing the project uncertainty. The numerical modeling in this paper 

is done in a two-period discrete time setting using martingale measures to 

enforce risk-neutral expectations. The cash flow uncertainty is modeled via 

a multiplicative binomial process. The analysis is concluded with an empirical 

validation of the circumstances under which these options can be created 

in most software development projects. The paper ends with suggestions for 

further research. 

3 - Pricing of foreign exchange options under the heston 

stochastic volatility model and the cir interest rates 

Rehez Ahlip, School of Computing and Mathematics, University 

of Western Sydney, University of Western Sydney„ 1797, 

Parramatta Campus, NSW, Australia, r.ahlip@uws.edu.au 

Foreign exchange options are studied in the Heston stochastic volatility model 

for the exchange rate combined with the Cox, Ingersoll and Ross dynamics for 

the domestic and foreign stochastic interest rates. The instantaneous volatility 

is correlated with the dynamics of the exchange rate return, whereas the domestic 

and foreign short-term rates are assumed to be independent of the dynamics 

of the exchange rate. The main result furnishes a semi-analytical formula for 

the price of the foreign exchange European call option. 

4 - The Use of Survival Analysis Techniques in studying a 

Bank’s Attractiveness 

Katerina Dimaki, Statistics, Athens University of Economics & 

Business, 76 Patission Street, 10434, Athens, Greece, 

dimaki@aueb.gr, Vassilis Angelis, Maria Mavri 

Customers choose a bank on the basis of its attractiveness as expressed through 

what we have defined as the bank’s Image. This paper uses Survival Analysis 

to monitor a bank’s attractiveness over time. In this sense banks are considered 

as patients whose health status is given by the values of their Image whereas 

treatment is defined as the set of actions taken in order to improve their Image. 

A bank is considered as a survivor as long as its Image follows an increasing 

trend or remains constant. The moment of change in trend direction from 

increasing to decreasing indicates its failure. 

� MB-26 


3.1.11 

Bioinformatics Applications of Machine 

Learning 



Chair: Burcu Gungor, Computer Engineering, Bahcesehir University, 

34353, Istanbul, Turkey, burcu.gungor@bahcesehir.edu.tr 

Chair: Fadime Uney-Yuksektepe, Industrial Engineering, Istanbul 

Kultur University, E5 Karayolu Londra Asfalti Uzeri, Atakoy 

Kampusu, 34156, Istanbul, Turkey, f.yuksektepe@iku.edu.tr 


1 - Application of advanced machine learning methods for 

tag snp selection in complex disease association studies 

Gürkan Üstünkar, Information Systems, Middle East Technical 

University, METU Informatics Institute, Ankara, Turkey, 

ustunkar@gmail.com, Sureyya Ozogur-Akyuz, 

Gerhard-Wilhelm Weber, Nazife Baykal 

In this study we propose two novel and efficient Tag SNP selection algorithms. 

Our proposed algorithms are motivated by prediction of Tagged SNPs and Pairwise 

Association among SNPs. We also introduce Infinite Kernel Learning 

based binary classification scheme that can be used for SNP-Complex Disease 

Association studies. We evaluate the performance of our proposed models 

compared to previous studies in the literature that used the data sets we use in 

our study (Crohn’s disease, Autoimmune disorder and Tick-borne encephalitis). 

We get competitive results within much more reasonable run times. 

2 - Bioinformatics Approaches to Associate Single Nucleotide 

Polymorphisms with Human Complex Diseases 

According to Their Pathway Related Context 

Burcu Gungor, Computer Engineering, Bahcesehir University, 

34353, Istanbul, Turkey, burcu.gungor@bahcesehir.edu.tr, Ceyda 

Sol, Ugur Sezerman 

Genome-wide association studies (GWAS) with millions of single nucleotide 

polymorphisms (SNPs) are popular strategies to reveal the genetic basis of human 

complex diseases. In this study, we use GWAS data provided by WTCCC 

for seven complex diseases and present how to derive SNP combinations associated 

with a specific disease, utilizing gene networks, protein protein interaction 

networks and pathway classification tools. With the whole-genome 

sequencing on the horizon, we show that the full potential of GWAS can only 

be achieved by integrating pathway-oriented analysis. 

� MB-27 


8.2.06 




Chair: Roman Muraviev, Mathematics, ETH Zurich, 

Hungerbergstrasse 5, 8046, Zurich, Switzerland, 

roman.muraviev@math.ethz.ch 

1 - Optimal Control of Financial Derivatives by Chaotic Dynamics 

Giacomo Patrizi, Statistica, Probabilita’ e Statistiche Applicate, 

La Sapienza, University of Rome, Piazza Aldo Moro 5, 00185, 

Rome, g.patrizi@caspur.it, Laura Di Giacomo 

A ’certainty equivalent’ optimal control policy for a Dynamic financial Derivatives 

system based on chaotic theory will be formulated. The representation 

of stochastic processes as chaotic data improves the precision of the control of 

derivatives policies. 

Theory and experimental results will be presented 

2 - Dynamic management of multiple real options 

Michi Nishihara, Graduate School of Economics, Osaka 

University, 1-7, Machikaneyama, Toyonaka, Osaka, 560-0043, 

Osaka, Japan, nishihara@econ.osaka-u.ac.jp 

We develop a model for management of multiple real options. Simultaneous 

exercise of options has positive synergy effects such as economies of scale and 

scope, while separate exercise of options benefits from project flexibility. We 

consider two styles of management; static and dynamic management. A firm 

under static management determines whether real options are exercised simultaneously 

or separately ex ante. In contrast, a firm under dynamic management 

makes the decision at the time of exercise. We reveal characteristics of the 

policies and demonstrate differences among them. 

3 - Utility Maximization with Additive Habits: Optimal Consumption 

and Wealth 

Roman Muraviev, Mathematics, ETH Zurich, Hungerbergstrasse 

5, 8046, Zurich, Switzerland, roman.muraviev@math.ethz.ch 

35


Utility function with additive habits assigned to a financial agent, takes into 

consideration her addiction to certain levels of past consumption, and constitutes 

a significant neo-classical preference device, that enables to measure appropriately 

the satisfaction from consumption. Despite of its extensive recent 

study, almost nothing is known about the structure of the optimal consumption 

paths in the setting of incomplete markets and random endowments. We 

derive a recursive procedure for solving this utility maximization problem and 

uncover various economical features of the optimal consumption stream, such 

as monotonicity, concavity and asymptotics for large levels of wealth. 

� MB-28 



Applications of stochastic programming to 

the energy sector - electricity 



Chair: Maria Teresa Vespucci, Department of Information 

Technology and Mathematical Methods, University of Bergamo, via 

Marconi, 5a, 24044, Dalmine (BG), Italy, mtvespucci@tin.it 

1 - A multi-stage stochastic electricity portfolio model with 

forward contracts 

Rosella Giacometti, Mathematics and Statistics, University of 

Bergamo, via dei Caniana 2, 24127, Bergamo, Italy, 

rosella.giacometti@unibg.it, Maria Teresa Vespucci, Marida 

Bertocchi 

In this paper we develop a stochastic multi-stage portfolio model for a hydropower 

producer operating in a competitive electricity market. The portfolio 

includes its own production and a set of forward contracts for future delivery or 

purchase of electricity to hedge against risks. Our findings show that the use of 

forward contracts for hedging purposes results in a risk reduction and in a more 

efficient use of the hydroplant, taking advantage of the possibility of pumping 

water and ending up with a higher final value of the reservoir. 

2 - Measures of performance of a stochastic model for the 

daily coordination of wind power plants and pumped 

storage hydro plants. 

Maria Teresa Vespucci, Department of Information Technology 

and Mathematical Methods, University of Bergamo, via 

Marconi, 5a, 24044, Dalmine (BG), Italy, mtvespucci@tin.it, 

Asgeir Tomasgard, Marida Bertocchi, Mario Innorta 

The integration in a power generation system of wind power plants and hydro 

plants with pumped storage allows to efficiently manage the uncertainty 

of wind power generation, as wind power in excess can be used for storing 

potential energy. A stochastic programming model is developed for the daily 

optimal coordination of this generation system. Approaches based on ARIMA 

models and on quantile regressions are used for constructing the scenario trees 

that represent the uncertainty of wind power production. Ex-ante and ex-post 

measures are used to evaluate the performance of the stochastic model. 

3 - Optimal day-ahead bidding strategy with futures and bilateral 

contracts. Scenario generation through factor 

models. 

Cristina Corchero, Estadistica i Investigacio Operativa, 

Universitat Politecnica Catalunya, c/Jordi Girona 31, Campus 

nord, Ed. C5, dspt 224, 08034, Barcelona, Spain, 

cristina.corchero@upc.edu, F.-Javier Heredia, M. Pilar Muñoz 

We propose a stochastic programming model that gives the optimal bidding, 

bilateral (BC) and futures contracts (FC) nomination strategy for a price-taker 

generation company in the MIBEL. The objective of the study is to decide the 

optimal economic dispatch of the physical FC and BC among the thermal units, 

the optimal bidding at day-ahead market (DAM) abiding by the MIBEL rules 

and the optimal unit commitment that maximizes the expected profits from the 

DAM. For the uncertainty characterization, we apply the methodology of factors 

models to forecast market prices in a short-term horizon. 

4 - Single and multi-settlement approaches to market 

clearing mechanisms under demand uncertainty 

36 

Javad Khazaei, Engineering Science, University of Auckland, 

R215-70 symonds st-, 1010, Auckland, Auckland, New Zealand, 

j.khazaei@auckland.ac.nz, David Young, Golbon Zakeri 

In the presence of uncertainty in generation or demand, there is a need that 

the market clearing mechanism deals with the deviations between the real and 

predicted quantities, and does so optimally. For this purpose, one or two settlement 

markets have been suggested in the literature. In this talk, we will discuss 

a single settlement stochastic programming mechanism and present its characteristics. 

We then compare this with some currently existing two settlement 

markets. The criterion is the social welfare obtained from these mechanisms at 

the equilibrium where we assume having rational players. 

� MB-29 


8.2.11 

Global Optimization of Graph Partitioning 

Stream: Boolean Programming 


Chair: Frauke Liers, Institut fuer Informatik, Universitaet zu Koeln, 

Pohligstrasse 1, 50969, Koeln, Germany, 

liers@informatik.uni-koeln.de 

1 - An Effective Exact Algorithm for k-way Graph Partitioning 

Frauke Liers, Institut fuer Informatik, Universitaet zu Koeln, 

Pohligstrasse 1, 50969, Koeln, Germany, 

liers@informatik.uni-koeln.de, Miguel Anjos, Bissan Ghaddar, 

Angelika Wiegele 

We consider the task of partitioning the nodes of a graph into k subsets such 

that the sum of weights of the edges joining nodes in different partitions is 

maximum. Applications for this NP-hard problem arise, e.g., in VLSI design 

and in theoretical physics. Based on the work of Ghaddar et al. (Annals of OR, 

to appear), we present an SDP-based branch-and-cut approach for its solution. 

We show that suitably combining polyhedral insights with a bundle method 

for solving the positive semidefinite relaxations leads to an effective solution 

algorithm. 

2 - Solving Non-convex Binary Quadratic Problems via 

Second-order Cone Programming 

Bissan Ghaddar, Management Sciences, University of Waterloo, 

University of Waterloo, 200 University Avenue West, N2L 3G1, 

Waterloo, Ontario, Canada, bghaddar@uwaterloo.ca, Miguel 

Anjos, Juan C. Vera 

We present a general framework based on polynomial programming for solving 

non-convex binary quadratic problems. Using this framework, we can re-derive 

previous relaxation schemes and provide new ones. In particular, we propose 

a second-order cone-based relaxation that can be strengthened by cut generation 

schemes. Computational results on the max-cut problem and the quadratic 

knapsack problem show that our second-order cone based-relaxation with valid 

inequalities is competitive in terms of bounds and time. 

3 - Integer Linear Graph Partitioning 

Matthias Peinhardt, Faculty of Mathematics (IMO), University 

Magdeburg, Universitätsplatz 2, 39106, Magdeburg, Germany, 

matthias.peinhardt@ovgu.de 

We consider the problem of partitioning the vertices of a graph into at most k 

clusters, such that the weight of the edges in the multicut is maximized. We 

explore the capabilities of a linear integer programming formulation, with particular 

emphasis on the exploitation of the sparsity of graphs. The chosen formulation 

requires to deal with symmetrical solutions, which we approach with 

several different techniques. Both theoretical and computational results using 

this formulation are presented. 

4 - On Handling Cutting Planes in Interior-Point Methods 

for Solving Semidefinite Relaxations of Binary 

Quadratic Optimization Problems 

Alexander Engau, Mathematical and Statistical Sciences, 

University of Colorado Denver, Campus Box 170, PO Box 

173364, 80217-3364, Denver, CO, United States, 

aengau@alumni.clemson.edu, Miguel Anjos, Anthony Vannelli

We describe an improved technique for handling large numbers of cutting 

planes when using an interior-point method for the solution of linear and 

semidefinite programming relaxations of combinatorial optimization problems. 

The approach combines an infeasible primal-dual interior-point algorithm with 

a cutting-plane scheme that and does not solve successive relaxations to optimality 

but adds and removes cuts at intermediate iterates based on indicators 

for cut violation and feasibility of the associated slacks. The slack variables of 

added cuts are initialized using a recently proposed interior-point warm-start 

technique that relaxes the interiority condition on the original primal-dual variables 

and enables a restart from the current iterate without additional centering 

or correction steps. Our computational tests on relaxations of the maximumcut 

and single-row facility layout problems demonstrate that this new scheme 

is robust for both unconstrained and constrained binary quadratic problems and 

its performance superior to solving only the final relaxation with all relevant 

cuts known in advance. 

� MB-30 


8.2.13 

MCDA II: Health 


other Issues 





Chair: Mara Airoldi, Management - OR group, London School of 

Economics and Political Science, Houghton Street, WC2A 2AE, 

London, United Kingdom, m.airoldi@lse.ac.uk 

1 - Multiple treatment multi-criteria drug benefit-risk assessment 

Gert van Valkenhoef, Faculty of Economics and Business, 

Universty of Groningen, PO Box 800, 9700 AV, Groningen, 

Netherlands, g.h.m.van.valkenhoef@rug.nl, Bert de Brock, Hans 

Hillege, Douwe Postmus, Tommi Tervonen, Tommi Tervonen 

Pharmaceutical decision making is based on assessing benefits and risks of 

alternative drugs, ideally by considering all available clinical evidence. In 

practice this is not achieved as the assessments are based on meta-analytical 

methods that allow only single criterion pair-wise comparisons. The recently 

proposed Mixed Treatment Comparison (MTC) models enable indirect comparisons 

within a set of alternative treatments. In this presentation we show 

how MTC can be used to construct stochastic multi-criteria models that allow 

to compare multiple treatments in terms of multiple criteria. 

2 - International perspectives on priority setting in health 

care using Multi-Criteria Decision Analysis (MCDA) 

Stuart Peacock, Canadian Centre for Applied Research in Cancer 

Control (ARCC), University of British Columbia, 675 West 10th 

Avenue, V5Z 1L3, Vancouver, BC, Canada, speacock@bccrc.ca 

Priority setting in health care involves managers and doctors making decisions 

about which drugs and health technologies to fund from their limited budgets. 

This paper will review the burgeoning use of Multi-Criteria Decision Analysis 

(MCDA) within the Program Budgeting and Marginal Analysis priority setting 

framework. Particular attention will be given to international consistencies and 

controversies in: the decision criteria used in priority setting; the methods used 

to elicit criteria weights; and, the type of aggregation rule used to combine 

criteria scores. 

3 - Design of Appointment Systems for Outpatient Clinics 

with Scheduled and Unscheduled Arrivals 

Maartje Zonderland, Stochastic Operations Research, University 

of Twente, Citadel H-127, Postbox 217, 7500AE, Enschede, 

Netherlands, m.e.zonderland@utwente.nl, Nikky Kortbeek, 

Richard Boucherie, Nelly Litvak 

Outpatient clinics and diagnostic testing facilities traditionally provide patients 

with individual appointments to balance workload. Disadvantages however, include 

patients needing to revisit the hospital, an involved planning process and 

potentially long access times. This study explores the viability of various walkin 

based policies. We present a stochastic method that finds the mixed strategy 

that optimally balances the benefits and drawbacks of pure appointment and 

walk-in policies. 


� MB-31 


8.2.15 

Societal Complexity, City and Economy 



Chair: Cathal Brugha, Management Information Systems, University 

College Dublin, Quinn School of Business, Belfield, 4, Dublin 4, 

Ireland, Cathal.Brugha@ucd.ie 

1 - The identity of a city 

Zbigniew Kesek, Architecture, University of Technology, 

Podchorazych1, Krakow, Poland, zbigniewkesek@interia.pl 

This paper will be concerned with the event considered as a creation of new 

opportunities for experiencing urban space; with the transition from the concept 

of a city as a static entity to the vision of a city as a kinetic entity. The 

main shift in thinking of town planners and architects, inspired by the theories 

of physicists and philosophers, consists in the belief, that apart from the flow 

of people and images, also architecture is a kinetic component of the city; that 

it is the event — and not form or style — that is the basis for thinking about 

architecture. 

2 - A multivariate study of social change in Canterbury 

Aijie Xie, Kent Business School, University of Kent, F4/G, The 

Pavilion, Woolf College, Giles Lane, CT2 7BQ, Canterbury, 

Kent, United Kingdom, ax3@kent.ac.uk, Cecilio Mar Molinero 

In 1979, a new government in the UK decided to allow tenants of council 

houses to buy their properties. Many took up mortgages and bought their 

houses, only to be caught by increased interest rates, and were forced to sell 

the houses they had just bought. This resulted in many social changes in urban 

areas. In this research we trace the changes that took place in Canterbury using 

the 1981 Census (before the changes), the 1991 Census (during the changes), 

and the 2001 Census (after the changes). We use three-way scaling techniques 

combined with regression analysis to deal with a large and complex data set. 

3 - Why Bubble Economy Occur and Bubble Economy Destroy 

Eizo Kinoshita, urban sicience, Meijo University, 

4-3-3,nijigaoka, 509-0261, kani, gifu, Japan, 

kinoshit@urban.meijo-u.ac.jp 

This paper shows that there are two different phases in economics. These are 

the primal and dual problems. In the primal problem phase, capital expenditures 

of private corporations grow, creating an impetus towards the maximization 

of profits (the spirit of capitalism, as explained by Max Weber). In this 

case, as Adam Smith once wrote, the "invisible hand of God’ works to lead the 

economy to a significant growth. This paper defines the concept of Economic 

Growth, Bubble Economy and Destruction of Bubble Economy. And this paper 

describes that why bubble Economy occur and Bubble Economy Destroy. In 

the process, the paper shows that Primal Economy exists before Bubble Economy 

and Dual Economy exist after Destruction of Bubble Economy. 

4 - The Actors of the Credit Crisis reflected by the Compram 

Methodology 

Dorien DeTombe, Methodology of Societal Complexity, Chair 

Euro Working Group, P.O.Box 3286, 1001 AB , Amsterdam, 

Netherlands, detombe@nosmo.nl 

The credit crisis is a complex societal problem in which many phenomena and 

actors are involved. If one wants to analyze the causes of this problem, try to 

stabilize the situation and prevent new fall backs, a multi-disciplinary approach 

is prescribed. A careful analysis based on the scientific methodology of societal 

complexity is needed in order to find how the credit crisis happened and how 

new crises can be prevented. Theories of multiple disciplines must be used by a 

multi disciplinary team to analyze the situation and to find sustainable options. 

This process can be accomplished by following the Compram Methodology 

of DeTombe. The Compram Methodology provides a framework for policy 

making which includes many methods and tools. The Compram Methodology 

is specialized to handle complex interdisciplinary world-wide problems and to 

offer a step-by-step approach of analyzing the problem, finding and implementing 

sustainable interventions and evaluating the effects. The Compram methodology 

offers a bird’s-eye view on the complexity of the problem and gives direction 

to policy makers to build their decisions on using a multi-disciplinary, 

multi-actor approach. In this article the credit crisis is discussed in relation with 

the Compram Methodology. Aspects of the credit crisis are described with an 

emphasis on the role of the actors. Based on the Compram Methodology direction 

can be provided for handling the credit crisis and avoiding future similar 

problems. 

37


� MB-32 


8.2.17 

OR in Forestry I 



Chair: Andrés Weintraub, University of Chile, Santiago, Chile, 

aweintra@dii.uchile.cl 

1 - Optimal weights in route distance generation 

Mikael Rönnqvist, Department of Finance and Management 

Science, Norwegian School of Economics and Business 

Administration, NO-5045 , Bergen, Norway, 

mikael.ronnqvist@nhh.no, Patrik Flisberg, Bertil Lidén 

Contracts between transporters and forest companies are often based on the 

driven distance. However, this distance can often be hard to agree on because 

of many attributes e.g. length, quality, width, speed limits, etc. We present 

an approach to find a set of optimal weights for more than 30 attributes. An 

important part is a set of detailed "Solution routes" where the forest companies 

and transporters have agreed. Then an optimization model is formulated where 

the main variables are the weights on each attribute. Results and experiences 

from an industrial implementation are reported. 

2 - Dynamic log yard designs for an improved coordination 

of sawmill and log yard operations 

Luc LeBel, Sciences du bois et de la foret, université laval, 2405 

de la terrrase, g1v 0a6, Quebec, Quebec, Canada, 

luc.lebel@sbf.ulaval.ca, Daniel Beaudoin 

Sawmills have traditionally kept high level of inventory in their yards for seasonal 

considerations. We attempt to find the optimal assortments to store 

through a Forward-Reserve Problem (FRP). The classical FRP was extended 

to a multi-period FRP in the context of a divergent process industry. The multiperiod 

context allows for changes in assignments to the forward area as market 

conditions change. In order to account for the divergent nature of the industry, 

the FRP formulation has been extended to anticipate production at the mill 

based on known demands and market anticipation functions. 

3 - A solution of the cable corridor layout problem for harvesting 

under steep slope conditions 

Leo Bont, Institute of Terrestrials Ecosystems, CH-8092 , Zürich, 

leo.bont@env.ethz.ch, Hans Rudorf Heinimann, Richard Church 

The spatial layout of forest harvesting systems is a task which highly affects 

operational efficiency. Under steep slope conditions cable systems have been 

used for timber extraction, being laid out by rules of thumb, especially in Europe. 

Here we introduce a modelling approach for the cable corridor layout 

problem, assuming that the road network is given. We formulated the problem 

as a set cover mixed integer programming problem. Our approach will be 

useful for assessing layouts for different cable systems and to investigate the 

effectivity of road network layout for cable systems. 

4 - Mixed Integer Programming Models to Evaluate Integrating 

Strategies for Value Chain Management: a Case 

Study of the Chilean Forest Industry 

Andrés Weintraub, University of Chile, Santiago, Chile, 

aweintra@dii.uchile.cl, Juan José Troncoso, Sophie D’Amours, 

Mikael Rönnqvist 

We present a mixed integer programming model to evaluate two different integration 

strategies in order to show the impacts of a fully demand driven integration 

of the value chain in the forest industry. To illustrate our thoughts, we use 

forest, economic and production information from a Chilean forest company. 

We compare two different integration strategies: the first one where the forest 

and the industry planning are decoupled and the second, were all parts of the 

value chain (forest, transportation, mills) are driven by final product demand. 

38 

� MB-33 


8.2.19 

Energy Policy and Planning 



Chair: Sandrina Pereira, IDMEC-IST, Avenida Rovisco Pais, Pav. 

Mecânica, 2 o andar, 1049-001, Lisboa, Portugal, 

sandrinapereira@ist.utl.pt 

1 - Decision Support System for Low Carbon Regions 

Sandrina Pereira, IDMEC-IST, Avenida Rovisco Pais, Pav. 

Mecânica, 2 o andar, 1049-001, Lisboa, Portugal, 

sandrinapereira@ist.utl.pt, Anildo Costa 

This paper analyse the usefulness of Scenario Building and Analysis - SBA and 

simulation models such as Agent Based Models — ABM, as methods to develop 

energy policies and pathways towards Low Carbon Region - LCR. SBA 

can illustrate the influence of technology in GHG reduction but technology evolution 

by itself will not be enough to achieve LCR. Also, energy consumption is 

influenced by consumer’s behaviour and social values which are better captured 

by ABM. A methodology to achieve LCR based on SBA (to assess technology) 

and ABM (to assess influence of society behaviour) will be described 

2 - Energy Indicators for Sustainable Development 

Athanasios Angelis-Dimakis, Chemical Engineering, National 

Technical University of Athens, Heroon Polytexneiou 9, 

Zografou Campus, 15780, Athens, Greece, 

angelis@chemeng.ntua.gr, George Arampatzis 

Energy planning for sustainable development at regional and national level is a 

basic priority for every country. The scope of this paper is to develop a set of 

indicators in order to assess the economic, social and environmental aspects of 

an energy system’s sustainable development. Those indicators give an overall 

picture of a country’s energy system. Furthermore, as they fluctuate over time 

they will be good markers of progress and underlying changes and will guide 

decision-making on investments in energy. The set of indicators will be applied 

to the case of the Greek Energy System. 

3 - An n-K Contingency-Constrained Unit Commitment 

Model via Robust Optimization 

Fabrício Oliveira, Industrial Engineering, PUC-Rio, Av. Marquês 

de S. Vicente s/n, Gávea, 22453-900, Rio de Janeiro, Rio de 

Janeiro, Brazil, fabricio.carlos.oliveira@gmail.com, Alexandre 

Street, José Manuel Arroyo 

This paper presents a new approach for the contingency-constrained unit commitment 

problem. The model incorporates an n-K security criterion by which 

power balance is guaranteed under any contingency state comprising the simultaneous 

loss of up to K units. Instead of considering all possible contingency 

scenarios, which would render the problem intractable, a novel scenario-free 

formulation based on robust optimization is proposed. Unlike scenario-based 

approaches, the robust model does not depend on the size of the set of contingencies, 

thus providing a computationally efficient framework. 

� MB-34 


8.2.23 

Solution Approaches for Lot-sizing 

Problems II 


Quantity 


Chair: Bernardo Almada-Lobo, Industrial Engineering and 

Management, Faculty of Engineering of Porto University, Rua Dr. 

Roberto Frias s/n DEIG, 4200-465, Porto, Portugal, 

almada.lobo@fe.up.pt 

1 - An Efficient Computational Method for Non-Stationary 

(R,S) Inventory Policy with Service Level Constraints 

Mustafa Dogru, Bell Labs, Alcatel-Lucent, 

BLANCHARDSTOWN INDUSTRIAL PARK, 15, DUBLIN,

Ireland, mustafa.dogru@alcatel-lucent.com, Armagan Tarim, 

Ulas Ozen, Roberto Rossi 

This paper provides an efficient computational approach to solve a mixed integer 

programming (MIP) model developed for calculating the parameters of an 

(R,S) policy in a finite horizon with non-stationary stochastic demand and service 

level constraints. Given the replenishment periods, we characterize the optimal 

order-up-to levels for the MIP model and use it to guide the development 

of a relaxed MIP model, which can be solved in polynomial time. Extensive 

numerical tests show that our method dominates the MIP solution approach and 

can handle real-life size problems in trivial time. 

2 - Multi-Product Single-Stage Lot Sizing with Random 

Yield, Imperfect Inspection, Process Compressibility, 

and Partial Backordering 

D. Moradinezhad, industrial engineering, iran university of 

science and technology, no. 9, Zamani Alley, Nejatollahi Ave, 

Karimkhan St„ Tehran, Tehran, Iran, Islamic Republic Of, 

dmcatalonia@gmail.com, M.b. Aryanezhad, E. Noorollahi, M. 

Karimi-Nasab, Seyed Mohammad Ghoreyshi 

Production planning with real world assumptions is a complicated issue for 

production managers. This paper develops an economic production quantity 

model for a multi-product single-stage production system under several working 

conditions / assumptions. The proposed model is proposed to determine 

decision variables such as cycle length, batch size of each product type for every 

production run, backorder, regular production rate, and rework rate simultaneously. 

Computational experiences induce that both model and the algorithm 

have high performance even in the worst cases. 

3 - Heuristic approaches for a practical lot splitting and 

scheduling problem 

Carina Pimentel, Departamento de Produção e Sistemas, 

Universidade do Minho, Campus de Gualtar, Portugal, 4710-057, 

Braga, Braga, Portugal, carina@dps.uminho.pt, Filipe Alvelos, 

António Duarte, J. M. Valério de Carvalho 

In this talk we present some heuristic approaches for a real world lot splitting 

and scheduling problem of a Textile factory. The problem consists of finding 

a weekly production plan for the knitting section of the company, establishing 

the quantities to produce of each component (organized in one or several lots), 

and where and when (starting/completion times) to produce them. Two important 

objectives to achieve are on time delivery of products and minimum levels 

of work-in-process inventory. We present some results for a set of randomly 

generated instances based on real world data. 

4 - Multi-plant, multi-period and multi-item capacitated lotsizing 

problem in beverage industry 

Luis Guimarães, DEMEGI, FEUP, Portugal, 

luis.santos.guimaraes@gmail.com, Bernardo Almada-Lobo 

In some process industries joint lotsizing and scheduling is known to be vital in 

order to achieve feasible and effective production plans. Moreover, when facing 

a multi-plant scenario coordination between plants can make a substantial 

difference towards more efficient planning. Inspired by a real case-study in the 

beverage industry we develop a novel formulation and a new heuristic for the 

multi-plant, multi-period and multi-item capacitated lotsizing problem where 

transfers between plants are allowed and sequence dependent setup times and 

costs are considered in a rolling horizon approach. 

� MB-35 


6.2.46 

Soft OR and Problem Structuring Methods I 



Chair: Colin Eden, Management, University of Strathclyde, 199 

Cathedral Street, G4 0QU, Glasgow, United Kingdom, 

colin@gsb.strath.ac.uk 

1 - Enabling Politically Feasible Agreements: what is a 

group? 

Colin Eden, Management, University of Strathclyde, 199 


colin@gsb.strath.ac.uk, Paul Nutt, Fran Ackermann 


Involving many group participants meets, (i) the need to tap into a range of 

different areas of expertise and perspectives, and (ii) the importance of gaining 

ownership for the outcome(s) among all those who can influence the implementation 

of agreements. This presentation discusses these issues. 

2 - Problem structuring for investing in an uncertain future 

Colin Eden, Management, University of Strathclyde, 199 


colin@gsb.strath.ac.uk 

Developing an appropriate problem structure for public policy issues that have 

very long term consequences (20-30yrs) has particular challenges. This presentation 

discusses the issues and possibilities in undertaking such a task. 

3 - Using the Theory of Constraints to resolve conflicts in 

a large public hospital 

Vicky Mabin, Victoria Management School, Victoria University 

of Wellington, PO Box 600, Wellington, 6140, Wellington, New 

Zealand, vicky.mabin@vuw.ac.nz, Sally Babington, Vanessa 

Caldwell, Julie Yee 

Hospitals typically struggle to meet demand for services within limited budgets. 

Significant and long-standing tensions exist. This paper provides an 

analysis of a large public hospital using the Theory of Constraints, to build 

an understanding of the many cause and effect and necessity relationships that 

exist in such complex organisational settings. In this case, the root cause of the 

long-standing conflicts in the hospital system are mapped, solutions suggested, 

tested and implementation planned. 

4 - Systematic and Comprehensive Supply Chain Environmental 

Management 

Joana M. Comas Marti, TOM, EPFL, Station 5, 1015, Lausanne, 

Switzerland, joana.comas@epfl.ch, Ralf W. Seifert 

Stakeholders are increasingly demanding businesses to manage environmental 

issues taking a supply chain or life cycle approach. Supply chain environmental 

management (SCEM) has arisen as a result. We contribute to the field with 

a framework for the systematic and comprehensive definition and assessment 

of SCEM strategies. We structure it in 3 dimensions: what, why and where, 

i.e. the action taken, the environmental impact being addressed with it, and the 

supply chain or life cycle stage where that impact takes place. We apply this 

framework to assess 12 sustainability leaders in 6 sectors. 

� MB-36 


3.1.05 

Fuzzy Goal Programming 1 


Intelligence 


Chair: Mariano Jimenez-Lopez, Economía Aplicada I, University of 

the Basque Country, Plaza de Oñati 1, 20018, San Sebastian, Spain, 

mariano.jimenez@ehu.es 

1 - A Fuzzy Multi-objective Approach for Master Planning 

in Ceramic Tile Supply Chains 

Josefa Mula, Business Management, Polytechnic University of 

Valencia, Escuela Politécnica Superior de Alcoy, Plaza Ferrándiz 

y Carbonell, 2, 03801, Alcoy, Alicante, Afghanistan, 

fmula@cigip.upv.es, David Peidro, Mareva Alemany, Francisco 

Lario Esteban 

We propose a fuzzy multi-objective linear programming (FMOLP) approach 

to model a centralized replenishment, production and distribution problem for 

ceramic tile supply chains. We also present an interactive solution methodology 

to convert this FMOLP model into an auxiliary crisp single-objective 

linear model and to find a preferred compromise solution in an interactive fashion. 

For illustration purposes, an example based on modifications of real-world 

industrial problems is presented. 

39


2 - An Application of Goal Programming in Personnel Selection 

Decisions 

Lourdes Canos, Organizacion de Empresas, Universidad 

Politecnica de Valencia, Crtra. Nazaret-Oliva s/n, 46730, Grao de 

Gandia (Valencia), Spain, loucada@omp.upv.es, Maria J. Canos, 

Vicente Liern 

The candidate selection process for a job can be seen as a decision problem 

in which some competences, valued under uncertainty, have to be optimized. 

When the number of candidates is high, the selection process has two or more 

phases. In the first stages of the process, establishing goals for some attributes 

makes a quicker and more effective election. In this paper we present some 

fuzzy logic based algorithms to solve this problem. We illustrate our approach 

with some numerical examples whose solution has been obtained by using software 

developed by the authors. 

3 - Weight and Parameter Space Exploration in Fuzzy Goal 

Programming 

Dylan Jones, Mathematics, University of Portsmouth, Lion 

Terrace, PO9 3HE, Portsmouth, Hampshire, United Kingdom, 

dylan.jones@port.ac.uk 

Fuzzy goal programming model is characterised by a number of weights and 

parameters that need to be set by the decision maker. This seminar explores 

effective means for setting those weights and parameters. The nature of weight 

space in different goal programming variants is discussed. Means of effectiveness 

exploration and categorisation of weight and parameter space are given. 

Practical means of conducting weight and parameter sensitivity analysis in 

fuzzy goal programming are discussed. 

4 - Ordinary Goal Programming with Fuzzy Aspiration Levels 

Mariano Jimenez-Lopez, Economía Aplicada I, University of the 

Basque Country, Plaza de Oñati 1, 20018, San Sebastian, Spain, 


Fuzzy sets help to express imprecise knowledge in a natural way. I mix ordinary 

goal programming with the representation of aspiration levels by fuzzy sets. I 

look for the minimum distance to these fuzzy sets, instead of, like in the case 

of fuzzy goal programming, looking for the better satisfaction degree. Therefore 

I accept solutions that surpass the tolerance thresholds. With this approach 

the shape of the penalty functions is generated intuitively and the membership 

degree of fuzzy aspiration levels can be used, in an interactive way, in order to 

look for a more equilibrated solution. 

� MB-37 


3.1.09 

Humanitarian Logistics 



Chair: Begoña Vitoriano, Estadística e Investigación Operativa I, Fac. 

CC. Matemáticas, Universidad Complutense de Madrid, Plaza de 

Ciencias, 3, Ciudad Universitaria, 28040, Madrid, Spain, 

bvitoriano@mat.ucm.es 

1 - Rule-based assessment of disaster’s consequences in 

a bipolar possibilistic framework 

J. Tinguaro Rodriguez, Department of Statistics and Operational 

Research, Complutense University of Madrid, Plaza de Ciencias, 

3, 28040, Madrid, Spain, jtrodrig@mat.ucm.es, Javier Montero, 

Begoña Vitoriano 

A decision support system useful to assess consequences of disasters with the 

very first available information just after the strike, usually incomplete and not 

fully reliable, is presented in this work. This DSS is based on a rule-based 

learning model focused on interpretability. The output of the system is a bipolar 

possibility distribution assessing both the evidential support for and the plausibility 

(in the sense of not being excluded) of the available linguistic outputs. 

This enables a more expressive representation of uncertainty and knowledge 

about inferred consequences of disasters. 

40 

2 - Measuring Risk Factors in Humanitarian Aid Distribution 

Begoña Vitoriano, Estadística e Investigación Operativa I, Fac. 

CC. Matemáticas, Universidad Complutense de Madrid, Plaza de 

Ciencias, 3, Ciudad Universitaria, 28040, Madrid, Spain, 

bvitoriano@mat.ucm.es, M. Teresa Ortuno, Gregorio Tirado 

Natural disasters are phenomena that strike countries, which sometimes request 

help from the rest of the world. International organizations respond organizing 

relief operations, which usually have to be developed in an atmosphere of uncertainty 

which risks the operation success. Humanitarian aid distribution is 

usually performed with limited resources to know if it is possible to deliver the 

planned amount, or under uncertainty about the road state, or, sometimes, under 

low security conditions with some risk of being hijacked. A goal programming 

model including these criteria is presented. 

3 - Combining distribution and recovery operations in Humanitarian 

Logistics 

Federico Liberatore, Kent Business School, University of Kent, 

Annexe, CT2 7PE, Canterbury, United Kingdom, 

fl51@kent.ac.uk, Begoña Vitoriano, M. Teresa Ortuno, Gregorio 

Tirado, Maria Paola Scaparra 

The distribution of emergency goods to a population affected by a disaster is 

one of the most fundamental operations in Humanitarian Logistics. In the case 

of a particularly disruptive event, parts of the distribution infrastructure (e.g. 

bridges, roads) can be damaged. This damage would make it impossible and/or 

unsafe for the vehicles to reach all the centers of demand (e.g. towns and villages). 

In this paper, we propose and solve the problem of planning for recovery 

of damaged elements of the distribution network, so that the consequent distribution 

scheme would be optimal. 

4 - Infrastructure and good governance as key elements to 

equity and economic growth in Sub Saharan Africa 

Moses Dowart, Department of Applied Mathematics, National 

University Of Science and Technology (NUST), 8327 unit k seke 

Chitungwiza Harare Zimbabwe, 263, Harare, Zimbabwe, 

Zimbabwe, mdowart@gmail.com 

Africa, a continent endowed with immense natural and human resources as well 

as great cultural, ecological and economic diversity, remained underdeveloped. 

In this paper we analyse infrastructure and good governance as key elements 

to equity and economic growth in sub saharan Africa. We construct a simple 

endogenous growth model with an exhaustible natural resource element, readdressing 

empirically the question of whether natural resources have a positive 

or negative effect on growth. 

Key Words: Economic Growth, Infrastructure, Good Governance, Sub Saharan 

Africa. 

� MB-38 


6.2.44 

Advances in Economical and Financial 

Theory 









1 - The visual exploration and grouping of EU member 

states according to Structural Indicators: a data-driven 

approach 

Jasna Soldic-Aleksic, Department of Statistics, Faculty of 

Economics, Kamenicka 6, 11000, Belgrade, Serbia, Serbia, 

jasnasoldic@eunet.rs

This study applies Kohonen Self-Organizing Map for a visual exploration of 

the convergence of EU member states in the light of Structural Indicators for 

different economic and social domains. Currently the SIs set comprises 79 

indicators. We want to acquire a general picture of the current status of EU 

member states. In this respect we have applied several standard univariate and 

multivariate data analysis techniques. The original SIs were replaced by fewer 

extracted factors. In addition, we produced six separate maps based on each SI 

dataset. 

2 - The correlation risk premium in the recent financial turmoil 

Chrysi Markopoulou, Management Science and Technology, 

Athens University of Economics and Business, 47A Evelidon str, 

Athens, Attica, Greece, cmarkopoulou@aueb.gr, Spyros 

Xanthopoulos, Apostolos Refenes 

Accurate estimation of correlation risk premium is of paramount importance in 

asset allocation and risk management. There is growing literature that discusses 

the pricing of the correlation risk as well as the magnitude of the correlation risk 

premium. Moreover, a number of studies have provided evidence of increasing 

correlation during periods of high volatility. In this paper we examine and evaluate 

the behavior, the properties and the forecasting performance of the implied 

correlation risk premium using data on the Dow Jones Industrial Index options 

during the financial crisis of 2007-2009. 

3 - The definition of customer churn in a non-contractual 

setting 

Susana San Matías, Departamento de Estadística e Investigación 

Operativa Aplicadas y Calidad, Universidad Politécnica de 

Valencia, Edificio 7A, Camino de Vera, s/n, 46022, VALENCIA, 

Spain, ssanmat@eio.upv.es, Mónica Clemente 

Churn is probably one of the most classical questions addressed by the field 

of marketing using data mining tools. The problem arises when a definition of 

churn has to be implemented in a non-contractual setting, because there is not 

a variable into the database that indicates when a customer is leaving. Several 

approaches to this issue can be found in the literature. Which is the most adequate 

definition of churn in such a context? In this paper we present, discuss 

and compare using objective criteria a large set of different definitions and we 

analyze the advisability of their application. 

� MB-39 


6.2.45 

Set-Valued Analysis for Control Problems 



Chair: Robert Baier, Department of Mathematics, University of 

Bayreuth, Chair of Applied Mathematics, D-95440, Bayreuth, 

Germany, robert.baier@uni-bayreuth.de 

1 - Discrete Approximation of Differential Systems with 

Variable Time Impulses 

Robert Baier, Department of Mathematics, University of 


Germany, robert.baier@uni-bayreuth.de, Tzanko Donchev 

Impulsive systems are differential equations or inclusions with jump conditions 

which force a piecewise defined solution to jump, if it hits one of the switching 

surfaces. To prove convergence order 1 for the set-valued Euler discretization, 

the beating phenomena must be avoided. An adapted Filippov-Gronwall lemma 

is the basis of this analysis. 

Several examples with numerical calculations including a set-valued version 

of the bouncing ball problem illustrate the results. Extensions to higher order 

Runge-Kutta methods with an adequate approximation of the jump times are 

discussed. 

2 - Collision avoidance using reachable sets 

Matthias Gerdts, Department of Mathematics, University of 

Wuerzburg, Am Hubland, 97074, Wuerzburg, Germany, 

gerdts@mathematik.uni-wuerzburg.de 


This talk discusses an application from the automotive industry and addresses 

the problem of detecting whether a collision with an obstacle will occur or not. 

An approach towards this problem is to use reachable set information for nonlinear 

control systems. For, if the reachable set of a car in a certain situation is 

known (at least approximately), then it is easy to decide whether a collision will 

occur or can be avoided by appropriate controls. The reachable sets are computed 

by appropriately formulated optimal control problems, which are solved 

by a direct discretization method. 

3 - Semiconcavity of the value function for a class of differential 

inclusions 

Peter Wolenski, Mathematics, Louisiana State University, LSU, 

70803, Baton Rouge, Louisiana, United States, 

wolenski@math.lsu.edu 

We consider the Mayer problem in optimal control in which the dynamics are 

in the form of a differential inclusion. The associated value function is known 

to be semiconcave if the velocity set has a smooth enough parametrization, but 

this is a difficult property to ascertain for general multifunctions. We provide 

new intrinsic sufficient conditions on a multifunction and the endpoint data so 

that the value function is semiconcave. 

4 - Operations on Sets and Approximation of Set-Valued 

Mappings 

Elza Farkhi, School of Math. Sciences, Tel-Aviv University, 

Haim Levanon Str., 69978, Tel Aviv, elza@post.tau.ac.il 

The approximation of set-valued maps appears in numerical control and optimization. 

Approximation operators on such maps are constructed in two ways: 

1. Indirectly, by single-valued representations of sets. Then operators on sets 

are reduced to their representing functions. We discuss examples of representations 

of convex or 1D sets and the corresponding set-valued approximations. 

2. Directly, by operations on sets. Here, the metric sum and average of sets 

are applied to numerical set-valued integration and approximation of special 

trajectories of differential inclusions. 

� MB-40 


6.2.52 

Network design 1 



Chair: Luis Gouveia, DEIO, University of Lisbon, Campo Grande, 

Bloco C6, 1749-016, Lisbon, Portugal, legouveia@fc.ul.pt 

1 - The Generalized Regenerator Location Problem 

Ivana Ljubic, Department of Statistics and Decision Support 


Austria, ivana.ljubic@univie.ac.at, Si Chen, S. Raghavan 

In the generalized regenerator location problem (GRLP), we are given a set S 

of potential regenerator locations and a set T of terminal nodes that want to 

communicate with each other. A signal can only travel a maximum distance 

before its quality deteriorates. Therefore, regenerators need to be installed at 

nodes from S to enable communication between nodes from T at minimum 

cost. We model the problem as the (node-weighted) directed Steiner forest 

problem (NWDSF). We provide several MIP formulations for both, NWDSF 

and GRLP, and compare them theoretically and computationally. 

2 - Benders decomposition for the hop-constrainted survivable 

network design problem 

Quentin Botton, Louvain School of Management - CORE, 

Université catholique de Louvain, place des doyens,1, 1348, 

Louvain-la-Neuve, Belgium, quentin.botton@uclouvain.be, 

Bernard Fortz, Luis Gouveia, Michael Poss 

Given a graph with nonnegative edge weights and a set of pairs of nodes Q, we 

study the problem of constructing a minimum weight set of edges so that the 

induced subgraph contains at least K edge-disjoint paths containing at most L 

edges between each pair in Q. We present the first formulation for the problem 

valid for any K and L. We use a Benders decomposition method to handle the 

big number of variables and constraints. We present a thorough computational 

study of various cutting plane and branch-and-cut algorithms on a large set of 

instances including the real based instances from SNDlib. 

41


3 - Combining Column Generation and Local Search 

Heuristics for the Load Balancing Optimization of 

Telecommunication Networks 

Dorabella Santos, Instituto de Telecomunicações - Pólo de 

Aveiro, Campus de Santiago, Aveiro, 3810-193, Aveiro, 

dorabella@av.it.pt, Amaro de Sousa, Filipe Alvelos 

We address the problem of routing a set of traffic flows over a capacitated 

telecommunications network (we assume single path routing). The aim is to 

optimize the network load balancing using the min-max optimization of link 

loads. We propose the combination of column generation with local search 

based heuristics where the columns given by column generation define the 

search space for the local search. Using the lower bounds of column generation, 

we show through computational results that our proposal is very efficient 

in providing good quality feasible solutions in short computing times. 

4 - A Branch-And-Cut Algorithm for the Partitioning-Hub 

Location-Routing Problem 

Daniele Catanzaro, Computer Science, Université Libre de 

Bruxelles, Boulevard du Triomphe CP210/01, 1050, Brussels, 

Belgium, dacatanz@ulb.ac.be, Martine Labbé, Aykut Ozsoy 

We introduce the Partitioning-Hub-Location-Routing Problem (PHLRP), a 

hub location problem involving graph partitioning and routing features. The 

PHLRP consists of partitioning a given network into sub-networks, locating at 

least one hub in each sub-network and routing the traffic within the network at 

minimum cost. This problem finds applications in deployment of an Internet 

Routing Protocol called Intermediate System - Intermediate System (ISIS), and 

strategic planning of LTL ground freight distribution systems. We describe an 

Integer Programming (IP) formulation for solving the PHLRP and explore possible 

valid inequalities to strengthen it. Computational experiments prove the 

effectiveness of the model and the valid inequalities which allow exact analysis 

for the PHLRP instances containing up to 20 vertices. 

� MB-41 


3.1.06 

Behavioral Models in Revenue Management 



Chair: Tatsiana Levina, School of Business, Queen’s University, 143 

Union str, K7L 3N6, Kingston, Ontario, Canada, 

tlevin@business.queensu.ca 

Chair: Yuri Levin, School of Business, Queen’s University, 143 


ylevin@business.queensu.ca 

1 - Strategic Bidders Club: Implications for Consumer 

Learning 

Tatsiana Levina, School of Business, Queen’s University, 143 


tlevin@business.queensu.ca, Yuri Levin, Jeff McGill, Mikhail 

Nediak 

We develop a model of strategic consumer learning of bidding strategies in a 

market for opaque products, obtain the optimal strategies for the consumers, 

and study their properties. The effects of consumer cooperation on learning are 

examined and compared under two different scenarios: exchange of information 

only and exchange of information with coordinated bidding. 

2 - Optimal Keyword Bidding to Allocate Expenditures 

Across Keywords: A Stochastic Optimization Analysis 

Ozgur Ozluk, College of Business, San Francisco State 

University, 1600 Holloway Avenue, 94132, San Francisco, CA, 

United States, ozgur@sfsu.edu, Susan Cholette, Mahmut Parlar 

Today, keyword advertising on the web pages of search engines is a predominant 

venue for advertising online. Advertisers bid for keywords, where bid 

price determines ad placement, which in turn affects the response function, 

defined as the click-through rate. Advertisers typically have a fixed daily budget 

that should not be exceeded, so an advertiser must allocate the budget as 

productively as possible by selecting which keywords to use and then deciding 

how much to allocate for each keyword. We construct and examine a stochastic 

model for this selection and allocation process. 

42 

3 - Empirical Investigation of Strategic Buyer and Seller 

Behavior in the Wait-or-Buy Game 

Anton Ovchinnikov, Darden School of Business, University of 

Virginia, 100 Darden Blvd, 22903, Charlottsville, VA, United 

States, Ovchinnikova@darden.virginia.edu 

We discuss a laboratory experiment — the wait-or-buy game - in which some 

subjects (business school students) are playing strategic buyers and others are 

playing revenue managers (sellers.) We describe the systematic biases in the 

behaviors of both buyers and sellers, as well as suggest how firms might adjust 

their revenue management strategies in order to capitalize on the buyers’ 

biases and protect themselves against sellers’ biases. We also discuss the implication 

of this empirical analysis on creating more accurate models of strategic 

consumer behavior. 

4 - Selling to Heterogeneous Customers with Uncertain 

Valuations under Returns Policies 

Qian Liu, IELM, HKUST, Hong Kong, qianliu@ust.hk, 

Wenqiang Xiao 

We consider a firm that sells a product with finite capacity to a random market 

size. Each customer’s valuation of the product includes an intrinsic value 

privately known by him before purchase and the ex ante uncertain product fitness 

revealed to him after purchase. In contrast to the established screening 

result with unlimited capacity, we show that offering a menu of return contracts 

simultaneously is not optimal with finite capacity. The firm is better off 

by offering a menu of return contracts sequentially, i.e., offering one at a time. 

� MB-42 


3.1.07 

Mathematical Programs with 

Complementarity Constraints 




Chair: Joaquim Judice, Dept. Mathematics, University of Coimbra, 

Largo D. Dinis, 3000, Coimbra, Portugal, Joaquim.Judice@co.it.pt 

1 - Strong Stability in MPCC 

Vladimir Shikhman, Dept. Mathematics, RWTH Aachen 

University, Templergraben 55, 52056, Aachen, Germany, 

shikhman@mathc.rwth-aachen.de, Hubertus Th. Jongen, Jan-J. 

Rückmann, Sonja Veelken 

We study mathematical programs with complementarity constraints (MPCC). 

Special focus will be on C-stationary points. Under LICQ we characterize 

strong stability of C-stationary points (in the sense of Kojima) by means of first 

and second order information of the defining functions. The strong stability of 

C-stationary points allows a possible degeneracy of bi-active Lagrange multipliers. 

We relate our results to the critical point theory for MPCC discussing 

the notions of a non-degenerate C-stationary point and its C-index. 

2 - On the formulation and solution of a traffic Linear Program 

with Complementarity Constraint 

Isabel Ribeiro, Engenharia Civil (SMF), FEUP, Rua Dr. Roberto 

Frias, s/n, 4200-465, Porto, Portugal, iribeiro@fe.up.pt, Maria 

Lurdes Simões 

A queuing system resulting from a signalized intersection regulated by pretimed 

control in urban traffic network is considered. In this talk, we show how 

Global Optimization and Complementarity can be used to find the optimal cycle 

length and green split allocation for an isolated signalized intersection. The 

model is formulated as a Linear Program with Complementarity Constraints 

(LPCC). A sequential complementarity algorithm for computing a global minimum 

for the LPCC is analysed. Computational experience is included to highlight 

the efficacy of this method for processing this problem. 

3 - The Asymmetric Eigenvalue Complementarity Problem: 

Theory and Algorithms 

Joaquim Judice, Department of Mathematics, University of 

Coimbra, Largo D. Dinis, 3000, Coimbra, Coimbra, Portugal, 

Joaquim.Judice@co.it.pt, Silvério Rosa

Given a real matrix A and a real (symmetric or asymmetric) Positive Definite 

matrix B, the Eigenvalue Complementarity Problem (EiCP) is an extension of 

the well-known Generalized Eigenvalue Problem GEiP(A,B) where the variables 

of the problem are required to be nonnegative and have to satisfy a complementarity 

constraint. In this talk the most difficult case is analysed, where 

at least one of matrices A or B of the EiCP is asymmetric. It is shown that the 

EiCP reduces to a Finite-Dimensional Variational Inequality, to a Mathematical 

Programming Problem with Linear Complementarity Constraints and to a 

Global Optimization Problem. Based on these reductions, necessary and sufficient 

conditions for the existence of a solution to the EiCP are established. A 

projected gap-function method and an enumerative algorithm are introduced for 

finding a solution to the asymmetric EiCP. The computation of several complementary 

eigenvalues and of the maximum and minimum of these eigenvalues is 

also discussed. Computational experience is reported to illustrate the efficiency 

of the algorithms to deal with the asymmetric EiCP. 

4 - Dynamics of equilibrium problems: a hybrid systems 

approach 

Monica-Gabriela Cojocaru, mathematics & statistics, University 

of Guelph, Guelph, Ontario, Canada, mcojocar@uoguelph.ca 

We relate variational inequalities, noncooperative games and hybrid dynamical 

systems so as to describe the _disequilibrium_ evolution of an equilibrium 

problem (e.g. a dynamic network equilibrium problem or a dynamic game). 

To describe it, we use a hybrid system with a switch & jump mechanism between 

continuous dynamic states given by a differential equation. The hybrid 

system also provides a way to analyze stability issues of hybrid solutions w.r.t. 

the problem’s equilibrium states. We apply the ideas to modelling population 

groups’ strategies playing a noncooperative vaccinating game. 

� MB-43 


8.2.02 


Stream: Algorithmic Decision Theory 


Chair: Alexis Tsoukiàs, CNRS - LAMSADE, Université Paris 


tsoukias@lamsade.dauphine.fr 

1 - Constructing indexes: poverty measurement 

Alexis Tsoukiàs, CNRS - LAMSADE, Université Paris 


tsoukias@lamsade.dauphine.fr, Vivien Kana 

In this paper we present a general approach on how to construct indicators 

aimed at being used in conceiving, implementing and evaluation public policies. 

The basic idea consists in using unsupervised and supervised classification 

procedures in such a way that the policy maker can construct an indicator 

useful for the specific policy to pursue. The case of "poverty measurement" is 

then discussed. 

2 - The quality of life in Milano: a rating analysis 

Alberto Colorni, Department of Industrial Design, delle Arti e 

della Comunicazione, Politecnico di Milano, c/o Metid, p.zza 

Leonardo da Vinci 32, 20133, Milano, Italy, 

alberto.colorni@polimi.it, Alessandro Luè 

The association MeglioMilano has been conducting for the past 20 years an 

Observatory of the Quality of Life in Milano. MeglioMilano created a year-byyear 

historical dataset in order to compare the city with itself. This year, the 

authors proposed a classification of the years based on based on the Electre tri 

outranking rating method. For the elicitation of the parameters (thresholds and 

criteria weights), 19 experts have been interviewed. The core of the method 

is to determine if there are enough reasons in favor and no vetoes as regards a 

statement like "year X is placed in class Y’. 

3 - The value of interconnecting water sources: a real option 

approach 

Chiara D’Alpaos, DIMEG, University of Padova, via Venezia 1, 

35131, Padova, Italy, chiara.dalpaos@unipd.it 


Technological innovations lead to the construction of water utilities characterized 

by a high operational flexibility and high irreversible sunk costs. The 

interconnection and integration between supply sources, in particular, enable 

the system to handle crisis in the provision of the service caused, for example, 

by pollution emergencies or peaks in day demand curves. This operational flexibility 

have an economic value since it gives the provider the option to strategically 

decide the optimal switching rule between two different water sources 

and maximize its profits accordingly. 

4 - Comparison of alternative approaches for calculation of 

human development indices 

Veronika Skocdopolova, Dept. of Econometrics, University of 

Economics Prague, Winston Churchill sq. 4, 130 67, Prague 3, 

Czech Republic, veronika.skocdopolova@vse.cz, Josef 

Jablonsky 

The Human Development Report (HDR) is an independent report which is published 

every year by the United Nations Development Programme (UNDP). 

Since the first Report was published in 1990, four composite indices for human 

development have been created. Computations of these indices are multicriteria 

decision problems. In this paper we compare the standard computational 

approaches with other quantitative methods based on multi-criteria decision 

making. All methods are applied on the newest data set published in 

2009. 

� MB-44 


8.2.03 

Portfolio Decision Analysis II 



Chair: Nikolaos Argyris, Management, Operational Research Group, 

London School of Economics and Political Science, Houghton Street, 

WC2A 2AE, London, United Kingdom, n.argyris@lse.ac.uk 




1 - CUT and CUTE: a new multicriteria approach for nonadditive 

concavifiable preferences 

Nikolaos Argyris, Management, Operational Research Group, 

London School of Economics and Political Science, Houghton 

Street, WC2A 2AE, London, United Kingdom, 

n.argyris@lse.ac.uk, José Rui Figueira, Alec Morton 

We propose a new preference aggregation-disaggregation approach for concavifiable 

preferences, Concave UTility (CUT). CUT defines a space of possible 

value functions consistent with a DM’s expressed preferences. There is an analogy 

between CUT and the well-known UTA procedure: CUT is more general 

as it does not require that preferences are additive. CUT can be used in an interactive 

setting; we call this usage CUTE (CUT Elicitation) and we show how 

CUT might be used in pre-ordering a finite set of discrete alternatives, linear 

programming and combinatorial optimization. 

2 - Using interactive heatmaps and parallel coordinate 

plots to support multi-criteria portfolio selections 

Christian Stummer, Department of Business Administration, 

University of Vienna, Bruenner Str. 72, A-1210, Vienna, Austria, 

christian.stummer@univie.ac.at, Elmar Kiesling, Johannes 

Gettinger, Rudolf Vetschera 

We compare interactive heatmaps and parallel coordinate plots as decision aids 

for multi-criteria portfolio selection by means of an experiment conducted with 

students at the University of Vienna. In this experiment, each participant used 

one of the two methods for exploring a solution space of Pareto efficient class 

schedules. We describe the experimental setting, summarize results with respect 

to subjective user evaluations and objective measures, and report findings 

on the impact of decision making styles on users’ attitudes towards the two 

methods. 

3 - Using Linear and Non-Linear Programming Methods for 

recommendation in multicriteria sorting problems 

Philippe Nemery, Department of Mathematics, University of 

Portsmouth, Portsmouth, United Kingdom, 

pnemery@gmail.com, Alessio Ishizaka, Dylan Jones 

43


Several multicriteria decision aid procedures have been proposed to assign a set 

of actions to predefined classes or groups. In this work, we are interested in the 

following problem. Given the initial performances of an action, the aim is to 

analyse how that performances of an action should be enhanced efficiently (i.e., 

with the minimum cost) in order to be assigned into a better category. On the 

other hand, we may also propose an efficient saving in performances without 

changing the assignment result. 

4 - Portfolio project management using ahp: a case study 

Amanda Silva, Produção, ITA, Brazil, 

amanditasimoes@gmail.com, Mischel Carmen N. Belderrain, 

Ademilton Santos, Eduado Inoue, Milton Chagas Júnior 

This work applies a Multi-criteria Decision Making method, the Analytic Hierarchic 

Process — AHP, for the portfolio project management. The case study 

was applied to foreign trade company in São Paulo, Brazil. The criteria considered 

are Maturity of the Project Area, Investment / Return and Product Quality. 

Application in a real case has shown that AHP is only useful for a preevaluation 

of the projects, since the final screening is dependent on subjective 

factors. 

� MB-45 


8.2.12 

Dynamic Programming II 

Stream: Dynamic Programming 


Chair: Lidija Zadnik Stirn, Biotechnical Faculty, University of 



1 - A Sequential Decision Probem with Partial Maintenance 

on Partially Observable Markov Process 

Toru Nakai, Faculty of Education, Chiba University, Yayoi 1-33, 

Inage-Ku, 263-8522, Chiba, Japan, t-nakai@faculty.chiba-u.jp 

We develop an optimal maintenance policy for some products. During their 

life cycle, conditions which causes some troubles changes. The decision-maker 

does not observe this condition directly, but information is obtained through a 

magnitude of a trouble. These conditions change according to a Markov process. 

As for information, a random variable exists for a magnitude of a trouble, 

and improve information by employing the Bayesian learning procedure. The 

decision-maker decides a level of repair or maintein a faulty component with 

cost which is different depending on the level. This problem is formulated as a 

sequential decision problem on a partially observable Markov process. The dynamic 

programming formulation implies a recursive equation about the optimal 

value, and observe monotonic properties for this value. 

2 - Solving the Top-percentile Traffic Routing Problem by 

Approximate Dynamic Programming 

Xinan Yang, School of Mathematics, Uniersity of Edinburgh, 

43/3 Prestonfield Avenue, EH16 5EQ, Edinburgh, Midlothian, 

United Kingdom, s0677435@sms.ed.ac.uk, Andreas Grothey 

This study investigates the optimal routing strategy under multi-homing in case 

where network providers charge ISPs according to top-percentile pricing. It 

is a multi-stage stochastic problem in which decisions should be made before 

knowing the traffic to be shipped. Solution approaches based on SIP or DP suffer 

from the curse of dimensionality which restricts their applicability. To overcome 

this we use Approximate Dynamic Programming (ADP) which exploit 

the structure of the problem to construct approximations of the value function 

in DP. Thus the curse of dimensionality is largely avoided. 

3 - Single-Leg Airline Revenue Management with Overbooking 

and Cancellations 

Nursen Aydin, Industrial Engineering, Sabanci University, 


nursenaydin@su.sabanciuniv.edu, S. Ilker Birbil, J.b.g. Frenk, 

Nilay Noyan 

Airline revenue management (ARM) is about identifying the seat allocation 

policies. In this study we propose new models for static and dynamic singleleg 

ARM problems with overbooking and cancellations. In the static case, we 

introduce computationally tractable models that give upper and lower bounds 

for the expected optimal revenue. In the dynamic case, we propose a new dynamic 

programming model based on two streams of events; booking requests 

and cancellations. We also conduct a set of simulation experiments to illustrate 

the performances of the proposed models. 

44 

4 - Linear Time Exact Algorithm for Capacitor Placement 

on Radial Distribution Feederers 

Christiano Lyra, School of Elec. and Comp. Engineering 

(FEEC), University of Campinas (UNICAMP), Av. Albert 

Einstein 400, Cidade Universitária, 13083-852, Campinas, São 

Paulo, Brazil, christiano@pq.cnpq.br, José Federico Vizcaino, 

Fábio Usberti 

The optimal capacitor allocation problem for radial power distribution feeders 

searches the best compromise between cost of capacitors and their benefits 

to decrease reactive flows on the network. Heuristic approaches dominate the 

scene in addressing this problem. A previous paper proposed the use of dynamic 

programming (DP) to find optimal solutions; however, with simplifying 

assumptions that restricted its application to simple examples. Here, the DP 

approach is extended to consider the main requirements of the problem and 

applied to real scale networks. It is a linear time exact algorithm. 

� MB-46 


8.2.14 

OR Challenges Related to the Recent and 

Future Disasters II 






1 - Coupling of Dynamic Network Flows and a Cellular Automaton 

for Evacuation Modeling 

Markus Thiemann, Mathematics, TU Kaiserslautern, 

Paul-Ehrlich-Strasse 14, Room 438, 67663, Kaiserslautern, 

Germany, thiemann@mathematik.uni-kl.de, Stefan Ruzika, 

Horst W. Hamacher 

We combine dynamic network flows and cellular automata (CA) by a coupling 

technique to obtain realistic predictions for pedestrian evacuation times. Output 

of each model is fed into the other, thus establishing a control cycle. As 

a result, the gap of evacuation times of both models is narrowed: the network 

flow model produces more realistic evacuation times by including data of microscopic 

pedestrian behavior. The CA benefits from route optimization resulting 

in lower evacuation times. These results were obtained within the project 

REPKA: Regional Evacuation, Planning, Control, Adaption. 

2 - How to very quickly solve a staffing and dispatching 

problem for fire and rescue services 

Tobias Andersson Granberg, Department of Science and 

Technology, Linköping University, Division of Communications 

and Transport Systems, SE-60174, Norrköping, Sweden, 

tobias.andersson@liu.se 

A staffing and dispatching problem for fire and rescue services is identified as a 

subproblem to a location model where fire and rescue service units and personnel 

are to be optimally positioned. A simple, very fast heuristic is used to solve 

this problem. The solution quality of the heuristic is important, and in this study 

heuristic solutions are compared to the optimal solutions, produced by solving 

the corresponding model. The results show that the heuristic mostly find the 

optimal solution, and that the exact solution method is too time consuming to 

be practical. 

3 - Logistics for Decision Support in Case of Natural Disasters 

in Developing Countries 

Christian Tesch, Technische Universität Dortmund, 

Leonhard-Euler-Str. 2, 44388, Dortmund, NRW, Germany, 

tesch@vsl.mb.tu-dortmund.de 

In this approach different scheduling heuristics are applied to logistic operations 

in disaster relief. The main objective is to improve the efficiency of handling 

relief goods in terminals. Based on limited resources for cargo handling in 

disaster areas, various weights and sizes of relief goods, improper truck arrival 

times and different departure times we apply and evaluate multiple strategies 

to schedule trucks in the unloading process. A real-time decision algorithm 

applicable in practical situations has been developed to handle the problem of 

uncertain and changing information.

4 - Multi-class Open Queueing Network with Nonpreemptive 

Priority Discipline 

Seongmoon Kim, School of Business, Yonsei University, 134 

Shinchon-dong Seodaemun-gu, 120-749, Seoul, Korea, Republic 

Of, kimsm@yonsei.ac.kr 

Emergency care centers show the characteristics of the open queueing network 

since the patients visit the processes in different orders. In addition, the patients 

are classified based on the levels of urgency to receive service with priority at 

some processes. We present a mathematical model for multi-class open queueing 

network with non-preemptive priority discipline. We present a case study 

based on actual data which have been collected at an emergency care center in 

a hospital. 

� MB-48 


8.2.04 

Ill-posed Variational Problems II 

Stream: Ill-posed Variational Problems - Theory, 

Methods and Applications 





1 - Improving a hybrid preconditioner for solving largescale 

normal equations systems arising from interior 

point methods 

Aurelio Oliveira, Computational & Applied Mathematics, State 

University Of Campinas, DMA IMECC UNICAMP, C. P. 6065, 

13081-970, Campinas, SP, Brazil, aurelio@ime.unicamp.br, 

Marta Velazco 

The normal equations systems are solved by the preconditioned conjugate gradient 

method in two stages. During the early interior point iterations the controlled 

incomplete Cholesky preconditioner is used. As an optimal solution is 

approached, the linear systems become highly ill-conditioned and the specialized 

splitting preconditioner is adopted. A new splitting preconditioner version 

approach that works with a singular submatrix is presented. Numerical experiments 

show that the new approach improves previous performance results for 

both robustness and time on some large-scale problems. 

2 - On Some Properties and Numerical Methods for Problems 

of Topology Optimization 

Wolfgang Achtziger, Department of Mathematics, University of 

Erlangen-Nuremberg, Chair of Applied Mathematics 2, 

Martensstrasse 3, 91058, Erlangen, Germany, 

achtziger@am.uni-erlangen.de 

Topology optimization is a meaningful tool in design optimization of mechanical 

structures. A challenge is the correct formulation and treatment of local 

constraints. For example, stress turns out to be discontinuous as a function of 

the design variables. This is avoided by formulations in design and state variables. 

The price is violation of constraint qualifications and failure of standard 

optimization methods. We present some new results on problem properties, 

optimality conditions, and numerical techniques (Some results together with T. 

Hoheisel, Ch. Kanzow, and Ch. Schuerhoff). 

EURO 24 - Lisbon 2010 MC-02 

Monday, 12:20-13:40 

� MC-01 


Aula Magna 




Chair: Marielle Christiansen, Department of Industrial Economics 

and Technology Management, Norwegian University of Science and 

Technology, Alfred Getz vei 3, N-7491, Trondheim, Norway, 

Marielle.Christiansen@iot.ntnu.no 

1 - Operations Research for Green Logistics 

Rommert Dekker, Erasmus University Rotterdam, Econometric 

Institute, Erasmus University Rotterdam, P.O. Box 1738„ 3000 

DR , Rotterdam, Netherlands, rdekker@few.eur.nl 

After years of talking, green logistics or environmental friendly supply chain 

management seems to have taken off in many companies. In fact, industry 

seems to embark on hundreds of new initiatives, although it may be more popular 

in certain countries than in others. That is on one hand surprising, because 

at first sight becoming green may cost industry money and we would expect 

more initiatives from governments. The attention for green logistics in the OR 

community, however seems to lag behind in this respect. In this overview we 

will first indicate the importance of green aspects to logistics and supply chain 

management and the drivers behind it. Next we will give an overview of various 

aspects of green logistics. We start with extending the traditional cost objective 

in quantitative logistic studies with environmental criteria, like emissions, 

in the form of CO2, NOx and PM. In a following step we discuss the physical 

supply chain drivers such as transportation, facilities and products and highlight 

the fundamental green choices considered in this respect. Thereafter we focus 

on the soft drivers, by considering logistic concepts, decision support systems, 

sourcing and pricing in the supply chain. In doing so, we will highlight the contributions 

of OR so far to green logistics. In fact OR has contributed already 

a lot to green logistics by reducing costs in supply chains, although it most 

cases the environmental aspects were not addressed explicitly. Next we focus 

on those problems where a high contribution of Operations Research can be expected. 

These problems deal with congestions, with coordination in transports 

and the use of price mechanisms to improve efficiency and they occur in a great 

variety of cases. Hence there are many opportunities for Operations Research, 

both methodologically, problem oriented as well as supporting discipline in the 

many logistic studies. 

� MC-02 


3.2.14 

Applications of Combinatorial Optimization 



Chair: Nelson Maculan Filho, COPPE / PESC, Universidade Federal 

do Rio de Janeiro, Rio de Janeiro, RJ, Brazil, maculan@cos.ufrj.br 

Chair: Carlile Lavor, Applied Mathematics, UNICAMP, 

IMECC-UNICAMP, 13083-859, Campinas, SP, Brazil, 

clavor@ime.unicamp.br 

1 - Discretization of the Molecular Distance Geometry 

Problem by defining artificial orderings on the molecule 

Antonio Mucherino, INRIA, Lille Nord Europe, France, 

antonio.mucherino@inria.fr, Carlile Lavor, Leo Liberti, Nelson 

Maculan Filho, Tiberius Bonates, Gabriel Paillard 

The Molecular Distance Geometry Problem (MDGP) can be defined as the 

problem of finding all the atomic positions of a molecule by exploiting the 

given distances between some of its atoms. The MDGP is NP-hard in general, 

although some polynomially solvable subclasses are known, e.g. when all the 

distances are given. We discuss artificial orderings on atoms of the molecule 

that restrict the search for molecular conformations (satisfying the given distances) 

on a finite discrete space. The importance of these orderings is related 

to the formulation of the MDGP as a combinatorial optimization problem. 

45

MC-03 EURO 24 - Lisbon 2010 

2 - Heuristic Method for the Multi-choice Multidimensional 

Knapsack Problem 

Raid Mansi, Departamento de Produção e Sistemas, 

Universidade do Minho, Portugal, raidmm@yahoo.fr, Cláudio 

Alves, J. M. Valério de Carvalho, Saïd Hanafi 

The Multi-choice Multidimensional Knapsack (MMKP) Problem is a variant of 

the knapsack problems. The proposed heuristic, which is an iterative process, 

is based on the resolution of a MMKP relaxation to provide an upper bound 

and generate a reduced problem. The resolution of this reduced problem improves 

a lower bound. Pseudo-cuts are added to the problem to eliminate search 

spaces already explored. The resolution of the MMKP relaxation is done using 

columns generation techniques. This Heuristic improves the value of 18 

instances on the 33 referred in recent work. 

3 - The maximum k-balanced subgraph of a signed graph 

Rosa Maria Figueiredo, Departamento de Matemática, 

Universidade de Aveiro, Campus de Santiago, 3800, Aveiro, 

Aveiro, Portugal, rosa.figueiredo@ua.pt, Daniele Catanzaro, 

Martine Labbé 

Let G=(V,E) be an undirected graph and let s be a function that assigns a sign 

to each edge in E: G together with s is called a signed graph. Consider a parameter 

k. A signed graph is k-balanced if V can be partitioned into k sets in 

such a way that positive edges are found only within the sets and negative edges 

go between sets. The maximum k-balanced subgraph problem is the problem 

of finding a subgraph of G that is k-balanced and maximum according to the 

number of vertices. We present a description of the polytope associated with 

this problem and propose an exact algorithm for its solution. 

4 - The coherence problem - polynomially solvable cases 

Gilberto Calvillo, Unidad Cuernavaca, UNAM, Av. Universidad 

S/N, Col. Chamilpa, 62210, Cuernavaca, Morelos, Mexico, 

gilberto@matcuer.unam.mx, David Romero 

Let G be a simple undirected graph whose edges are colored black or blue. Assume 

further that a real positive number (weight) is assigned to each edge of 

G. 

The coherence problem, relevant in Paul Thagard’s studies related to cognition 

processes and decision-making theory, consists in finding a bipartition (L,R) of 

the vertices of G, so as to maximize A+B+C, where A is the sum of weights 

of black edges with both endpoints in L, B is the sum of weights of blue edges 

having one endpoint in L and the other in R, and C is the sum of weights of 

black edges with both endpoints in R. 

The coherence problem is NP-hard because it generalizes the well-known maxcut 

problem. Hence, several heuristics like neural networks and simulated annealing 

have since been proposed for its solution. 

In this talk a polynomial time algorithm will be presented to exactly solve special 

cases of the coherence problem and some of its extensions. 

� MC-03 


3.2.15 

Metaheuristics for the VRPTW 



Chair: Caroline Prodhon, University of Technology of Troyes, 12 rue 


Chair: Philippe Lacomme, Université de Clermont-Ferrand, 63177, 

Clermont Ferrand, France, placomme@sp.isima.fr 

1 - Heuristic methods for the Vehicle Routing Problem with 

Time Windows and Split Deliveries 

Reghioui Mohamed, Logistics and Computer Science, Ecole 

Nationale des Sciences Appliquées de Tétouan, Mhannech II, 

B.P : 2121, 93000, Tétouan, Morocco, m.reghioui@uae.ma, 

Christian Prins, Abdellah El Fallahi 

This work summarizes our investigation into various heuristic methods to 

solve the vehicle routing problem with time windows and split deliveries (SD- 

VRPTW). The SDVRPTW is similar to the well studied Vehicle Routing Problem 

with Time Windows (VRPTW), except that a customer can be serviced 

by more than one vehicle. The paper proposes constructive heuristics and a 

two-stage metaheuristic based on the combination of several components of 

different metaheuristics. Comparisons with exact methods and a tabu search 

algorithm prove the efficiency of the proposed approaches. 

46 

2 - Granular Variable Neighborhood Search for the Team 

Orienteering Problem with Time Windows 

Jan Melechovsky, LOSI, UTT, 12, rue Marie Curie, 10000, 

Troyes, France, jan.melechovsky@utt.fr, Labadi Nacima, Renata 

Mansini, Roberto Wolfler-Calvo 

This note deals with the Team Orienteering Problem with Time Windows 

(TOPTW). The objective is to build m tours visiting a subset of vertices in order 

to maximize the total collected profit. Each tour must respect a given time 

limit. We present a metaheuristic solving the TOPTW. The method is based on 

the idea of exploring reduced granular rather than complete neighbourhood in 

order to gain the efficiency while preserving the quality of obtained solutions. 

3 - Variable neighborhood search for the time-dependent 

vehicle routing problems with time windows 

Stefanie Kritzinger, Department of Business Administration, 

University of Vienna, Bruennerstr. 72, 1210 Vienna, 1210, 

Vienna, Vienna, Austria, stefanie.kritzinger@univie.ac.at, Karl 

Doerner, Richard Hartl, Fabien Tricoire 

We solve a time-dependent vehicle routing problem with time windows (TD- 

VRPTW) with variable neighborhood search. In the TD-VRPTW the time and 

the cost of traversing an arc are deterministic, but both depend on the time at 

which the traversing takes place. Besides the popular impacts in VRPs the timedependency 

gains more and more importance. It makes VRPs much harder because 

there is no simple way to compute the duration of a route, which has an 

impact on solution feasibility and quality. 

4 - Vehicle Routing Problem with Time Windows using 

SOMA 

Ivan Brezina, Department of Operations Research and 

Econometrics, University of Economics, Dolnozemska 1, 85235, 

Bratislava, Slovakia, brezina@euba.sk, Juraj Pekár, Zuzana 

Cicková 

Self Organizing Migration Algorithm (SOMA) (Zelinka, 1999) belongs to the 

class of Evolutionary Algorithms. It is based on geometric principle and even 

though no individuals are created using parents crossover, the position of the 

individuals in the search space are changed during the evolution. SOMA is also 

able to solve the routing problems from the NP-class efficiently. This algorithm 

was used for solving the Vehicle Routing Problem with Time Windows mainly 

for the practical purposes to specify the service of a node (the collection and 

distribution of used products might be limited by e.g. opening hours, on site 

collection is realized only on collection days etc.). 

� MC-04 


3.2.13 

Parallel machine scheduling with 

metaheuristics 






1 - A literature review for the simulated annealing method 

with the implementation on the parallel machines since 

2000 

Pelin Alcan, The Department of Industrial Engineering, Yildiz 

Technical University, Mechanical Faculty„ Yildiz, 34349, 

Istanbul, Turkey, pelinalcan@gmail.com, Huseyin Basligil 

Performing a detailed literature research has revealed that there are numerous 

heuristic methods applied to parallel machine problems. It is seen that, among 

these heuristics, simulated annealing yields the best solution in the shortiest 

time, which is also the case with the other probabilistic approaches. Creating 

this work we have benefited from the databases of many journals. Research 

results show that there is a little work which is the simulated annealing method 

with the implementation of the parallel machine has been noticed.

2 - Scheduling independent jobs on identical parallel machines: 

An effective approach 

Francesca Vocaturo, Dipartimento di Economia e Statistica, 

Università della Calabria, Via Pietro Bucci - Cubo 0C, 87036, 

Arcavacata di Rende (CS), Italy, vocaturo@unical.it, Giuseppe 

Paletta 

We consider the nonpreemptive scheduling of independent jobs on identical 

parallel machines with minimum makespan objective. We propose an effective 

approach, consisting of a construction phase and an improvement phase. The 

constructive algorithm generates and combines partial solutions of the problem. 

We provide a bound on the performance ratio of this algorithm, by distinguishing 

between "prerun’ and "postrun’ value. Innovative local search procedures, 

based on the concept of partial solutions too, are used in the improvement 

phase. The effectiveness of our approach is evaluated through computational 

tests on benchmark instances. 

3 - Speeding up a Rollout Algorithm for Complex Parallel 

Machine Scheduling 

Marco Pranzo, Dipartimento di Ingegneria dell’Informazione, 

Università di Siena, Siena, Italy, pranzo@dii.unisi.it, Michele 

Ciavotta, Carlo Meloni 

Rollout algorithms are easy to implement but they often require high computation 

time. We present some variants of the basic rollout scheme aimed at 

limiting the computational effort while preserving the overall solution quality. 

Namely, we propose dynamic heuristics pruning, candidates reduction mechanisms 

and the hybridization with a local search procedures. A tight lower bound 

is used to certify the quality of the generated solutions. A campaign of computational 

experiments, carried out also on instances from a real manufacturing 

plant, shows the effectiveness of this approach. 

� MC-05 


3.2.16 

Multi-objective metaheuristics 



Chair: Walter Habenicht, Business Administration, University of 

Hohenheim, Lst. fuer IBL(510A), 70593, Stuttgart, Germany, 

walter.habenicht@uni-hohenheim.de 




1 - Pareto-Search in Discrete Vector Optimization Problems. 

Walter Habenicht, Business Administration, University of 

Hohenheim, Lst. fuer IBL(510A), 70593, Stuttgart, Germany, 

walter.habenicht@uni-hohenheim.de 

In this paper we deal with discrete vector optimization problems with large sets 

of efficient solutions. We assume that the efficient set has been identified (for 

example by some metaheuristic approach) and the non dominated set has been 

stored in a special data structure called quad tree. In order to organize efficiently 

a searching process in outcome space, namely in the non dominated set, 

we discuss different neighborhood definitions. These neighborhoods differ in 

the computational complexity of identifying the neighbors and in their ability 

to avoid suboptimal solutions. In all neighborhoods the process of identifying 

the neighbors is supported by the underlying data structure. 

2 - Multiobjective optimization using the Lorenz dominance 

for a buffers sizing problem 

Hicham Chehade, Charles Delaunay Institute, University of 

Technology of Troyes, 12 Rue Marie Curie, 10000, Troyes, 

France, chehadeh@utt.fr, Lionel Amodeo, Farouk Yalaoui 

In this work, a multiobjective buffers sizing is studied. Two objectives are considered: 

the maximization of the throughput rate and the minimization of the 

buffers total size. A new resolution approach based on the Lorenz dominance 

relationship is introduced. The method is called L-archive which is based on 

a genetic algorithm with a niching strategy and uses the Lorenz dominance to 

identify the set of non dominated solutions. The L-archive algorithm is compared 

to the SPEA2 algorithm of Zitzler and the computational experiments 

show the advantages of our method. 


3 - Comparison of multiobjective cooperative and classical 

evolutionary algorithms for global supply chain optimisation. 

Maksud Ibrahimov, Computer Science, The University of 

Adelaide, Plaza Building, The University of Adelaide, SA 5000, 

Adelaide, South Australia, Australia, 

maksud.ibrahimov@adelaide.edu.au, Neal Wagner, Sven 

Schellenberg, Arvind Mohais, Zbigniew Michalewicz 

This paper discusses global optimization of supply chain (SC) operations. Often 

organizations lack communication between SC silos with optimization restricted 

at the silo level which does not always lead to the global optimum. A 

two-silo SC was built based on the combination of vehicle routing and scheduling 

problems. Three approaches were used: a classical evolutionary approach 

and two approaches based on cooperative co-evolution one with multi-objective 

optimization and one without. A real-world problem is presented involving an 

Australian sheet steel business. 

� MC-06 


8.2.30 

DEA Methodology III 



Chair: Wade Cook, Schulich School of Business, York University, 

Management Science, Room S337M, 4700 Keele Street, M3J 1P3, 

Toronto, Ontario, Canada, wcook@schulich.yorku.ca 

1 - Multiple Variable Proportionality in Data Envelopment 

Analysis 

Wade Cook, Schulich School of Business, York University, 

Management Science, Room S337M, 4700 Keele Street, M3J 

1P3, Toronto, Ontario, Canada, wcook@schulich.yorku.ca, Joe 

Zhu 

Data envelopment analysis (DEA) provides an optimization methodology for 

deriving an efficiency score for each member of a set of peer decision making 

units. Under the original DEA model of Charnes, Cooper and Rhodes (1978) it 

was assumed that there are constant returns to scale (CRS). This idea was later 

extended by Banker, Charnes and Cooper (1984) to the more general case that 

allowed for variable returns to scale (VRS). In both of these structures, it is assumed 

that the returns to scale (RTS) classification, consistent with the classical 

definition, applies to the entire (input, output) bundle. In many settings it can 

be the case that the output bundles can be separated into distinct components 

wherein an RTS-type behavior may be different for one component than for another. 

We refer to such situations as involving multiple variable proportionality 

(MVP). Examples of MVP occur when there are different product groupings in 

a manufacturing facility, different wards in hospitals, and so on. Identification 

of such differential behavior can provide management with important insights 

regarding the most productive scale size (MPSS) in each of those components. 

In the current paper we introduce DEA-based tools that address those situations 

where MVP exists. 

2 - Technical Efficiency based on Cost Gradient Measure 

Miki Tsutsui, Socio-economic Research Center, Central 

Research Institute of Electric Power Industry, 2-11-1 Iwadokita, 

Komae-shi, 2018511, Tokyo, Japan, miki@criepi.denken.or.jp, 

Kaoru Tone, Yuichiro Yoshida 

We propose a new model named cost gradient measure (CGM), which enables 

us to measure more price-conscious technical efficiency. In the CGM model, 

projection to the efficiency frontier is defined by the cost gradient direction of 

each decision making unit, which is the normal to the cost plane and the steepest 

dissent direction of total input cost. This will be reasonable for company 

managers. Furthermore, we can derive CCR from CGM formula, which clearly 

implies CCR is under strong restriction of proportionality. Even if cost data is 

not available, CGM can be solved under several assumptions. 

3 - Centralised DEA model for target setting when articulation 

of partial ideal targets are considered 

Gabriel Villa, Departament of Industrial Management, University 

of Seville, Camino de los Descubrimientos, s/n, Seville, Spain, 

gvilla@esi.us.es, Sebastián Lozano, David Canca 

47


DEA models calculate the level of inputs/outputs to render inefficient units efficient. 

This article tries to identify these operating points in a case in which 

the organisations are able to articulate a set of ideal input/output targets that 

the DMU would wish to achieve. A central entity (CE) is the responsible of 

the allocation of some inputs of DMUs and could estimate and enforce to meet 

the amount of inputs and outputs expected for DMUs.We propose a set of DEA 

models that assigns the level of inputs and outputs to each DMU to be efficient, 

according to the ideal levels established by CE. 

4 - On the choice of weights profiles in cross-efficiency 

evaluations 

Nuria Ramón, Centro de Investigación Operativa, Universidad 

Miguel Hernández, Avda. Universidad, 03200, Elche, Spain, 

nramon@umh.es, Inmaculada Sirvent, Jose L. Ruiz 

The literature has claimed that the differences between the weights profiles 

that different DMUs use in their assessments (self-evaluations) may be a concern. 

This is particularly relevant in cross-efficiency evaluations, since the 

cross-efficiency score of a given DMU is usually calculated as an average of 

its ratings obtained with the weights profiles provided by all the DMUs. We focus 

in this paper on the choice of these weights trying to avoid large differences 

between the profiles that the different DMUs provide. We also pay attention to 

problems such as the zero weights and to the possible strategies that can be 

used in the choice of the weights profiles to be considered. 

� MC-07 


8.2.47 

Recent Advances in the Use of 

Mathematical Programming 






Chair: Kaisa Miettinen, Dept. of Mathematical Information 

Technology, University of Jyvaskyla, P.O. Box 35 (Agora), FI-40014, 

University of Jyvaskyla, Finland, kaisa.miettinen@jyu.fi 

Chair: Jussi Hakanen, Dept. of Mathematical Information 

Technology, University of Jyväskylä, P.O. Box 35 (Agora), FI-40014, 

University of Jyväskylä, Finland, jussi.hakanen@jyu.fi 

1 - The last buy decision in situations with an alternative 

item 

Simme Douwe Flapper, Technische Universiteit Eindhoven, 

Netherlands, s.d.p.flapper@tue.nl 

Many different last buy/final buy/last series situations can be found in practice, 

some triggered by suppliers, some triggered by customers. First, we present 

a classification of the different last buy/final buy/last series situations found in 

practice. Next we focus on the many situations where an alternative item is/or 

will become available. A mathematical model to support the last buy decision 

in the latter situations is presented, and the results obtained for a case study are 

discussed. Finally, an overview of areas for further research is given. 

2 - Corporate Investment Choice and Exchange Option between 

Production Functions 

Jean-luc Prigent, ThEMA, University of Cergy-Pontoise, 33, Bd 

du Port, 95011, CERGY-PONTOISE, France, 

jean-luc.prigent@u-cergy.fr, Olfa Bouasker 

We examine the strategy in resource allocation of a firm which must choose 

between several production functions using the real option approach. We look 

for the option value of the managerial flexibility between production functions. 

First, we determine the values of exchange options when the underlying asset 

is no more a geometric Brownian motion but equal to the sum of the net present 

value and of the corresponding growth / delay options. Second, we give a general 

model to evaluate Profit and Loss. We provide a valuation formula and 

show how risk aversion modifies the optimal choice. 

48 

� MC-08 


6.1.36 

Airside Airport Operations 



Chair: Jason Atkin, School of Computer Science, University of 

Nottingham, Jubilee Campus, Wollaton Road, NG8 1BB, 

Nottingham, Notts, jaa@cs.nott.ac.uk 

Chair: Chris Potts, School of Mathematics, University of 

Southampton, Highfiled, SO17 1BJ, Southampton, Hampshire, 

United Kingdom, C.N.Potts@soton.ac.uk 

1 - Realistic runway scheduling problems 

Jason Atkin, School of Computer Science, University of 


Nottingham, Notts, jaa@cs.nott.ac.uk, Edmund Burke 

Since the runway(s) can often for the bottleneck for the arrival and departure 

systems at airports, achieving high quality arrival and departure sequences is 

extremely important. We will discuss the similarities and important differences 

between these problems, showing why a solution method for one will not necessarily 

be useful for the other despite the similarity of the models. We will 

then extend this illustration to discuss why the real problems actually involve 

more than pure sequencing and present some more realistic models and results 

for London Heathrow. 

2 - Models and Algorithms for Scheduling Aircraft Landings 

Chris Potts, School of Mathematics, University of Southampton, 

Highfiled, SO17 1BJ, Southampton, Hampshire, United 

Kingdom, C.N.Potts@soton.ac.uk, Julia Bennell, Mohammad 

Mesgarpour 

There is a great interest in optimizing the usage of airport runways due to an 

anticipated increase in numbers of flights during the next decade. Although a 

significant amount of previous research on runway scheduling has been carried 

out, very little of it has been implemented. This paper studies the scheduling 

of aircraft landing on a single runway. A generic mixed integer linear programming 

model is presented. Also, we propose some fast heuristics that are 

suitable for use by air traffic controllers. Finally, initial computational results 

and future work directions are discussed. 

3 - Comparison of airport ground movement algorithms 

Stefan Ravizza, School of Computer Science, University of 


Nottingham, United Kingdom, smr@cs.nott.ac.uk, Jason Atkin, 

Edmund Burke 

The airport ground movement optimisation problem links other important airport 

airside operations together. With the expected increase in air traffic, ground 

movement operations will become increasingly important. It is essential to understand 

the characteristics of ground movement optimisation problems and 

solution approaches, since airport layouts differ significantly. We will compare 

exact and heuristic approaches for a number of supplied benchmark scenarios 

and critically analyse the differences between them and the advantages and 

disadvantages of each. 

4 - A comparison of constructive algorithms for baggage 

sorting station allocation 

Amadeo Ascó, ASAP, School of Computer Science and IT, 

University of Nottingham, C87, Jubilee Campus, Wollaton Road, 

NG8 1BB, Nottingham, Nottinghamshire, United Kingdom, 

aaz@cs.nott.ac.uk, Jason Atkin, Edmund Burke 

The correct allocation of limited airport resources can greatly affect the quality 

of service provided by airlines and airports to their customers. Efficient 

baggage handling is already important, and likely to become more so with the 

expected increases in civil air traffic. We will define the (previously neglected) 

baggage sorting station allocation problem and provide a model. We will then 

describe a number of constructive algorithms for solving this innately multiobjective 

problem and compare them against each other and previous work, 

illustrating the trade-offs between the objectives.

� MC-09 


6.2.53 

Complementarity Problems, Variational 

Inequalities and Equilibrium 



Chair: Sandor Zoltan Nemeth, School of Mathematics, The 

University of Birmingham, The Watson Building, Edgbaston, B15 

2TT, Birmingham, United Kingdom, nemeths@for.mat.bham.ac.uk 

Chair: Florian Potra, Mathematics & Statistics, University of 

Maryland, MD 21250 , Baltimore, United States, potra@umbc.edu 

1 - Weighted Complementarity Problems: Theory and Applications. 

Florian Potra, Mathematics & Statistics, University of Maryland, 

MD 21250 , Baltimore, United States, potra@umbc.edu 

We introduce the notion of a weighted linear complementarity problem which 

generalizes the classical notion of a linear complementarity problem. We show 

that such problems arise in a natural way in market equilibrium and other contexts. 

We present some general theoretical results concerning the properties of 

the solution set of such problems, and present a class of interior point methods 

for their numerical solution. We give general conditions that insure the polynomial 

complexity and superlinear convergence of the interior point methods. 

2 - Consistent Conjectures in Mixed Oligopoly 

Vyacheslav Kalashnikov, Systems and Engineering, ITESM 

(Monterrey Technological Institute), ITESM, Ave. Eugenio 

Garza Sada 2501 Sur, 64849, Monterrey, Nuevo Leon, Mexico, 

kalash@itesm.mx, Vladimir Bulavsky, Nataliya Kalashnykova 

A model of mixed oligopoly with conjectured variations equilibrium (CVE) 

is considered. The agents’ conjectures concern the price variations depending 

upon their production increase or increase. We establish existence and uniqueness 

results for the CVE (called an exterior equilibrium) for any set of feasible 

conjectures. To introduce the notion of an interior (or consistent) equilibrium, 

we develop a consistency criterion for the conjectures and prove the existence 

theorem for such an equilibrium. For the extension of our results to the case 

of non-differentiable demand functions, we also investigate the behavior of the 

consistent conjectures in dependence upon a parameter. 

3 - Properties of Schur complements in Euclidean Jordan 

algebras 

Roman Sznajder, Mathematics, Bowie State University, 14000 

Jericho Park Road, 20715-9465, Bowie, Maryland, United 

States, rsznajder@bowiestate.edu 

We study the concept of Schur complement in the setting of Euclidean Jordan 

algebras and describe Schur determinantal and Haynsworth inertia formulas. In 

addition, we prove an analogue of the Crabtree-Haynsworth quotient formula 

and show that any Schur complement of a strictly diagonally dominant element 

is strictly diagonally dominant. We also introduce the concept of Schur product 

of a real symmetric matrix and an element of Euclidean Jordan algebra when its 

Peirce decomposition with respect to a Jordan frame is given. An Oppenheim 

type inequality is proved in this setting. 

4 - Iterative methods for nonlinear complementarity problems 

on isotone projection cones 

Sandor Zoltan Nemeth, School of Mathematics, The University 

of Birmingham, The Watson Building, Edgbaston, B15 2TT, 

Birmingham, United Kingdom, nemeths@for.mat.bham.ac.uk 

We present a recursion related to a nonlinear complementarity problem defined 

by a closed convex cone in a Hilbert space and a continuous mapping defined 

on the cone. If the recursion is convergent, then its limit is a solution of the 

nonlinear complementarity problem. In the case of isotone projection cones 

sufficient conditions are given for the mapping so that the recursion to be convergent. 

EURO 24 - Lisbon 2010 MC-10 

� MC-10 


6.2.56 

Graphs and Applications 



Chair: Bernard Ries, Warwick Business School, University of 

Warwick, CV4 7AL, Coventry, Switzerland, 

Bernard.Ries@wbs.ac.uk 

1 - Fractionally co-strongly perfect claw-free graphs with 

an application to wireless networking 

Yori Zwols, Department of Industrial Engineering and 

Operations Research, Columbia University, Rm. 313, School of 

Engineering and Applied Sciences, 500 West 120th Street, 

10027, New York, NY, United States, yz2198@columbia.edu 

Strongly perfect graphs have been studied by several authors (e.g. Berge, 

Duchet, Ravindra, Wang). This talk deals with a fractional relaxation of strong 

perfection. Motivated by a wireless networking problem, we consider clawfree 

graphs that are fractionally strongly perfect in the complement. We obtain 

a forbidden induced subgraph characterization and an approximate structure 

theorem, and we display graph-theoretic properties of such graphs. This is 

joint work with Maria Chudnovsky and Bernard Ries. 

2 - Profit Maximizing Pricing on graphs. 

Rajiv Raman, DIMAP and Department fo Computer Science, 

University of Warwick, Department of Computer Science, 

University of Warwick, CV4 7AL, Coventry, West Midlands, 

United Kingdom, R.Raman@warwick.ac.uk 

We are given a graph, whose edges we wish to price. We are given a set of 

buyers, each with a budget and a specific path the buyer is interested in buying. 

The buyer will buy the path if she the sum of the prices of the edges in the 

path is at most her budget. Our objective is to set prices to the edges in order 

to maximize the profit we obtain by selling the paths to the buyers who can 

afford them. I will present complexity and approximation algorithms for this 

problem. This is joint work with K. Elbassioni, S. Ray and R. Sitters. 

3 - Coloring vertices of triangle-free graphs 

Bernard Ries, Warwick Business School, University of Warwick, 

CV4 7AL, Coventry, Switzerland, Bernard.Ries@wbs.ac.uk 

The vertex coloring problem is known to be NP-complete in the class of 

triangle-free graphs. Moreover, it remains NP-complete even if we additionally 

exclude a graph F which is not a forest. We study the computational complexity 

of the problem in (triangle; F)-free graphs with F being a forest. From known 

results it follows that for any forest F on 5 vertices the vertex coloring problem 

is polynomial-time solvable in the class of (triangle; F)-free graphs. We show 

that the problem is also polynomial-time solvable in many classes of (triangle; 

F)-free graphs with F being a forest on 6 vertices. 

4 - Computing identifying codes in chosen graph classes 

and dynamic scenarios 

Adrian Kosowski, Dept. of Algorithms and System Modeling, 

Gdansk University of Technology, ul. Narutowicza 11/12, 

80-233, Gdansk, Poland, adrian@kaims.pl 

An identifying code in a graph is a subset of vertices whose intersection with 

the open neighborhood of each vertex is a distinct, non-empty set. Identifying 

codes are applied, among others, in fault-detection in networks and for locating 

fires in facilities. 

In this talk we present some new structural and computational results for the 

problem of determining minimal identifying codes in some graph classes inspired 

by real-world network topologies. We also consider algorithms for locally 

updating identifying codes in dynamic settings, in which nodes are allowed 

to join or leave the network. 

49


� MC-11 


8.2.38 

Energy Market Modeling 



Chair: Steven Gabriel, Civil & Env. Engin./ Applied Math and 

Scientific Computation Program, University of Maryland, 1143 

Martin Hall, 20742, College Park, MD, United States, 

sgabriel@umd.edu 

1 - Analysis of a Possible Natural Gas Cartel 

Steven Gabriel, Civil & Env. Engin./ Applied Math and 

Scientific Computation Program, University of Maryland, 1143 

Martin Hall, 20742, College Park, MD, United States, 

sgabriel@umd.edu, Knut Rosendahl 

In this presentation we present an analysis of the global gas market under several 

possible cartels involving the Gas Exporting Countries Forum (GECF). 

We make use of the World Gas Model, a large-scale complementarity model 

for determining Nash-Cournot based market equilibria. 

2 - Using Real Options to Evaluate Optimal Funding Strategies 

for Carbon Capture and Storage (CCS) Projects in 

the European Union 

Jeremy Eckhause, Civil and Environmental Engineering, 

University of Maryland/LMI, 2000 Corporate Ridge, 22102, 

McLean, Virginia, United States, jeckhause@lmi.org, Johannes 

Herold 

A barrier to large scale implementation of CCS is the lack of demonstration 

projects that validate the technology. A few projects in the EU are under development 

to use CCS on a large scale. Taking a funding agency’s perspective, 

we employ a real options framework to select an optimal project portfolio. We 

solve stochastic dynamic programs to obtain funding strategies in order to maximize 

success by a target year. The model demonstrates the reduction of risk 

in the multi-stage competition, while considering knowledge spillover. State 

space, computational complexity and runtimes are analyzed. 

3 - A Model for Oligopolistic Natural Gas Markets 

Ibrahim Abada, Electricite de France/Univ. Paris, Université 

Paris Ouest, Nanterre - La Défense, 200, Avenue de la 

République, 92001, Nanterre, France, ibrahim.abada@edf.fr, 

Steven Gabriel, Olivier Massol, Vincent Briat 

In our model, the interaction between certain market players is posed as a generalized 

Nash-Cournot competition. We take into consideration the long-term 

contracts aspects. The producers sell their gas to a set of independent traders 

who sell it then to end-users. Storage and transportation aspects are taken care 

of by global operators. We use a system dynamics approach to model possible 

fuels substitution between the consumption of coal, oil and natural gas. We describe 

some of the theoretical aspects as well as preliminary numerical results 

for the European natural gas market. 

4 - Cartelisation in the Natural Gas Market: the Stability Issue 

Olivier Massol, Center for Economics and Management, IFP - 

IFP School, 228-232 Avenue Napoléon Bonaparte, 92852 , 

Rueil-Malmaison, France, olivier.massol@ifp.fr, Stéphane 

Tchung-Ming 

The creation of the Gas Exporting Countries Forum (GECF) has motivated 

numerous discussions. In that context, Egging et al. (2009) have recently proposed 

a numerical model to measure the market power that could potentially be 

exerted by a coalition of gas exporters. As the GECF is often described as "an 

informal association" with an unstable membership, an investigation focussed 

on the cartel stability might be needed. This is precisely the goal of our contribution 

that illustrates how a numerical model (a simple mixed complementarity 

problem) can provide some policy-relevant conclusions. 

50 

� MC-12 


8.2.39 

AHP 03 


Processes 


Chair: Anabela Pereira Tereso, Production and Systems Department, 

University of Minho, Campus de Azurém, 4800-058, Guimarães, 

Portugal, anabelat@dps.uminho.pt 

1 - Multicriteria Decision Aid: Evaluation and Comparison 

of the Main Tools 

Anabela Pereira Tereso, Production and Systems Department, 

University of Minho, Campus de Azurém, 4800-058, Guimarães, 

Portugal, anabelat@dps.uminho.pt, Cristina Seixedo 

Good decision making is increasingly more important to organizations. This 

work presents a review of the literature in multicriteria decision aid with a reference 

to the main techniques available in the area. We also present a research 

on the software tools available in this field. These software tools were then 

characterized and classified and their main characteristics summarized in the 

present work. This work is part of a project that has as final goal to implement 

a software tool using AHP to help on the selection of the right multicriteria 

decision aid software. 

2 - Environmental Risk Assessment for Roadway Transportaion 

of LNG 

Emel Topuz, Environmental Engineering, ITU, ITU Ayazaga 

Yerleskesi Insaat Fakultesi, Cevre Muhendisligi Bolumu Maslak, 

34469, ISTANBUL, Turkey, topuze@itu.edu.tr, Ilhan Talinli, 

Atakan Oztekin 

The objective of this study is to apply an environmental risk assessment approach 

for roadway transportation of Liquid Natural Gas (LNG) by using fuzzy 

AHP. Environmental and transportation factors are scored and compared with 

fuzzy numbers by experts and fuzzy inference methodology is applied. The 

outputs of this study helped to understand the source of risk and its magnitude 

which can be used to develop risk management for transportation of LNG. 

Moreover, quantified environmental risk magnitude is significant to use for further 

decision analyses in which environmental factors included. 

3 - Designing a Prioritization Model for Investment Plans 

under Uncertainty Using Interval Comparison Matrices 

Seyed Hesameddin Anvar, Industrial Engineering, Yazd 

University, Yazd, Yazd, Iran, Islamic Republic Of, 

hesam.anvar@gmail.com, Masoud Narenji, Mohammad Fathian 

In hierarchical MCDM methods, one of the main steps is to weigh criteria and 

compute each alternative weight according to the defined criteria. One of the 

most common weighting criteria methods is to apply the pairwise comparison 

matrices. In this paper we model uncertainty in investment plans prioritization 

by using interval comparison matrices as inputs for Lexicographic Goal 

Programming (LGP) and Two-stage Logarithmic Goal Programming (TLGP) 

methods. Finally a numerical example for real location problem is solved for 

each method and compared with Analytic Hierarchy Process (AHP) approach. 

4 - A Two Phased Multi Criteria Approach for Aviation Accident 

Analysis 

Ahmet Kandakoglu, Department of Industrial Engineering, 

Istanbul Technical University, Macka, 34367, Istanbul, Turkey, 

kandakoglu@itu.edu.tr, Erol Yucel, Y. Ilker Topcu 

This study proposes a two phased approach based on the 5M (Man, Machine, 

Medium, Mission and Management) model and the Analytic Hierarchy Process 

(AHP) method for quantitatively analyzing the aviation accidents. While 

the 5M model is used to identify all the causal factors in an accident, the AHP 

method is applied to quantify these factors using the subjective judgments of 

experts. This approach provides both an analytical framework to assess which 

factor influences the accident most and valuable insights for the preventive actions 

to reduce the accident risk in the safety management process.

� MC-13 


2.2.21 

Managing disruptions in facility location 

models 



Chair: Maria Paola Scaparra, Kent Business School, University of 

Kent, The University, CT2 7PE, Canterbury, United Kingdom, 

M.P.Scaparra@kent.ac.uk 

1 - Node Fails, Facility Fails: Facility Location Problem for 

Emergency Evacuation 

Rongbing Huang, School of Administrative Studies, York 

University, M3J 1P3, Toronto, Ontario, Canada, 

rhuang@yorku.ca, Mozart Menezes, Seokjin Kim 

In the p-median and p-center problems, people take advantage of the facility 

located at the same site. However, in the situation of some natural disasters 

like hurricane Katrina, or the recent earthquake in Haiti, the whole city may 

become functionless. Therefore, customers can’t rely on the facility located at 

the same place. We compare this problem with the center problem and analyze 

the problem on some simple networks. An efficient algorithm is provided for 

the problem on the general network. 

2 - Incorporating recovery time issues in facility protection 

planning 

Chaya Losada, Business School, Kent University, CT2 7PE, 

canterbury, United Kingdom, cl243@kent.ac.uk, Maria Paola 

Scaparra 

We consider a p-median system whose facilities are vulnerable to external disruptions. 

Disrupted facilities are not operative for a given length of time, the 

recovery time. We propose a protection model which identifies the optimal allocation 

of a protection budget among the system facilities so as to minimize 

the impact of potential disruptions over a planning horizon. We assume that 

protection reduces the facility recovery time. The proposed model is a mixed 

integer bi-level problem with integer variables in both levels. We solve it efficiently 

by decomposition methods. 

3 - A Bilevel Fixed Charge Location-Protection Model for 

Facilities under Imminent Attack 

Deniz Aksen, College of Administrative Sciences and 

Economics, Koç University, Rumelifeneri Yolu, Sariyer, 34450, 

Istanbul, Turkey, daksen@ku.edu.tr, Necati Aras, Nuray Piyade 

In a public service network customers always travel to the nearest facility to 

get service. A system planner has to determine the locations of facilities each 

of which can be opened either in the protected or unprotected mode. The associated 

fixed costs differ from one candidate site to another. Protected facilities 

are immune against an attacker who is capable of destroying a certain number 

of unprotected facilities in the worst-case scenario. Partial protection or interdiction 

is not possible. We formulate this problem as a static Stackelberg game 

between the system planner and the attacker. 

4 - An integer programming formulation for the nonuniform 

p-median problem with unreliable facilities 

Maria Paola Scaparra, Kent Business School, University of Kent, 

The University, CT2 7PE, Canterbury, United Kingdom, 

M.P.Scaparra@kent.ac.uk, Jesse O’Hanley, Sergio García Quiles 

We consider the problem of locating p unreliable facilities on a network so as 

to minimize the expected sum of the weighted travel distances between facilities 

and customers. Facilities can have different probabilities of failure and 

customers have prior information on the operational status of the facilities. We 

present a novel mixed integer programming formulation for this problem and 

report numerical experiments on some benchmark data sets. 


� MC-14 


2.2.15 

Price and Capacity Planning in Supply 

Chains 



Chair: Esma Gel, Industrial Engineering, Arizona State University, 

P.O. Box 5906, 85287-5906, Tempe, AZ, United States, 

esma.gel@asu.edu 

1 - Dynamic price and lead time quotation for make-toorder 

systems with contract customers and spot purchasers 

Esma Gel, Industrial Engineering, Arizona State University, P.O. 

Box 5906, 85287-5906, Tempe, AZ, United States, 

esma.gel@asu.edu, Ahmet Hafizoglu, Pinar Keskinocak 

We consider dynamic price and lead time quotation for a make-to-order company 

with demand from contract customers as well as spot purchasers. Contract 

customers are offered a uniform price and lead time, and upgraded service 

that prioritizes their orders. Spot purchasers are subject to dynamically quoted 

price and lead times, which are accepted or rejected by the spot purchasers with 

known probability. We discuss the potential of dynamic quotation for such environments, 

various properties of optimal control policies, and the optimal mix 

of contract customers and spot purchasers. 

2 - An Approach to Robust Capacity Allocation in Supply 

Chains 

Thomas Makuschewitz, IPS - Intelligent Production and 

Logistics Systems, BIBA - Bremer Institut für Produktion und 

Logistik GmbH at the University of Bremen, Hochschulring 20, 

28359, Bremen, Bremen, Germany, mak@biba.uni-bremen.de, 

Bernd Scholz-Reiter, Fabian Wirth, Michael Schönlein 

A perturbation can change the dynamic behavior of a supply chain. As a consequence 

planned capacity levels might no longer be sufficient to handle the 

work load. In this talk we present an approach to robust capacity allocation 

that ensures robustness of a supply chain with respect to perturbations of customer 

demand. As a measure for the robustness we use the stability radius that 

reflects the size of the smallest perturbation that destabilizes the supply chain. 

Based on results concerning this measure we set up an optimization problem 

for the capacity allocation at each production facility. 

3 - A comparative study of near-optimal and flexible 

strategic remanufacturing capacity planning policies in 

CLSCs 

Efstratios Athanasiou, Mechanical Engineering, Aristotle 

Univesity of Thessaloniki, Laskaratou 11, 54646, Thessaloniki, 

Greece, efsatha@auth.gr, Patroklos Georgiadis 

We study strategic remanufacturing capacity planning policies in a two-product 

CLSC. The system’s response (transient flows, actual remanufacturing capacity 

level, profit) is studied throughout the two-product life cycle, under alternative 

scenarios regarding the market preference and taking explicitly into account 

the end-of-use product returns according to a residence time distribution. We 

adopt a dynamic planning tool based on System Dynamics methodology. The 

comparative study of near-optimal and flexible policies provides useful insights 

regarding remanufacturing capacity decisions. 

4 - Supply Chain Management: planning decisions and 

economical performance criteria 

Ana Cristina Santos Amaro, Instituto Sup. Contabilidade e 

Administração de Coimbra, ISCAC, Instituto Politécnico de 

Coimbra, Quinta Agrícola, Bencanta, 3040 Coimbra, Portugal, 

3040, Coimbra, cristinaamaro@sapo.pt, Ana Paula 

Barbósa-Póvoa 

In the global economy efficient supply chains, SCs, are required to guarantee 

profitability within uncertain markets. In this paper, a centralized managing 

strategy is proposed to model SC planning decisions based on different economical 

criteria evaluation, while analysing partnership scenarios within the 

global chain as well as different product portfolios demand and prices. A 

Mixed Integer Linear Programming approach (MILP) is developed to model 

SC planning decision scenarios. The applicability of the proposed approach is 

illustrated by a real case-study taken from the pharmaceutical sector. 

51


� MC-15 


2.2.12 

Green Vehicle Routing Problems 



Chair: Tolga Bektas, University of Southampton, School of 

Management, Highfield, SO17 1BJ, Southampton, -, United 

Kingdom, T.Bektas@soton.ac.uk 

1 - Analysis of Travel Times and CO2 Emissions in Time- 

Dependent Vehicle Routing 

Ola Jabali, Eindhoven University of Technology, 5600 MB , 

Eindhoven, Netherlands, O.Jabali@tue.nl, Tom Van Woensel, 

Ton de Kok 

We consider the time-dependent VRP from two standpoints; one seeks to optimize 

exclusively on travel time; the other does so on total CO2 emissions. 

We also present a cost-based model that optimizes on a weighted average of 

travel time, emission and fuel costs. As the amount of CO2 emissions is correlated 

with vehicle speed, the emissions-based optimization is done by limiting 

vehicle speed. The emission per kilometer as a function of speed attains its 

minimum at speed v*. We show that limiting vehicle speed to v* might be 

sub-optimal. Solutions are obtained via a tabu search procedure. 

2 - Fuel Emission Optimization in Vehicle Routing Problems 

with Time-Varying Speeds 

Jiani Qian, Dept. Management Science, Lancaster University, 

b204 house 3, graduate college,lancaster university, la2 0pf, 

Lancaster, United Kingdom, j.qian@lancaster.ac.uk, Richard 

Eglese 

The objective is to produce routes and schedules for a fleet of vehicles that minimize 

the fuel emissions in a road network where speeds depend on time. A new 

heuristic approach selects potential routes and then determines the speeds on 

each road for every vehicle that has been assigned a list of customers. This approach 

is compared to a time-increment based dynamic programming method 

in searching for the least polluting route between two customer nodes. To solve 

the full VRP, the new heuristic algorithm is embedded into a tabu search algorithm. 

3 - An Analytical Approach to Medical Waste Management 

in Istanbul 

Tülin Aktin, Industrial Engineering Department, Istanbul Kültür 

University, Ataköy Campus, E-5 Karayolu Uzeri, Bakırköy, 

34156, Istanbul, Turkey, t.aktin@iku.edu.tr, Ilayda Karabulut, 

Suzan Yavuztürk 

In this study, the vehicle routing problem encountered in Istanbul during the 

collection of medical wastes from hospitals with bed capacities over 20 is analyzed. 

Due to the geographical location of the city, the problem is examined in 

two parts: European side (PE) and Asian side (PA). Two MILP models (PE and 

PA) are developed to find the routing of vehicles that collect medical wastes 

from hospitals, as well as, to determine the location/allocation of new disposal 

sites on both sides of the Bosphorus. The study is implemented using real data 

obtained from Istanbul Metropolitan Municipality. 

4 - The Pollution-Routing Problem 

Tolga Bektas, University of Southampton, School of 

Management, Highfield, SO17 1BJ, Southampton, -, United 

Kingdom, T.Bektas@soton.ac.uk, Gilbert Laporte 

This talk will present the Pollution-Routing Problem (PRP), an extension of 

the classical Vehicle Routing Problem (VRP) with a broader objective function 

that accounts for the amount of greenhouse gas emissions, fuel, travel times 

and their costs. Mathematical models will be described for the PRP with or 

without time windows. Results of computational experiments performed on realistic 

instances will be presented to shed light on the tradeoffs between various 

parameters such as vehicle load, speed and total cost, and to offer insight into 

economies of ‘environmental-friendly’ vehicle routing. 

52 

� MC-16 


2.2.14 

Rescheduling in railway operations 



Chair: Gabrio Curzio Caimi, Netzdesign und Fahrplan, BLS Netz 

AG, Genfergasse 11, 3001, Bern, Bern, Switzerland, 

gabrio.caimi@bls.ch 

Chair: Marco Laumanns, IBM Research Zurich, 8803, Rueschlikon, 

Switzerland, mlm@zurich.ibm.com 

1 - Comparing Multi-Criteria Decision Making Methods for 

a Railway Dispatching Problem 

Anne Binder, Dresden University of Technology, 01069, 

Dresden, Germany, Anne.Binder@mailbox.tu-dresden.de, 

Thomas Albrecht, Christian Gassel 

In the literature on railway rescheduling, the optimization function consists of 

maximum 3 components: train delays, passenger discomfort and energy consumption. 

This multi-criteria problem is solved by reducing it to one cost function 

consisting of the weighted sum of the individual criteria. Here, the use of 

other methods for multi-criteria decision making is examined for a straightforward 

dispatching problem on a junction. It is proposed to use conflict probability 

as 4th criterion. By analyzing the Pareto front the sensitivity of the solution 

towards the weighting factors is shown. 

2 - The effectiveness of rescheduling strategies: evidences 

from two case studies in Italy 

Giorgio Medeossi, Department of Civil and Environmental 

Engineering, University of Trieste, Piazzale Europa 1, 34127, 

Trieste, Italy, giorgio.medeossi@phd.units.it, Giovanni Longo 

The accuracy and performance of real time rescheduling algorithms is rapidly 

increasing. Before implementing them in control centres, it is important to 

quantify their benefits and understand their requirements. In this study, the effects 

of different rescheduling strategies in the nodes of Florence and Venice 

are reproduced by means of stochastic micro-simulation, pointing out the relationship 

between infrastructure topology, timetable structure and the objective 

functions. Despite the apparently similar topology, rescheduling strategies perform 

significantly different. 

3 - Dispatching support system for a main station 

Martin Fuchsberger, D-Math, Insitute for Operations Research, 

ETH Zurich, Raemistrasse 101, 8092, Zurich, Zurich, 

Switzerland, fumartin@ifor.math.ethz.ch, Marco Laumanns 

A dispatching support system for the station area of Berne, Switzerland is presented. 

The model assigns to each train a path based on a variety of alternative 

track routes, station platforms, speed profiles and arrival or departure times. 

While respecting safety regulations based on detailed blocking time theory, the 

optimal assignment is minimizing a weighted combination of delay propagation 

and number of broken connections. The model is iteratively adapted to the 

current state situation and solved in a rolling time horizon framework to cope 

with the continuously changing online situation. 

4 - Coping with extremely variable train delay on a complex 

railway junction: lesson learned from Beograd case 

study 

Giovanni Longo, DICA, University of Trieste, Via Valerio, 6/1, 

34127 , Trieste, Italy, longo@dica.units.it, Giorgio Medeossi, 

Zorica Milanovic 

In large and complex railway nodes, different train categories and services 

share the same infrastructure. Moreover, they usually have different stochastic 

behaviour, and different reliability targets. However, from the infrastructure 

Manager perspective, a global reliability should be maintained, also through a 

adequate timetabling strategy. In this paper, starting from the analysis of the 

real behaviour of different train services in the Belgrade junction, a number of 

possible timetable structures are compared through stochastic micro-simulation 

in terms of reliability.

� MC-17 


1.3.14 

Online Planning in Vehicle Routing and 

Scheduling 



Chair: Stefan Bock, WINFOR (Business Computing and Operations 

Research) Schumpeter School of Business and Economics, 

University of Wuppertal, Gaußstraße 20, D-42097 Wuppertal, 42097, 

Wuppertal, NRW, Germany, sbock@winfor.de 

Chair: Francesco Ferrucci, WINFOR (Business Computing and 

Operations Research), University of Wuppertal, Gaussstrasse 20, 

42119, Wuppertal, NRW, Germany, fferrucci@winfor.de 

1 - A Dynamic Vehicle Routing Problem with Multiple Trips 

Michel Gendreau, MAGI and CIRRELT, École Polytechnique, 

C.P. 6079, succ. Centre-ville, H3C 3A7, Montreal, Quebec, 

Canada, michel.gendreau@cirrelt.ca, Nabila Azi, Jean-Yves 

Potvin 

We consider a routing problem in which vehicles performing multiple shortduration 

trips serve customer requests that arrive dynamically according to a 

known stochastic process. In this problem, requests may be turned down as 

they arrive, but the associated revenues are lost. The objective is to maximize 

the profits (revenues from served requests - transportation costs). Our solution 

method relies on a sample of demand scenarios, which is used to evaluate opportunity 

costs related to future requests. Computational results on randomlygenerated 

test instances will be reported and analyzed. 

2 - Schedule Nervousness Reduction in Online Vehicle 

Routing 

Joern Schoenberger, Operations Research and Supply Chain 

Management, RWTH Aachen, Templergraben 64, 52056, 

Aachen, Germany, schoenberger@or.rwth-aachen.de, Stefan 

Voss, Herbert Kopfer 

A vehicle schedule determines the arrival times of vehicles at customer sites. 

As soon as additional customer-specified requests become known a scheduler 

changes visiting sequences of vehicles and/or shifts requests among the available 

vehicles having to prevent significant variations of once announced arrival 

times at customer sites. We propose a penalty-based modeling approach for 

online vehicle routing problems that addresses explicitly the keeping of previously 

decided arrival times. Within computational simulation experiments, we 

demonstrate the general applicability of our approach. 

3 - Solving the Dynamic Ambulance Relocation Problem 

using Approximate Dynamic Programming 

Verena Schmid, Faculty of Business, Economics and Statistics, 

University of Vienna, Bruenner Strasse 72, 1210, Vienna, 

Austria, verena.schmid@univie.ac.at 

For emergency service providers it is crucial to locate ambulances such that 

patients can be reached fast enough in case of emergency. The problem can 

be formulated using a dynamic program. We solve the dynamic ambulance 

relocation problem using approximate dynamic programming methods. Ambulance 

redeployment decisions will be made real-time. Our computational 

experiments show that the policy obtained outperforms greedy strategies used 

traditionally by ambulance service providers. 

4 - An analysis of the impact of dynamic vehicle routing in 

service delivery systems 

Vitória Pureza, Engenharia de Produção, Universidade Federal 

de São Carlos, Via Washington Luiz, km 235, 13565-905, Sao 

Carlos, Sao Paulo, Brazil, vpureza@dep.ufscar.br, Daniel 

Lazarin 

In this work, we analyze the impacts resulting from the incorporation of dynamic 

vehicle routing and scheduling in service delivery systems for which 

meeting the delivery due dates is a priority issue. Specifically, we propose a 

constructive/deconstructive heuristic for the Dynamic Vehicle Routing Problem 

in order to obtain routes in real time. Sets of instances randomly generated 

from the data supplied by a drink company in Brazil are used to assess the 

relative performance of the heuristic to other methods. 


� MC-18 


1.3.15 

Stochastic Modeling and Simulation II 



Chair: Zeev (Vladimir) Volkovich, Ort Braude Academic College, 

Yehiam 6, 21823, Karmiel, Israel, zeev@actcom.co.il 

1 - Approximation Algorithms for the General Multi Assignment 

Problem 

Ron Adany, Computer Science, Bar-Ilan University, 52900, 

Ramat-Gan, Israel, adanyr@cs.biu.ac.il, Sarit Kraus, Fernando 

Ordonez 

We consider the problem of assigning personalized advertisements to viewers 

in order to maximize revenue. Each viewer has a limited capacity and each 

ad has a given length and a revenue that is obtained if it is assigned to a given 

number R of different viewers. We focus on the Ads Packing problem assuming 

a subset of ads which can be packed is given. We present two bi-criteria approximations 

algorithms: Extra-Packing a (1,2)-approximation algorithm and 

Deep-Search-Replacer, an (R+1,1)-approximation algorithm. Both algorithms 

were obtained by rounding the relaxed LP solution of the problem. 

2 - A randomized algorithm for estimation number of clusters 

Zeev (Vladimir) Volkovich, Ort Braude Academic College, 

Yehiam 6, 21823, Karmiel, Israel, zeev@actcom.co.il, Oleg 

Granichin, Dmitry Shalymov, Renata Avros 

We address to the well known machine learning problem concerning estimation 

of clusters amount in a given dataset. Our approach offered in the framework of 

the common "elbow" methodology such that the true number of clusters is recognized 

as the discontinuity point of the differential risk function. A randomized 

optimization algorithm is applied to allocate this position. Appropriate 

theoretical study and numerical experiments provided demonstrate high ability 

of the proposed method together with its relatively low complexity cost in the 

case of big suggested number of clusters. 

3 - On the number of coincidence of two homogeneous 

random walks with positive paces. 

V. N. Surikov, Institution of Russian Academy of Sciences 

Dorodnicyn Computing Centre of RAS, 12345, Moscow, 

Russian Federation, svna@zao-zt.ru, I. A. Kravchenko 

Objects identification tasks arise in many practical problems connected to the 

genetic objects’ recognition and complex information systems’ monitoring. 

There are some original results that can be used to separation hypotheses’ 

about membership of two independent samples to the same object in the case 

of noised observation, or to stochastic close diferent states of the same object. 

The most commonly results are on the renewal theory’s fundamental postulates 

based, and can be used to the large-scale technical systems’ operational 

simulation. 

4 - Synergic process of speech signal energy transmission 

Vladimir Zhuravlev, Zaporozhzhya Technical University, 22 

Nagornaya Street, office 301, 04107, Kiev, ws_50@mail.zp.ua, 

Dorovskykh Anatoliy 

To explain contradictions of imperfect adequacy of speech and hearing process 

theories the math model of speech signal (SS) energy generation and receiving 

is posed. The model is based on synergic analysis of SS informational 

components in communication channel energy transfer speed and communication 

channel energy substance carriers — air molecules average quadratic 

speed ratio. Theoretical, estimating and experimental researches, which indirect 

prove the SS synergic properties, were carried out. The report is available 

at http://kudin.net/r/index.php/20100204-scientific-library.html 

53


� MC-19 



Mathematical Finance 



Chair: Diogo Pinheiro, CEMAPRE - ISEG, Technical University of 

Lisbon, Rua do Quelhas, 6, 1200-781, Lisboa, Portugal, 

dpinheiro@iseg.utl.pt 

1 - Optimal Life Insurance Purchase, Consumption and Investment 

Decisions with Multiple Risky Securities 

Isabel Duarte, Universidade do Minho, Rua Monte do Barral, 37, 

Real, 4700-280, Braga, Portugal, isabelduarte@math.uminho.pt, 

Diogo Pinheiro, Alberto A. Pinto, Stanley Pliska 

We analyse the problem faced by a wage earner looking for optimal consumption, 

investment, and insurance purchase strategies in order to maximize expected 

utility of consumption, of the size of the estate in the event of premature 

death, and of the size of the estate at the time of retirement for an underlying 

financial market composed by one risk-free security and an arbitrary finite 

number of risky securities whose diffusive term is given by a multi-dimensional 

Brownian motion. We use dynamic programming methods to obtain explicit 

solutions for the case of CRRA utility functions. 

2 - Risk and beliefs updating mechanisms for contingent 

claims valuation in incomplete markets 

Diogo Pinheiro, CEMAPRE - ISEG, Technical University of 

Lisbon, Rua do Quelhas, 6, 1200-781, Lisboa, Portugal, 

dpinheiro@iseg.utl.pt, Alberto A. Pinto, Athanassios 

Yannacopoulos, Stylianos Xanthopoulos 

We review two scenarios for the determination of asset prices in an incomplete 

market. One scenario is based on the update of the attitude towards risk of 

the agents involved on the trading of such assets, whereas the other scenario is 

based on the update of their beliefs about the future states of the world. Furthermore, 

we describe dynamic mechanisms that model the convergence of the 

buyer’s and the seller’s prices of a contingent claim in an incomplete market 

to a unique price and discuss the stability and robustness of these mechanisms 

with respect to random perturbations. 

3 - Behavioural scenarios for contingent claims valuation 

in multiperiod incomplete markets 

Nuno Azevedo, Mathematics, Cemapre, Rua do Quelhas n. o 6, 

1200-781, Lisboa, Portugal, ncazevedo@gmail.com, Diogo 

Pinheiro, Alberto A. Pinto, Athanassios Yannacopoulos, 

Stylianos Xanthopoulos 

We describe behavioural scenarios for the determination of asset prices in incomplete 

multi-period financial markets. All these scenarios assume that the 

participating agents have some initial beliefs about the distribution of the future 

states of the world which the agents may be willing to update in order to 

reach an agreement about the price of a given asset. This final price is determined 

by optimizing some relevant quantity which can be seen as a measure of 

satisfaction or regret of the agents involved in the trade of such an asset. 

4 - Real Options and Signaling in Strategic Investment 

Games 

Takahiro Watanabe, Business Administration, Tokyo 

Metropolitan University, Minamioosawa 1-1, Hachiouji_city, 

1920397, Tokyo, Japan, contact_nabe08@nabenavi.net 

An investment game with an incumbent and an entrant is examined. The profit 

flows involve two uncertain factors: (1) the basic level of demand of the market 

observed only by the incumbent and (2) the fluctuation of the demand described 

by a geometric Brownian motion which is common to both firms. In our model, 

the incumbent enters into the market earlier than the entrant, so the timing of 

the incumbent’s investment may reveal the information of the demand level. I 

characterize this signaling effect and investigate the real option values of both 

firms. 

54 

� MC-20 


1.3.33A 




Chair: Ramon Alvarez-Valdes, Statistics and Operations Research, 

University of Valencia, Faculty of Mathematics, Doctor Moliner 50, 

46100, Burjassot, Spain, ramon.alvarez@uv.es 

1 - The container loading problem with practical considerations 

Maria Teresa Alonso, University of Valencia, 46100, Burjassot, 

mai.alonso@gmail.com, Francisco Parreno, Ramon 

Alvarez-Valdes, Jose Tamarit 

We have developed a hybrid algorithm for the three-dimensional container 

loading problem taken into account practical considerations. Embedded in a 

GRASP structure we have designed a constructive algorithm inspired on Ngoi’s 

box placement and a VND procedure for the improvement phase using increasingly 

aggressive moves. 

2 - Coordination of the Vehicle Routing with Time Windows 

and the Container Loading 

Olinto Araújo, CTISM, Universidade Federal de Santa Maria, 

Brazil, olinto@densis.fee.unicamp.br, Vinícius Armentano 

We address the integrated problem of the vehicle routing with time windows 

and container loading, involving practical constraints that are often found in 

freight transportation. A tabu search algorithm is proposed for the vehicle routing 

problem, and the loading problem is solved by a multi-start random constructive 

heuristic with a load arrangement based on generalized cuboids that 

fit in given empty spaces. Computational tests on instances from the literature 

are reported. 

3 - A model for the Container Stowage Problem 

Ana Moura, Economics, Management and Industrial 


Santiago, 3810-193 , Aveiro, Portugal, ana.moura@ua.pt, Paulo 

Triunfante Martins, António Andrade-Campos, Victor Lobo 

This work presents a model for a fleet of containerships. Three different challenges 

will be addressed: how to select the fleet navigation routes; how to distribute 

the cargo by the available vessels; how to stow cargo on the vessels. The 

cargo at each port must be delivered within a time limit by containerships, each 

of them with different characteristics. The main subject is the ship’s stowage 

plan because it has a great influence on port handling and vessels stability. In 

this work, a model for the problem resulted by the integration of CSP and VRP 

is presented and solved. 

4 - Cutting And Packing Algorithms Research Framework 

Denis Nazarov, Ufa State Technical University of Aviation, 

Russian Federation, denis.nsc@gmail.com 

Cutting And Packing Algorithms Research Framework 

(http://caparf.googlecode.com/) is an open-source cross-platform framework 

under GPLv3 license. It provides convenient "testing scenario" 

model, various reports generators and easy way for developing your own 

algorithms. Caparf consists of problems definition (for each type of problems), 

algorithms for solving particular type of problems or even a number of types 

(including lower/upper bounds, exact methods) and test instances (raw data 

and generators). 

� MC-21 


6.2.47 

Optimization Modeling I 



Chair: Robert Fourer, Industrial Engineering and Management 

Sciences, Northwestern University, 2145 Sheridan Road, 

60208-3119, Evanston, IL, United States, 

4er@iems.northwestern.edu

1 - Attacking Hard Mixed-Integer Optimization Problems 

through the AMPL Modeling Language 

Robert Fourer, Industrial Engineering and Management 

Sciences, Northwestern University, 2145 Sheridan Road, 

60208-3119, Evanston, IL, United States, 

4er@iems.northwestern.edu, David M. Gay 

There are many tricks for formulating complex optimization models by use of 

integer variables, but what’s to be done when even the most advanced solvers 

can’t produce results in reasonable time? A series of examples show how substantial 

improvements in performance can be achieved through carefully focused 

troubleshooting and experimentation facilitated by the power and flexibility 

of the AMPL modeling language and its solver interfaces. 

2 - A Structure-Conveying Modelling Language for Mathematical 

and Stochastic Programming 

Andreas Grothey, School of Mathematics, University of 

Edinburgh, The King’s Buildings, West Mains Road, EH9 3JY, 

Edinburgh, United Kingdom, A.Grothey@ed.ac.uk, Jacek 

Gondzio, Marco Colombo, Kristian Woodsend, Jonathan Hogg 

We present a structure-conveying algebraic modelling language for mathematical 

programming. The proposed language extends AMPL with object-oriented 

features that allows the user to construct models from submodels. Unlike traditional 

modelling languages, the new approach does not scramble the block 

structure of the problem, and thus enables the passing of this structure to the 

solver. Solvers that exploit block linear algebra can therefore directly take advantage 

of the problems structure. The language contains features to conveniently 

model stochastic programming problems. 

3 - Stochastic Optimization: Recent Enhancements in Algebraic 

Modeling Systems 

Lutz Westermann, GAMS Software GmbH, Eupener Str. 

135-137, 50933, Koeln, Germany, lutz@gams.com, Michael 

Bussieck 

With all the uncertainty in data there is considerable demand for stochastic optimization 

in operational, tactical and strategical planning. Nevertheless, there 

exist a fairly small number of commercial applications building on stochastic 

optimization techniques. After a decade without much progress on the software 

side, modeling system providers have recently entered the second round 

of making it easier to go from a deterministic model to a stochastic version of 

such a model. We will review the different concepts emphasizing on recent 

enhancements in the GAMS system. 

4 - Rapid application development with OptimJ, a practitioner’s 

experience report 

David Gravot, rostudel, 57, rue d’alleray, 75015, Paris, France, 

France, dgravot@rostudel.com, Patrick Viry 

We report practical experience with the OptimJ modeling language on a number 

of customer projects led by Rostudel Operations Research. Rostudel has 

extensive experience with tools such as OPL, COMET, GAMS, AMPL, MPL, 

AIMMS. Recurrent problems when developing applications are the poor integration 

of optimization and data modeling, preprocessing and integration with 

software applications. OptimJ overcomes these difficulties by building directly 

upon the Java language. We focus on practical examples of complex and sparse 

data collections that frequently occur in optimization applications. 

� MC-22 



OR in Education I 



Chair: Huiling Zhao, Kent Busincess School, University of kent, 19 

holter Mill, CT2 8SP, Canterbury, Kent, United Kingdom, 

hz34@kent.ac.uk 

1 - Structured Process for systematic review of literature 

and the establishment of the References in Multi- 

Criteria Decision Aid 

William Vianna, Production Engineering, Federal University at 

Santa Catarina (UFSC) - BRAZIL, Rua Cônego Bernardo, 100. 


ap. 202, Trindade, 88036570, Florianópolis, Santa Catarina, 

Brazil, wpwilliam@hotmail.com, Edilson Giffhorn, Leonardo 

Ensslin 

The objective of this paper is explore the use of a Structured Process for systematic 

review of literature and the establishment of the References in Multi- 

Criteria Decision Aid applications. This is carried in two research projects that 

deal with "Meta-evaluation of Educational Systems’ and "Performance Indicators’. 

The results are the selection of authors, journals and publications of 

impact that allow the validation of References, and consequently, the development 

of analysis and data interpretation. 

2 - The Environmental Impact of Free school choice in Kent 

(UK) 

Huiling Zhao, Kent Busincess School, University of kent, 19 

holter Mill, CT2 8SP, Canterbury, Kent, United Kingdom, 

hz34@kent.ac.uk, Kim Parker, Cecilio Mar Molinero 

In Kent (UK) there is a selective school system: academic results at the age 

of eleven determine the type of secondary school that children can attend and 

this may influence their life outcomes. Some parents are prepared to drive 

their children long distances from home to a "good’ school causing congestion, 

pollution, and social cost. We allocate children to schools based on distance. 

By comparing model results with the actual, we assess environmental impact. 

The results confirm the well established view that social background is a major 

determinant in the dynamics of free school choice. 

3 - Social Responsibility Management System: The case 

study of ESTG-IPG 

Ana Rosa, Guarda Polythecnic Institute, 6300-559, Guarda, 

acristina@ipg.pt, Rute Abreu, Constantino Rei 

Higher education is a strategic imperative, because neither CSR nor quality 

management system will subordinate to the other. The paper is centred in the 

new role of corporate social responsibility practices to achieve transparency, 

accountability and sustainability. The empirical evidence is based on a case 

study of the Escola Superior de Tecnologia e Gestão and it develops an exploratory 

analysis. The results show the application Portuguese Standard NP 

4469-1:2008 showing the new public management practices. 

� MC-23 


6.2.49 


Combinatorial Optimization III 



Chair: Luis Paquete, Department of Informatics Engineering, 

University of Coimbra, Polo II, 3030-290, Coimbra, Portugal, 

paquete@dei.uc.pt 

1 - Comparing two territory partitions in districting problems:Indices 

and practical issues 

Fernando Pereira, Matemática, Universidade da Beira Interior, 

Covilha, Portugal, fpereira@noe.ubi.pt, José Rui Figueira, 

Vincent Mousseau, Bernard Roy 

Planning is an activity of critical importance, with direct relevance for urban 

planners. The ramifications of such decisions generally have significant effect 

on peoples’ lives. The current paper deals with the comparison between territorial 

maps. The theoretical problem involves the comparison of two partitions 

in a connected, undirected, and planar graph. In considering this problem, we 

introduce three new indices to compare territory partitions: compatibility, inclusion, 

and distance, all of which have importance for real-world planning 

situations. 

2 - A Parallel Evolutionary Algorithm with Multiple Reference 

Points for Multi-objective Combinatorial Optimization 

Arnaud Liefooghe, LIFL - CNRS - INRIA, Université Lille 1, 

Lille, France, arnaud.liefooghe@lifl.fr, José Rui Figueira, 

El-Ghazali Talbi, Andrzej Wierzbicki 

55


When approximating the efficient set, multiple reference points can be uniformly 

distributed within a region that covers the Pareto frontier. Here, a 

preference-based evolutionary algorithm, based on an achievement scalarizing 

function, is defined for each reference point, so that all the algorithms can easily 

be launched in a parallel and distributed environment. Computational experiments 

are performed on a bi-objective flow-shop scheduling problem. Our results 

show that the parallelization of evolutionary multi-objective optimization 

algorithms based on reference point methods gives very promising and statistically 

significantly better results than comparable non-parallel approaches. 

3 - The Tree-Dimensional Bin-Packing with 90 o rotations 

Marisa Figueiredo, Department of Informatics, CISUC, 

University of Coimbra, Departamento de Engenharia Informática 

da FCTUC, Pólo II, Pinhal de Marrocos, 3030-290, Coimbra, 

Portugal, msbfigueiredo@gmail.com, Ana Almeida 

An interesting multi-criteria variant for the 3-Dimensional Bin Packing problem 

(3BPP) consists of determining the minimum number of three-dimensional 

rectangular bins that are required for orthogonally allocating a given set of 

three-dimensional rectangular items without overlap and minimizing the occupied 

space: the 3BPP-min problem. We propose a new approach for the 

3BPP-min problem where 90 degrees rotations are allowed in order to obtain 

more compact packing. Computational results show the effectiveness of the 

new approximation algorithm. 

4 - Multicriteria movement synchronization scheduling 

problems and algorithms 

Zbigniew Tarapata, Faculty of Cybernetics, Military University 

of Technology, Kaliskiego Str.2, 00-908, Warsaw, Poland, 

zbigniew.tarapata@wat.edu.pl 

The paper deals with multicriteria optimization model of movement scheduling 

for many objects to synchronize their movement. The model consists of 

two parts: disjoint path planning model for K objects; movement synchronization 

model in some intermediate nodes. We use two criteria for disjoint path 

planning: sum and maximum of achieving times of destination nodes by all objects. 

For synchronous movement two categories of criteria are defined: time 

of movement and "distance’ of K moved objects from movement pattern. Some 

algorithm for solving the problem and its properties is considered. 

� MC-24 


6.2.50 

Bioinformatics III 


Medicine 





Chair: Marta Szachniuk, Institute of Bioorganic Chemistry, PAS, 

Noskowskiego 12/14, 61-704, Poznan, Poland, 

Marta.Szachniuk@cs.put.poznan.pl 

1 - A Novel Framework to Elucidate Core Classes in a 

Dataset 

Daniele Soria, IMA, School of Computer Science, University of 


Nottingham, United Kingdom, dqs@cs.nott.ac.uk, Jonathan 

Garibaldi 

A new framework to extract representative groups from a dataset will be presented. 

It specifies the application of different clustering methods, then statistical 

and visualisation techniques are used to characterise the results, and 

core classes are defined by consensus clustering. Classes may be verified using 

supervised classification to obtain rules which may be useful for new data. 

The framework is validated over a set of biological markers for breast cancer 

patients. The resultant classes are well separated and characterised by low, 

medium and high levels of markers. Clinically, the groups distinguish patients 

with poor overall survival from those with low grade and better survival. 

2 - Infrastructure behind the EPipe framework 

56 

Peter Sackett, Center for Biological Sequence Analysis, 

Technical University of Denmark, Kemitorvet, Building 208, 

DK-2800, Kongens Lyngby, Denmark, pws@cbs.dtu.dk 

The talk will describe the EPipe framework, which is designed to detect functional 

differences between similar protein sequences, e.g. isoforms derived 

from alternative splicing, sequences indicating locus variation, or protein families 

from one or more organisms. EPipe draws on large databases and many, 

computationally intensive, protein feature prediction tools. In the post-genomic 

era with the 1000$ genome around the corner, EPipe can be used to meet 

the challenge of revealing functional variation, but the computational effort 

involved in handling large scale submissions is considerable. 

3 - Graph theoretic approaches to clustering 

Roberta De Asmundis, Dept. of Mathematics and Applications, 

University of Napoli FEDERICO II, Compl. MSA, Via Cintia, 

80126, Napoli, Italy, robertadeas@hotmail.com, Paola Festa, 

Mario Guarracino 

Spectral clustering has been recently proposed as a clustering algorithm simple 

to implement and that can be solved efficiently by standard linear algebra 

software. When data are given in form of a similarity graph, the clustering can 

be restated as the problem of finding a partition of the graph such that edges 

between different groups have a very low weight and the edges within a group 

have high weight. In a graph this can be related to a minimum cut problem 

which is NP hard. In this work we explore the possibility to decrease the computational 

complexity of a de facto standard algorithm for graph partitioning 

providing evidence of its performance on real world problems. 

4 - Predicting the evolution of Biomedical Ontologies 

Catia Pesquita, Department of Informatics, Universidade de 

Lisboa, Edifício C6 Piso 3, Campo Grande, 1749 - 016, Lisboa, 

Lisboa, Portugal, cpesquita@xldb.di.fc.ul.pt, Francisco Couto 

Ontologies represent one of the most relevant breakthroughs in bioinformatics, 

providing a semantic model that supports tasks such as text and data mining, 

database interoperability, annotation, and automated reasoning. However, the 

task of maintaining biomedical ontologies up to date is daunting due to the 

dynamical nature of biomedical knowledge. Here, we present preliminary results 

on the prediciton of the evolution of the Gene Ontology, one of the most 

proeminent biomedical ontologies, comparing different sets of features across 

several versions. 

� MC-25 


6.2.48 

Risk Management in Operations 



Chair: Stefan Spinler, Production Management, Beisheim School of 

Management, WHU, Burgplatz 2, 56179, Vallendar, Germany, 

sspinler@whu.edu 

1 - A mixed-integer linear programming model for 

sovereign debt issuance 

Malek Abdel Jawad, Mathematical Sciences, Brunel University, 

John Crank Building„ Kingston Lane„ UB8 3PH, Uxbridge, 

Middlesex, United Kingdom, mapgmma@brunel.ac.uk, Paresh 

Date, Alessandra Canepa 

Governments borrow funds to finance the excess of cash payments or interest 

payments over receipts, usually by issuing fixed income debt. The goal of 

this work is to determine the composition of the portfolio issued by the government,to 

minimize the cost of servicing debt while controlling risk and maintaining 

market liquidity. We show that this debt issuance problem can be modelled 

as a stochastic optimization problem with a receding horizon. The stochastic 

nature of interest rates is modelled using Kalman filter. We demonstrate the 

utility of our approach by back-testing on UK debt issuance. 

2 - Pricing of Full Service Repair Contracts 

Sebastian Huber, Chair of Logistics Mgmt., WHU Otto Beisheim 

School of Management, Burgplatz 2, 56179, Vallendar, 

Germany, sebastian.huber@whu.edu, Stefan Spinler 

Investment products (e.g. industrial trucks) are often sold with a fixed fee repair 

contract. The service provider has to set a price offer to heterogeneous 

customers optimizing his portfolio profit for contracted and on-call service. We 

combine price willingness based on risk aversion under mean-variance utility 

with a cost model based on stochastic failures and repair cost. Moral hazard is 

constrained by a guaranteed service time with a penalty based on a queue. We 

find the optimal price depending on customer cost expectation and incentivizing 

the provider to inform customers truthfully.

3 - Game-theoretic analysis of forward market under uncertainty 

Alexander Vasin, Operations Research, Moscow State 

University, Leninskie Gory, MGU, VMK faculty, 119991, 

Moscow, Russian Federation, vasin@cs.msu.su, Agata Sharikova 

This paper considers an oligopoly with fixed marginal cost. The outcome at 

both the forward and the spot market is a Cournot outcome dependent on demand 

and supply at the market. We assume that the outcome at the spot market 

is random and determine optimal behavior of consumers depending on reservation 

prices and risk-aversion parameters. Producers aim to maximize their 

profits via choosing subgame perfect equilibrium (SPE) of the two-stage game 

as their strategies. We show that there exists an SPE in correlated mixed strategies. 

The random variable determines one of two possibilities for spot market: 

either "bear market’, or "bull market’. We compare this SPE with Nash equilibria 

of one-stage markets 

4 - Retailer Product Portfolio Selection & Pricing Problem 

Soumojit Kumar, Operations management, Indian Institute of 

Management, Calcutta, Annexe-204, Indian Institute of 

Management Calcutta, Diamond Harbour Road, Joka, 

Kolkata-700104, 700104, Kolkata, West bengal, India, 

soumojitk08@iimcal.ac.in, Uttam Sarkar 

Retailers with limited storage capacity have to decide to stock the right mix of 

product variants to optimize their dual objective of maximizing profit as well 

as retain market share. The optimal portfolio solution is determined and the 

pricing of each variant is also a decision variable. The retailer has a certain 

amount of market share and can vary the prices of the variants between a maximum 

retail price and a lower bound. Excess inventory is sold at a salvage value 

usually less than the cost of the variant making a loss. Heuristics are developed 

to reach the optimal point. 

� MC-26 


3.1.11 

Machine Learning to help people with 

dissabilities 



Chair: Emilio Parrado-Hernandez, Signal Processing and 

Communications, Universidad Carlos III de Madrid, Avenida de la 

Universidad 31, 28911, LEGANES, emipar@gmail.com 

Chair: Jaisiel Madrid-Sanchez, Technosite, 28037, Madrid, 

jmadrid@technosite.es 

1 - Automatic data processing for web accessibilty personalization 

Olatz Arbelaitz, Computer Architecture and Technology, 

University of the Basque Country, Manuel Lardizabal 1, 20018, 

Donostia, Euskal Herria, Spain, olatz.arbelaitz@ehu.es, Julio 

Abascal, Javier Muguerza, Iñigo Perona 

People with physical, sensory or cognitive restrictions have difficulties to take 

advantage of the facilities offered by the web. The barriers can be surmounted 

using systems that are able to adapt the interfaces to the users’ characteristics. 

Most actual user models are designed based on the preconceived characteristics 

of users with disabilities. The use of machine learning techniques to build 

user models will allow to take into account their real capabilities since those 

techniques will extract knowledge from the information obtained from web 

navigation logs belonging to the users. 

2 - Identification of gait alterations with a wearable inertial 

system 

Andreu Catala, CETPD, Universitat Politecnica de Catalunya, 

Rambla de l’Exposicio 59-69, 08800, Vilanova i la Geltru, Spain, 

andreu.catala@upc.edu, Albert Samà, Diego Pardo 

Gait perturbations due to chronicle diseases like Parkinson or as a consequence 

of the age as in elderly people are an important cause of disability. Recognizing 

and monitoring of gait alterations is crucial in order to accomplish better 

diagnosis, better rehabilitation design and also fall prevention. We apply PCA 

to an organized set of time series provided by the wearable inertial sensor used 

during the patient daily life activities. It allows us to discriminate the internal 

dynamics of the system and correlate them with the actual spatio temporal 

properties obtained during gait 


3 - Improving accesibility with machine learning 

Manuel Ortega, Research, Technosite, C/ Albasanz 16, planta 3, 

28037, Madrid, Spain, mortega@technosite.es 

People with disabilities and elderly people face great barriers when trying to 

access digital information. Interfaces and contents are usually not designed to 

fit their needs. This could be overcome through automatic generation of interfaces. 

This solution can be addressed as an optimization problem. Machine 

Learning can also provide tools that adapt contents to users with special needs 

and is useful for user modelling and context-varying system adaptations. 

� MC-27 


8.2.06 




Chair: Laura Di Giacomo, Statistica Probabilità Statistiche Applicate, 

Universita La Sapienza, p.le Aldo Moro 5, 00185, Roma, Italy, 

lauradg@caspur.it 

Chair: André Salles, Industrial Engeneering, Federal University of 

Rio de Janeiro - UFRJ, Av. Ataulfo de Paiva, 348 ap. 501 - Leblon, 

22440033, Rio de Janeiro, Rio de Janeiro, Brazil, as@ufrj.br 

1 - Markowitz principles for multi-period portfolio selection 

problems with bankruptcy condition 

Thamayanthi Chellathurai, Treasury and Risk Management, 

Canadian Imperial Bank of Commerce, Capital and Credit Risk 

Analytics, CIBC, 21 Melinda Street, CCE- 9th Floor, M5L 1A2, 

Toronto, Ontario, Canada, thamay.c@gmail.com 

The multi-period portfolio selection problem is formulated as a Markowitz 

mean-variance optimization problem in terms of time-varying means, covariances, 

higher order and inter-temporal moments of the asset prices. To preclude 

possible arbitrage opportunities at future trading dates, a bankruptcy condition 

is enforced approximately. The expected return of the portfolio is dependent 

not only on the means of the asset prices at discrete times, but also on the 

higher order and inter-temporal moments. Numerical results are presented for 

some test problems. 

2 - Optimal Control of Exchange Traded Funds Instruments 

(ETF) 

Laura Di Giacomo, Statistica Probabilità Statistiche Applicate, 

Universita La Sapienza, p.le Aldo Moro 5, 00185, Roma, Italy, 

lauradg@caspur.it 

An optimal control policy is foundamental for the management of ETF throught 

formal mathematical programming system. The formulation of the algorithm 

will be presented and extensive implementation would be discussed. 

3 - An Approximate DP Approach to Benchmark Practicebased 

Heuristics for Natural Gas Storage Valuation 

Nicola Secomandi, Tepper School of Business, Carnegie Mellon 

University, 5000 Forbes Avenue, 15213, Pittsburgh, PA, United 

States, ns7@andrew.cmu.edu, Guoming Lai, Francois Margot 

The valuation of the real option to store natural gas is a practically important 

problem. Traders value this option heuristically because its exact valuation is 

at odds with the high-dimensional price evolution models that they use. We 

develop a novel and tractable approximate dynamic programming method that 

coupled with Monte Carlo simulation computes lower and upper bounds on the 

value of storage. We use these bounds to benchmark heuristics used in practice. 

4 - Liquidity Risk in the Brazilian Stock Market: An Empirical 

Evidence 

Alvaro Costa, Federal University of Rio de Janeiro, 22440033, 

Rio de Janeiro, corletto@gmail.com, André Salles 

The market liquidity refers to the capacity of buying or selling an asset quickly, 

without discounts and in a big quantity. A lot of research published in the literature 

of finance analyzes the market liquidity, the liquidity risk and their interrelations 

with the firms’ cost of capital. Some of this research deals with the 

liquidity of the financial assets, in particular the liquidity risk of these assets. 

This work purpose to verify the firms’ liquidity risk from the determination of 

time series of the liquidity, of financial asset selected, were used heteroscedasticity 

models of volatility. The sample used consists stocks negotiated in the 

Brazilian stock market. 

57


� MC-28 



Applications of stochastic programming to 

the energy sector - gas 



Chair: Marte Fodstad, Department of Industrial Economics and 

Technology Management, Norwegian University of Science and 

Technology, Alfred Getz veg 3, 7491, Trondheim, Norway, 

martefo@iot.ntnu.no 

1 - Evaluating the impact of forecasting techniques on a 

stochastic optimization model for a gas sale retailer 

Juan Ramon Trapero Arenas, Management Science, Lancaster 

University, LA1 4YX, Lancaster, United Kingdom, 

j.traperoarenas@lancaster.ac.uk, Nikolaos Kourentzes, Francesca 

Maggioni 

Deregulation of natural gas market in recent years has increased the necessity 

of accurate forecasting and optimization tools. Forecasting gas consumption 

is crucial to establish optimal strategies for companies dealing with gas retail 

commercialisation. Temperature and gas consumption data collected from an 

Italian retailer are used to compare different forecasting approaches based on 

univariate and multivariate methods. The influence of the forecasting performance 

is assessed in terms of looses in the gas seller’s budget. 

2 - A stochastic portfolio optimization model for the natural 

gas supply chain 

Marte Fodstad, Department of Industrial Economics and 


Technology, Alfred Getz veg 3, 7491, Trondheim, Norway, 

martefo@iot.ntnu.no, Kjetil Midthun, Frode Rømo, Asgeir 

Tomasgard 

We present a portfolio optimization model for tactical planning in the natural 

gas supply chain. The model takes the perspective of a large producer optimizing 

its portfolio of production rights, transportation rights, take-or-pay 

contracts and spot market opportunities. The objective is to maximize profits 

while contract obligations are satisfied. Spot prices and contract obligation and 

prices are stochastic parameters. We will describe the model, show examples 

of typical model effects and results on realistic data. 

3 - Hydrogen Production Facility Network Design with 

Stochastic Wind Energy Supply and Nodal Pricing 

Jorge Barnett Lawton, PhD Student, MIT-Zaragoza International 

Logistics Program, Zaragoza Logistics Center, Calle de Bari 55, 

Portal 5, PLAZA, 50197, Zaragoza, Spain, jbarnett@zlc.edu.es, 

Mozart Menezes, Jarrod Goentzel 

We analyze the profit maximization problem for an integrated firm generating 

electricity from wind and producing hydrogen through electrolysis, operating 

on a single node of an electric transmission network under nodal pricing. We 

present the firm’s optimal actions for any given market scenario, and formulate 

the firm’s expected profit function. We then use these results to address the 

hydrogen production location/capacity problem. 

4 - Solving stochastic equilibrium problems with stochastic 

gradient methods: analysis of collaborative service 

provision in telecommunication sector 

Alexei Gaivoronski, Industrial Economics and Technology 

Management, Norwegian University of Science and Technology, 

Alfred Getz vei 1, 7491, Trondheim, Norway, 

alexei.gaivoronski@iot.ntnu.no, Denis Becker 

In this paper we develop algorithms for solution of stochastic multilevel equilibrium 

problems which belong to the family of stochastic gradient methods. 

Such problems arise in environments composed by multitude of actors engaged 

in complex relations of competition and collaboration. Numerical experiments 

confirm the efficiency of the proposed techniques. We apply this methodology 

to the analysis of markets for ICT products and services in the process of transformation 

from young markets with few actors to mature markets with large 

number of participants. 

58 

� MC-29 


8.2.11 

Algorithmic Applications of Boolean 

Functions 



Chair: Endre Boros, RUTCOR, Rutgers University, 08854, 

Piscataway, New Jersey, United States, 

Endre.Boros@rutcor.rutgers.edu 

1 - A Boolean theory of tonal signatures 

Bruno Simeone, Dept. of Statistics, University of Rome La 

Sapienza, Piazzale Aldo Moro 5, 00185, Rome, Italy, 

bruno.simeone@uniroma1.it, Isabella Lari, Gilbert Nouno, 

Malik Mezzadri 

We explore the musical concept of tonal signatures developed by the French 

jazz flute player and composer Malik Mezzadri. A tonal signature of a scale S 

is a minimal subset of notes within S that is not contained in any scale S’ different 

from S. We present a set covering model to find a smallest signature. We 

show that the signatures of a scale are the prime implicants of a suitable monotone 

Boolean function represented by a CNF. On this ground, we introduce a 

more general notion of Boolean signature, depending on a Boolean operator. 

Short pieces will be played on this new harmonic concept. 

2 - A Classification Problem and the Maximal Weight 

Archipelago Subgraph Problem 

Bela Vizvari, Industrial Engineering, Eastern Mediterranean 

University, Gazimagusa, Mersin 10, Turkey, 

vizvaribela@gmail.com, Bruno Simeone 

Motivation. The relation of two stocks can be expressed by the correlation of 

their prices. In an ideal clustering of stocks there are positively correlated ones 

in each cluster and different clusters are connected by negative correlations. 

The ideal clustering can be achieved only by neglecting some correlations. 

Archipelago Graph. A signed graph is an archipelago if all negative edges are 

connecting different components of the graph of the positive edges. 

Results. Some edges are to be deleted from a weighted signed graph being not 

an archipelago. Two questions are answered in connection with that problem: 

how high weights must be deleted and how to obtain an archipelago subgraph 

of maximal weight. 

3 - A polynomial algorithm for dualizing monotone 

Boolean functions arsining in geometry 

Khaled Elbassioni, Department 1:Algorithms and Complexity, 

Max-Planck-Institut für Informatik, Campus E1 4, 66123 , 

Saarbrücken, Saarland, elbassio@mpi-inf.mpg.de 

We consider the problem of dualizing a monotone Boolean function whose 

CNF representation corresponds to a geometric hypergraph. We show that, 

when the hypergraph admits a balanced subdivision, a recursive decomposition 

can be used to obtain efficiently all minimal terms in the DNF representation 

of the function. We apply this framework to get efficient parallel algorithms 

for dualizing monotone Boolean functions whose CNF’s are induced by a set 

of points and a set of geometric objects, such as half-spaces, hyper-rectangles 

and balls, in fixed dimension. 

4 - Minimization of quadratic pseudo-Boolean functions 

Ivo Rosenberg, Math.Stat, Université de Montréal, C.P. 6128 

suc.centre ville, H3C3J7, Montréal, Qué., Canada, 

rosenb@dms.umontreal.ca, Calvin Mbuntcha-Wuntcha 

The problem is to find the minimum value of an n-variable quadratic polynomial 

on the unit n-dimensional hypercube. For this NP-complete problem - 

which includes several classes of NP-complete problems - we propose a heuristic 

based on steepest descent curves through the interior of the hypercube. At 

a point where such a curve definitely leaves the hypercube we freeze certain 

coordinates of the polynomial to 0 or 1 and continue the steepest descent from 

the point on the reduced polynomial. Repeated application leads to a point of 

local minimum.

� MC-30 


8.2.13 

MCDA II: Health and Environment 


other Issues 





Chair: David Collier, Golders Associates, OX33 1ER, Oxford, 

United Kingdom, David_Collier@golder.co.uk 

1 - Strategic appraisal of environmental risks: a contrast 

between the UK’s Stern Review on the Economics 

of Climate Change and its Committee on Radioactive 

Waste Management 

Alec Morton, Management/ Operational Research, London 


England, United Kingdom, a.morton@lse.ac.uk, Simon Dietz 

We compare two high-profile strategic policy reviews undertaken for the UK 

government: radioactive waste management and climate change. The Stern 

Review on the Economics of Climate Change was largely an exercise in expert 

modelling within a cost-benefit framework. The Committee on Radioactive 

Waste Management, on the other hand, followed a much more explicitly deliberative 

and participative process, using MCDA to bring together scientific 

evidence and stakeholder and public values. We ask why the two reviews were 

different, and whether the differences are justified. 

2 - Lessons from Two Recent Major MCDA Projects 

David Collier, Golders Associates, OX33 1ER, Oxford, United 

Kingdom, David_Collier@golder.co.uk, Michael Egan 

The paper contrasts two projects involving substantial specialist teams. One 

was based on a conventional scored and weighted multi-attribute decision analysis. 

In the other, the MADA framework was used largely without quantification 

to structure the problem and understand the main decision drivers. We 

will discuss the lessons for the use of MCDA frameworks and stakeholder involvement 

and draw conclusions about the appropriate use of quantification, 

the structured integration of MCDA outputs with other strands of information 

and management insight, and the role of decision specialists. 

3 - Facilitated health care priority setting at a time of economic 

crisis 

Mara Airoldi, Management - OR group, London School of 

Economics and Political Science, Houghton Street, WC2A 2AE, 

London, United Kingdom, m.airoldi@lse.ac.uk, Nikolaos 

Argyris, Alec Morton 

At times of economic crisis it is particularly difficult for diverse stakeholders to 

agree on resource allocation. In this paper we present our experience to support 

the English National Health Service setting priorities and facing the challenge 

of disinvestments. The approach is decision analytic, participative and allows 

using clinical and epidemiological evidence systematically. Particular attention 

is given to the difficulty of obtaining a list of potential disinvestments from 

stakeholders on one hand, and the impracticality of evaluating all interventions 

currently provided on the other. 

4 - On the design of custom packs: grouping of medical 

disposable items for surgical procedures 

Brecht Cardoen, Vlerick Leuven Gent Management School & 

Faculty of Business and Economics, Katholieke Universiteit 

Leuven, Reep 1, B-9000, Gent, Belgium, 

brecht.cardoen@vlerick.be, Mario Vanhoucke 

In order to limit the non-operative time in the operating theater, hospitals or 

surgeons often stimulate the use of custom packs in which medical disposable 

items are grouped. In this paper, we examine how many different types of 

custom packs have to be introduced for a given population of surgeries and surgeons, 

and how these packs should be configured. We develop a mixed integer 

linear programming approach that guides these decisions, taking into account, 

for example, the resulting number of touch points or the value of waste. We 

report on our experience in applying the solution methodology for the design 

of custom packs in an average-sized hospital in Belgium. We examine the solution 

quality that is obtained and discuss the major issues we encountered during 

the configuration process. 


� MC-31 


8.2.15 

Societal Complexity and Stakeholders 



Chair: Marcos Estellita Lins, Production Engineering, Federal 

University of Rio de Janeiro, Rua Belisário Távora 80 ap 506, 

Laranjeiras, 22245-070, Rio de Janeiro, Rio de Janeiro, Brazil, 

lins@pep.ufrj.br 

1 - Multi-actor multi-criteria analysis (MAMCA) as a means 

to cope with societal complexity 

Klaas De Brucker, Faculty of Economics and Management, 

Hogeschool-Universiteit Brussel (HUB), Stormstraat 2, 

BE-1000, Brussels, Belgium, klaas.debrucker@hubrussel.be, 

Cathy Macharis 

In this contribution it will be shown that the dynamics of stakeholder management, 

in combination with the implementation of MCA, can be used to cope 

with problems of societal complexity. Special attention will be given to the 

design of the value structure that is instrumental for this purpose. Several alternative 

approaches will be contrasted, such as the traditional value structure 

versus the multi-actor value structure (used in MAMCA). Regarding the latter, 

different methods for exploring (or aggregating) stakeholder objectives will be 

discussed and illustrated using real-life case-studies. 

2 - Systemic Intervention - Methodological Pluralism in 

Managing the Problem Situations 

Slavica P. Petrovic, Faculty of Economics, University of 

Kragujevac, D. Pucara 3, 34 000, Kragujevac, Serbia, 

pslavica@kg.ac.rs 

As a theoretical, methodological and applicable development within critical 

systems thinking, Systemic Intervention is based on the ideas of process philosophy 

and the theory of boundary critique. The resulting three-stage methodology 

is focused on: critique - reflection on, and choice between, boundaries; 

judgment - judgment about which theories and methods can be most appropriate; 

and action - the implementation of methods to create - at least local 

- improvement. In managing the problem situations, the Creative design of 

methods provides a strategy for selecting, designing and mixing methods during 

intervention. 

3 - Dialogical self, Theory of Mind and Meta-cognition as a 

contribution 

Marcos Estellita Lins, Production Engineering/ Operational 

Research, Federal University of Rio de Janeiro, Rua Belisário 

Távora 80 ap 506, Laranjeiras, 22245-070, Rio de Janeiro, Rio 

de Janeiro, Brazil, estellita@pep.ufrj.br 

Validation is important to applying formal models to complex societal problems. 

since unwitting pre-decisions reflect issues that are left out of the research, 

and can be crucial to the validity of the research. This work proposes 

the use of concept or cognitive maps to express complexity and conflictive nature 

of the social problems. We have borrowed multiagency concepts from 

Dialogical Self theory and Theory of Mind in order to provide a phenomenological 

support to Problem Structuring Methods. Looking at the big picture 

helps promoting outside validity. Some applications are provided. 

4 - Recent OR Advances on Societal Complexity, stream 

Methodology of Societal Complexity 

Julien Cotret, Computer science, Lirmm, LIRMM UMR 5506 - 

CC 477 161 rue Ada 34095 Montpellier Cedex 5 France, 

montpellier, France, julien.cotret@laposte.net 

Problems encountered within the e-democracy are closely related to the web 

2.0 phenomenon. One objective of the e-democracy is the opinion construction 

and/or extraction on the web to support a political decision. One efficient way 

is to have citizens debate on subjects or texts. The first issue is the analysis of 

the produced content, the second one is to improve the participation. Being able 

to manage the flow of generated data, models and tools become a necessity. In 

this paper, we present the "Annotation" as a practical and effective model on 

which we will base our argument. 

59


� MC-32 


8.2.17 

OR in Forestry II 



Chair: Isabel Martins, Departmento de Matemática, Instituto Superior 

de Agronomia, Centro de Investigação Operacional, Tapada da 

Ajuda, 1349-017 , Lisbon, Portugal, isabelinha@isa.utl.pt 

Chair: Miguel Constantino, University of Lisbon, FCUL-DEIO-CIO, 

Bloco C2 Piso 2 Campo Grande, 1749-016, Lisbon, Portugal, 

miguel.constantino@fc.ul.pt 

1 - Assessing sustainable management in forest Mediterranean 

ecosystems in Southern Portugal. 

Brigite Botequim, Department of Forestry, The Institute of 

Agronomy of Lisbon, Tapada da Ajuda, 1349-017, Lisbon, 

Portugal, bbotequim@isa.utl.pt, Paulo Borges, Miguel 

Constantino, Jose Borges 

Addressing sustainability concerns in Mediterranean forest ecosystems management 

is a complex task. This paper focuses on techniques for oak scenario 

analysis. Both a linear and integer programming models and Decision Support 

System architecture aim at addressing the complexity of such forest management 

problems and promote its sustainability. The mathematical model includes 

objectives such as net present value, cork and timber flows and carbon 

stocks. Results are discussed for a large-scale application encompassing over 1 

million ha of cork and holm oak forest ecosystems in Southern Portugal. 

2 - Tree search for forest harvest scheduling problems 

subject to area and connectivity constraints 

Teresa Neto, Matemática, Instituto Politécnico de Viseu, R. 

Coronel Fonseca, Lote 2, Carneiria, 2005-010 Santarém, 

Santarém, Portugal, tneto@mat.estv.ipv.pt, João Pedro Pedroso, 

Miguel Constantino, Isabel Martins 

This work presents a tree search method for finding good feasible solutions, in 

reasonable times, for forest harvest scheduling problems subject to area constraints 

(on clearcuts and habitats) and connectivity constraints. For the connectivity 

constraints, an index is used. The method is a search process inspired 

in branch-and-bound, designed specifically for this problem. In each branch, a 

partial solution leads to two children nodes, corresponding to harvesting or not 

a given stand, in a given period. Pruning is based on constraint violations or 

unreachable objective values. 

3 - Integer and multiobjective programming approaches 

for modelling target volume flows in harvest scheduling 

subject to maximum area restrictions 

Maria da Conceicao Fonseca, Departamento de Estatistica e 

Investigação Operacional, Universidade de Lisboa, Faculdade de 

Ciencias and Centro de Investigação Operacional, Bloco C/6 

Campo Grande, Cidade Universitária, 1749-016, Lisboa, 

Portugal, mdfonseca@fc.ul.pt, Isabel Martins, Mujing Ye, 

Miguel Constantino, Jorge Cadima 

A typical requirement in harvest scheduling is a non-declining flow of timber 

harvested over the planning horizon. Therefore, two objectives should be 

considered: maximize the profit and at the same time obtain an even flow of 

timber harvested along the planning horizon. This goal can be achieved either 

by integer programming approaches using volume constraints to impose such 

a non-declining yield or by a multiobjective programming approach. All the 

approaches are compared from a computational point of view. 

� MC-33 


8.2.19 

Electric Vehicles 



Chair: Andrei Neboian, Chair of Logistics Management, WHU, 

Burgplatz 2, 56179, Vallendar, Germany, andrei.neboian@whu.edu 

60 

1 - Fleet renewal with electric vehicles - a real options approach 

Andrei Neboian, Chair of Logistics Management, WHU, 

Burgplatz 2, 56179, Vallendar, Germany, 

andrei.neboian@whu.edu, Stefan Spinler, Paul Kleindorfer 

We evaluate commercial fleet renewal strategies with electric vehicles (EVs) 

using real options. We show that the fleet operator can hedge against uncertainties 

of the battery price and fuel price by having an option to replace certain 

amount of diesel vehicles with EVs every year. We use a discrete time lattice 

based option model, with two underlying stochastic processes, taking into account 

technology learning via experience curve. CO2 price will be modeled as 

a Geometric Brownian motion subject to regime switching. We derive optimal 

replacement strategies which maximize total project value. 

2 - Risk analysis to Electric Vehicles Fast Charging Stations 

Jorge Borges, MIT PORTUGAL SES / IN+, INSTITUTO 

SUPERIOR TECNICO (IST) - UNIVERSIDAD DE LISBOA 

(UTL), Lisbon, Portugal, jorge.g.borges@gmail.com, Christos 

Ioakimidis 

To drive an electric vehicle (EV), it is necessary an infrastructure system that 

can provide charging when the vehicle is parked as well as ongoing charge (to 

refill the EV power in a fast process, like Fast charging stations). Fast Charging 

is intended to perform similar to a commercial gasoline service station, aiming 

to achieve a 50% charge in an EV battery in 10 to 15 minutes. But as gasoline 

stations must be refuelled, Fast charging stations must also be supplied by the 

electric grid and therefore are dependent by the electricity price. This work 

makes a risk analysis to a fast charging station. It was developed a business 

model that simulates the fast charging price from the driver’s point of view. 

3 - Assessing the Power Sector-Related Environmental 

and Cost Impacts of Plug-In Hybrid Electric Vehicles in 

Germany 

Reinhard Madlener, Faculty of Business and Economics / E.ON 

Energy Research Center, RWTH Aachen University, 

Mathieustrasse 6, 52074, Aachen, Germany, 

rmadlener@eonerc.rwth-aachen.de, Christoph Mazur 

In the coming years the diffusion of PHEV among private households will increase. 

PHEV offer individual transport services less dependent on fossil fuels, 

so that variable costs for those households and emissions will decrease. Still, 

increased demand for electricity created by PHEV will significantly influence 

the total daily grid load (and thus capacity requirements and power generation 

cost). Due to the country-specific energy mixes and charging strategies, the 

impact on total emissions (from vehicles and power plants) will vary as well. 

Our study focuses on the situation in Germany. 

� MC-34 


8.2.23 

Production and Inventory decisions with 

recycling 


Quantity 


Chair: Simone Zanoni, Dipartimento di Ingegneria Meccanica, 

Università di Brescia, via Branze 38, 25123, Brescia, Italy, Italy, 

zanoni@ing.unibs.it 

1 - The value of flexibility of different sourcing strategies 

for spare parts after end-of-production 

Rainer Kleber, Faculty of Economics and Management, 

Otto-von-Guericke University Magdeburg, POB 4120, 39016, 

Magdeburg, Germany, rainer.kleber@ovgu.de, Karl Inderfurth 

The provision of spare parts for after sales service requires high flexibility. Options 

available provide different levels of flexibility: placing a final order at 

the end-of-production is not flexible at all, extra procurement is often charged 

with high unit cost and minimum lot size requirements, and remanufacturing is 

limited by the availability of used components. In a numerical study we combine 

the above options to strategies and investigate to which extent quantity, 

time and stock related flexibility properties of the different sourcing options 

contribute to a strategy’s profitability.

2 - Multi-product economic lot scheduling problem with remanufacturing 

(ELSPR) and yield loss 

Simone Zanoni, Dipartimento di Ingegneria Meccanica, 

Università di Brescia, via Branze 38, 25123, Brescia, Italy, Italy, 

zanoni@ing.unibs.it, Laura Mazzoldi, Ivan Ferretti 

We study a variant of the multi-product economic lot scheduling problem 

with remanufacturing (ELSPR). The system includes three limited capacity 

resources: after a disassembly and testing activities, product components are 

disposed or involved in remanufacturing operations (considering a probabilistic 

success of recovery), while new products are manufactured on a separate 

line. A simple heuristic is also applied to provide near-optimal solutions to the 

related economic lot scheduling problem: numerical study shows applicability 

of the heuristic and the savings that can be obtained. 

3 - Environmental collaboration in Closed-loop Supply 

Chain with dynamic returns 

Pietro De Giovanni, Management Science, ESSEC BS, av. B. 

Hirsch, Cergy Pontoise, 95800, Cergy, France, France, 

pietro.degiovanni@essec.fr 

In a closed-loop supply chain, the players coordinate their strategies to building 

up the goodwill and exploit it for marketing and operational purposes. The 

goodwill dynamic enhances the return rate whose behavior is investigated under 

infinite time horizon. We compare the players’ strategies and outcomes under 

wholesale price and reverse revenue sharing contracts, showing that coordination 

in environmental management is successful only when returns residual 

value is sufficiently large while the inefficiency due to a two-parameter contract 

is not too damaging. 

4 - Assessing the impact on optimal production capacities 

in a closed-loop logistics system of the assumption 

that returns are stochastically independent of sales 

Ernest Benedito, Universitat Politècnica de Catalunya, Spain, 

ernest.benedito@upc.edu, Albert Corominas, Albert Corominas 

We study a production system with reverse logistics such that new and recovered 

products are indistinguishable, manufacturing and storage capacities are 

limited and where product returns depend on previous sales. With these assumptions, 

the optimal manufacturing and storage capacities are hard to calculate, 

mainly because of the relation between sales and returns. In order to 

improve the tractability of the problem, we do the calculations assuming that 

sales and returns are stochastically independent, and then we analyse how this 

hypothesis influences the outcome of a given policy. 

� MC-35 


6.2.46 

Soft OR and Problem Structuring Methods II 



Chair: Fran Ackermann, University of Strathclyde, United Kingdom, 

fran@mansci.strath.ac.uk 

1 - ’Getting messy’ with problems: teaching students 

problem structuring methods 

Fran Ackermann, Management Science, Strathclyde University, 

40 George Street, G1 1QE, Glasgow, United Kingdom, 

fran.ackermann@strath.ac.uk 

For many teaching PSMs within the constraints of a degree program can be a 

challenge. For example, effort is needed in managing expectations — particularly 

as the system doesn’t give ’a clear right answer’. And then there is the 

requirement of working ’live’ with groups - a daunting task to many! Finally 

adding the extra consideration of distance learning just exacerbates the situation. 

However PSMs clearly provide benefit to many and are a valuable part of 

an OR’ers armoury. This paper reflects on a selection of these challenges, and 

provides some insights into their management. 

2 - On Verification of Cognitive-Map-Based Models of Illstructured 

Situations 

Svetlana Kovriga, Instute of Control Sciences of Russian 

Academy of Sciences, 65, Profsouznaya st., 117997, Moscow, 

Russian Federation, mackinder@mail.ru, Nina Abramova 


The inevitable participation of experts in structuring and solving problems in 

ill-structured situations by means of cognitive maps causes human-induced 

risks. In order to reduce these risks the verification method that applies a number 

of criteria of the cognitive-map-based models validity is proposed. This 

method is illustrated by a number of examples including detection of risks of 

causal influences false transitivity. The risks of this kind are typical for cognitive 

mapping. This result refutes the stereotypical conception of universality of 

causal influences transitivity principle. 

3 - Improving Surgical Center Management of a University 

Hospital through combined Soft and Hard O.R. approaches 

Leonardo Pessoa, CASNAV - Centro de Análises de Sistemas 

Navais, Praça Barão do Ladário s/n o - Ilha das Cobras, Centro, 

20091-000, Rio de Janeiro, Rio de Janeiro, Brazil, 

lampessoa@gmail.com, Marcos Estellita Lins, Angela Silva, 

Roberto Fiszman, Milena Diniz 

This paper focuses on the Surgical Center of "HUCFF’, which is a benchmark 

among Brazilian teaching hospitals. However, low surgical production is a 

major concern for managers. The aim of this work is to point out ways to 

increase performed surgeries, yielding benefits for patients, students and also 

researchers. Soft and Hard O.R. tools are used, striving for better assessment. 

Direct experimentation could decrease system performance, exposing patients 

to negative effects. Thus, simulation has special affinity in this case. Use of 

cognitive maps include health-staff points of view, supporting simulation experiments. 

Results concern both quantitative and qualitative recommendations. 

4 - Selection of executive aircraft using tools to support 

decision: Cognitive Maps and Decision Tree 

Eliseu Zednik, COMAER, ITA, CTA, 12.246-021, São José dos 

Campos, São Paulo, Brazil, zednikff@ita.br, Mischel Carmen N. 

Belderrain 

This work is the use of Cognitive Maps and its transition to a model Multicriteria 

Decision Support or problems of the type of decision trees. This study 

focuses on the importance of problem structuring applied to problems of type 

selection Portfolios. Through this kind of tool is obtained PVF Viewpoints fundamental 

clarify on which product best meets the requirements of the desirable 

maker. Thus, the decision-making process will have a greater chance of success 

especially in scenarios typically complex, with greater potential for mental 

representation of perceived reality. 

� MC-36 


3.1.05 

Genetic Fuzzy Systems; Fuzzy Goal 

Programming 2 


Intelligence 


Chair: Fernando Gomide, DCA—FEEC—UNICAMP, 13083-970 , 

13083-970 Campinas — SP, Brazil, gomide@dca.fee.unicamp.br 

Chair: Mariano Jimenez-Lopez, Economía Aplicada I, University of 

the Basque Country, Plaza de Oñati 1, 20018, San Sebastian, Spain, 


1 - Adaptive Meta-heuristics with Genetic Fuzzy Systems 

Vitor Marques, DCA, FEEC, Unicamp, 13083-970, Campinas, 

SP, Brazil, vmarques@gmail.com, Fernando Gomide 

The paper introduces a genetic fuzzy system (GFS) to control meta-heuristics. 

The system adaptively chooses parameters values and runs simultaneously with 

the meta-heuristics. The GFS learns a fuzzy rule-base for parameter selection. 

A rule selection procedure reduces system complexity. Learning trades-off exploration 

and exploitation during search. A case study with vehicle routing 

with time windows illustrates the approach using genetic algorithm and tabu 

search. Experimental results show that the GFS improves solution quality and 

reduces users effort considerably. 

61


2 - A Web Based DSS for Improved Innovation Management 

Kemal Kilic, Faculty of Engineering and Natural Sciences, 

Sabanci University, Orhanli, Tuzla, 34956, Istanbul, 

kkilic@sabanciuniv.edu 

The webbased DSS relies on a questionnaire and generates two reports. The 

first report benchmarks the company with others in terms of various innovation 

determinants (InnoDets). The second report provides focused policy suggestions. 

Hence, only those determinants that the company has room further 

improvement, which significantly improve the innovativeness, should be identified. 

In order to determine the significance of the InnoDets a fuzzy system 

modeling based feature weighting algorithm which utilizes genetic algorithms 

is proposed. This is a rare study that ranks and rates various InnoDets. 

3 - Fuzzy Goal Programming for Portfolio Selection 

Verónica Cañal, Applied Economics, University of Oviedo, 

Avda. del Cristo s/n, 33006, Oviedo, Spain, Spain, 

vcanal@uniovi.es, Bilbao-Terol Amelia, Mar Arenas-Parra, 

Maria Victoria Rodriguez-Uria 

In this paper, we are working with the problem of portfolio selection, taking 

into account not only their returns and risk but also their ethical profile. The 

financial criteria considered are the expected return and the return performance 

relative to a pre-specified benchmark. It is supposed that these criteria are not 

exactly known and also that the investor’s preferences about ethical features of 

the portfolio are known in an imprecise way. As fuzzy logic provides an useful 

framework to deal with this kind of uncertainty, criteria are incorporated into 

Fuzzy Goal Programming model. 

4 - Heuristic Procedures in Goal Programming: An Application 

to Fuzzy Portfolio Selection Problems 

Lourdes Canos, Organizacion de Empresas, Universidad 

Politecnica de Valencia, Crtra. Nazaret-Oliva s/n, 46730, Grao de 

Gandia (Valencia), Spain, loucada@omp.upv.es, Maria J. Canos, 

Vicente Liern 

In the fuzzy portfolio selection problem, it is quite usual to start by solving 

a crisp problem whose optimal solution is subsequently improved. In our approach, 

the optimal return and risk provided by the crisp problem are considered 

as the investor’s minimum aspiration levels in a fuzzy goal programming 

problem; then, we obtain advantageous portfolios. We present a heuristic procedure 

based on genetic algorithms that allows us to work with large instances 

of the problem. We present some numerical results by using data from the 

IBEX35 index and the Spanish Stock Exchange Interconnection System. 

� MC-37 


3.1.09 

OR Applications for Public Policy 

Assesment in Developing Countries 



Chair: Claudia Rave, Energy Institute, National University of 

Colombia, AA 1027, crr 80 No 65 - 223, BL M2 of 112, 57, 

Medellín, Colombia, claudia.rave@gmail.com 

1 - Modeling approach for externality analysis of urban development 

in Medellin metropolitan area 

Claudia Rave, Energy Institute, National University of Colombia, 

AA 1027, crr 80 No 65 - 223, BL M2 of 112, 57, Medellín, 

Colombia, claudia.rave@gmail.com, Patricia Jaramillo, Marcela 

Perez, Elizabeth Zapata 

A prospective modeling approach was developed to generate origin-destination 

matrices connected to changes in urban service infrastructure elements on the 

cities. The system integrates an expert system model to estimates attraction 

travel weights for each element (facilities, industries, malls) or region (urbanization 

and densification), and re-estimates the distribution entropy model. The 

application is useful to define pattern changes due to new urban developments 

and its different impacts on the city related to externalities over environment, 

final energy use and traffic congestion 

62 

2 - Reverse logistics and environmental management in 

the oil supply chain management: A case study Peru 

Gladys Maquera, Facultad de Ingeniería y Arquitectura, 

Universidad Peruana Unión, Salida Arequipa, km. 6. 

Chullunquiani, Psje. Alfonso Ugarte Mz C-1 Lt 14, Juliaca, 

Puno, Peru, nelidagladys@yahoo.com, Iliana Gutiérrez, Jorge 

Maquera, Santos Valerio Príncipe Anticona 

This work identifies the main environmental impact of the oil chain. It is proposed 

that production cycles can be implemented by Reverse Logistics techniques, 

providing a recycling way into the market products used by the production 

processes. Therefore, a benefit it is obtained in several areas such as: 

citizenship society, saving economic natural resources of energy, employment 

generation, etc. A case study was implemented in four cities of Peru and it was 

found that legal boundaries, environmental management, logistics, storage and 

transportation of solid waste is not good enough. 

3 - Evaluation of the computer system and risks of information 

technology for Peruvian entities 

Mercedes Bustos, Lima, Volcan Cia. Minera SAA, AV.Gregorio 

Escobedo 710, Lima, Lima, Peru, mbustos@volcan.com.pe 

This study was conducted for two entities (private sector of an international 

rank and government of Peru) to evaluate their systems and risks of IT to improve 

informatics services. We used COBIT model for auditing management 

and control of information systems Maturity Model, measures of development 

degree of the processes of the organization (CMU, USA), and COSO methodology 

for assessment and risk analysis. We conclude that the the entities should 

implement the recommendations structured in a plan to improve performance 

in IT. 

4 - Modeling platform development for scenario analysis 

and decision making, LAD-t 

Claudia Rave, Energy Institute, National University of Colombia, 

AA 1027, crr 80 No 65 - 223, BL M2 of 112, 57, Medellín, 

Colombia, claudia.rave@gmail.com, Juan Esteban Restrepo, 

Fabian Giraldo, Elizabeth Zapata, Patricia Jaramillo 

A prospective platform has been developed to allow scenario planning and 

prospective indicators estimation and analysis. The platform is a metamodeling 

software development, for math models interaction trough a model coordinator 

and a central database. An application was implemented for cities occupation 

patterns modeling, configured by optimization and simulation models (GAMS, 

Vensim) and an interface integrated with GIS. The platform is versatile for 

models improvement, adding and updating and has been a powerful tool to 

support decision making and cooperative work 

� MC-38 


6.2.44 

Recent Advances in the Economics 

Supported by OR I 






Chair: Stefan Pickl, Department for Computer Science, Universität 

der Bundeswehr München, heisenbergstr. 39, 85577, 

Neubiberg-München, Bavaria, Germany, stefan.pickl@unibw.de 




1 - Flexible Capacity Strategy in an Asymmetric Oligopoly 

Market with Competition and Demand Uncertainty 

Yang Liu, Department of Logistics and Maritime Studies, 

HongKong Polytechnic University, Department of Logistics and 

Maritime Studies, The Hong Kong Polytechnic University, Hung 

Hom , Hong Kong, 852, Hong Kong, Hong Kong, 

lgt.yangliu@polyu.edu.hk, Chi To Ng

We construct an asymmetric oligopoly competition model with flexible and inflexible 

firms under demand uncertainty. All firms carry out a decision-making 

operation process spanning stages of capacity, production and pricing. Flexible 

firms can postpone production decisions until observing the actual demand, 

whereas in-flexible firms cannot. We characterize the unique competition equilibrium 

and verify the unique costing threshold to determine the optimal strategy. 

Further, increasing production cost benefits flexible firms under certain 

conditions, but is always harmful to in-flexible firms. 

2 - Forecasting by Econometric Models to Supporting 

Management 

Tunde Dobrodolac, Faculty oif Economics, Segedinski put 9-11, 

24000, Subotica, Serbia, tinde@ef.uns.ac.rs 

In contemporary environment characterized by dynamic structure of factors and 

the non-predictability of the relations existing among them, one of the central 

problem is to selected strategic goals. Forecasting and prediction as precedents 

of the process of planning include the investigation of future events. Application 

of econometric models can be successfully used to predict the future 

development of economic processes, supporting so strategic decision-making. 

3 - Transnational Interregional I-O Table Analysis for Policy 

Making 

Mei-Chen Lo, National United University, Taiwan, 

meichen_lo@nuu.edu.tw, Ozaki Toshimasa, Gwo-Hshiung 

Tzeng 

The I-O tables-received were used to consider different scenarios of the regional 

development. We analyzed FDI strategies via a scenario of increasing 

consumption up to normative standards for policy making. The authors have 

calculated the scenario of IT product price increase in the region up to the freemarket 

level. All the simulations were made to estimate multiplicative effects 

caused by upsizing industries in the regional economy. The basic picture and 

framework of the Transnational Interregional I-O Table between China and Taiwan 

as example is introduced. 

4 - Entry deterrence under scope economies 

Margarida Catalão-Lopes, DEG, Instituto Superior Técnico, Av. 

Rovisco Pais, 1049-001, Lisbon, Portugal, mcatalao@ist.utl.pt, 

Cesaltina Pires 

In this paper we develop a model where the incumbent may expand to a second 

market so as to signal the existence of scope economies and deter potential 

entry. We show that the incumbent only expands to another market when 

scope economies are large enough. Thus expansion is indeed a signal of larger 

economies of scope and for certain parameter values it leads to entry deterrence. 

We characterize the unique Perfect Bayesian Equilibrium (PBE) for the 

various parameter values and show that the PBE may involve accommodation, 

entry deterrence or a mixed strategy equilibrium. 

� MC-39 


6.2.45 

Recent Advances in Optimal Control Theory 






Chair: Lyudmila Kuzmina, Kazan Aviation Institute, Kazan State 

Technical University of A.N.Tupolev’s name, Adamuck, 4-6, 420015, 

Kazan-15, Russian Federation, Lyudmila.Kuzmina@ksu.ru 

1 - Higher Order Simple Adaptive Control 

Isaac Yaesh, Control, IMI/ASD, P.O.B. 1044/77, 47100, Ramat 

Hasharon, Israel, iyaesh@imi-israel.com, Uri Shaked 

A dynamic version of the Simple Adaptive Control (SAC) method is considered. 

The new method, referred to as Higher Order SAC (HOSAC), is derived 

by applying an augmentation method allowing the transformation of the dynamic 

output-feedback control problem to one of static control which is then 

solved using SAC. Sufficient conditions in terms of linear matrix inequalities 

are derived relating the closed-loop stability to the Almost Passivity of the 

plant. These may be applied to plants with polytopic uncertainties. Two examples 

illustrate the new controller’s superior performance. 


2 - A variational calculus based method for optimal control 

problems 

Ana García-Bouso, Statistics and Operations Research, 

Universidad Rey Juan Carlos de Madrid, C/Tulipán s/n, 28933, 

Móstoles, Madrid, Spain, ana.bouso@urjc.es, Alberto Olivares, 

Ernesto Staffetti 

Classical indirect optimal control approaches are based on the Euler-Lagrange 

necessary condition in its differential form. An indirect numerical scheme derived 

from an integral version of the Euler-Lagrange equation is used, the underlying 

idea being to reformulate the optimal control problem into an unconstrained 

variational one. Stationary points of this variational calculus equivalent 

problem are computed by means of a numerical method obtained approximating 

its true solution by virtue of a family of cubic B-spline basis functions. This 

leads to a non-linear equations system. 

3 - Optimal Control under Probability Constraint 

Guy Cohen, CERMICS, Ecole des Ponts-Paristech, Cité 

Descartes, Champs sur Marne, 6-8, avenue Blaise Pascal, 77455, 

Marne la Vallée, France, guy.cohen@mail.enpc.fr, Pierre 

Carpentier, Jean-Philippe Chancelier 

The problem is to drive a spatial vehicle to a target at the final time while 

minimizing e.g. the fuel consumption. This is a classical optimal control problem. 

However temporary stochastic failures of the engine may prevent reaching 

the target after the engine usage is recovered. Therefore, a stochastic optimal 

control problem is formulated under the constraint of ensuring a minimal probability 

of hitting the target. This problem is solved by dualizing the probability 

constraint and using an Arrow-Hurwicz stochastic algorithm. 

4 - A.M.Lyapunov theory methodology and modelling 

problems 

Lyudmila Kuzmina, Kazan Aviation Institute, Kazan State 

Technical University of A.N.Tupolev’s name, Adamuck, 4-6, 

420015, Kazan-15, Russian Federation, 

Lyudmila.Kuzmina@ksu.ru 

The work is developing approximate methods for nonlinear analysis of largescale 

systems dynamics. A.M.Lyapunov methodology, N.G.Chetayev stability 

postulate combined with asymptotic approach allow to establish the effective 

method in modelling problem of complex systems, for qualitative analysis, 

control, synthesis. Constructed approach, founded on stability/singularity postulates, 

with generalization of parametric stability, is creating efficient method 

in fundamental problems for dynamic systems with subsystems of different nature, 

for approximate theories substantiation, important for Knowledge theory. 

� MC-40 


6.2.52 

Routing in telecommunication networks 



Chair: Amaro de Sousa, Instituto de Telecomunicações, Universidade 

de Aveiro, Campus Universitário, 3810-193, AVEIRO, Portugal, 

asou@ua.pt 

1 - COSE-MS: An algorithm for SRLG diverse routing 

Teresa Gomes, Electrical and Computer Engineering, University 

of Coimbra / INESC Coimbra, Polo II da Univ, de Coimbra, 

Pinhal de Marrocos, 3030-290, COIMBRA, Portugal, 

teresa@deec.uc.pt, Luís Fernandes 

A Share Risk Link Group (SRLG) is the set of arcs in the network that share a 

common physical resource subject to failure. The problem of finding a SRLG 

diverse path pair is NP-Complete. We will present a short revision algorithm 

COSE for (min-min) SRLG diverse routing. Then we will propose algorithm 

COSE-MS, for solving the min-sum problem. This is achieved by introducing 

Suurballe algorithm and the Modified Suurballe’s Heuristic in COSE. Finally 

we will compare COSE-MS performance with an algorithm which enumerates 

SRLG diverse paths, by non decreasing cost of their total (additive) cost. 

63


2 - Algorithms for Inverse-Multiplexed Routing with Minimal 

Capacity Requirements and Differential Delay Accumulation 

João Santos, Networks and Systems, DWDM R&D, Nokia 

Siemens Networks Portugal S.A., Rua Irmaos Siemens, 1, 

Alfragide, 2720-093, Amadora, Portugal, joao.santos@nsn.com, 

João Pedro, Paulo Monteiro, João Pires 

The combination of virtual concatenation with multipath routing in the optical 

transport network enables substantial capacity savings in comparison with 

single-path approaches. However, routing over paths with different propagation 

times introduces differential delay between the concatenated streams. Hence, 

the authors propose and evaluate an integer linear programming model and 

three other heuristic methodologies for solving the joint routing problem for 

virtually-concatenated traffic by simultaneously minimizing the maximum link 

load and the accumulated differential delay in the network. 

3 - Infeasibility in Inverse Shortest Path Routing Problems 

Mikael Call, Department of Mathematics, Linköping University, 

Linköpings universitet, 581 83, Linköping, Sweden, 

mikael.call@liu.se, Kaj Holmberg 

Traffic engineering in IP networks often requires that all traffic is routed in accordance 

with a shortest path routing (SPR) protocol, e.g. OSPF or IS-IS. This 

implies that some routing patterns are not eligible. We analyze LP models of an 

associated inverse SPR subproblem. This yields a description of infeasibility in 

terms of combinatorial structures formed by tentative routing patterns. Further 

examination of the inverse SPR problem and the LP models yields additional 

insights, e.g. the extremal structure, graph search (separation) algorithms and 

alternative problem formulations. 

4 - Models for Optimal Survivable Routing with a Minimum 

Number of Hops: Comparing Disaggregated with Aggregated 

Models 

Pedro Patrício, Mathematics, University of Beira Interior, Rua 

Fernando Henriques da Cruz, 9, 6200-093, Covilhã, Portugal, 

pedrofp@mat.ubi.pt, Luis Gouveia, Amaro de Sousa 

In this work we address two traffic engineering problems. We determine how 

traffic commodities must be routed in order to maintain an optimal network 

performance, minimizing the average or the maximum number of routing hops 

in each commodity, while ensuring some desired survivability guarantees. We 

study disaggregated and aggregated Integer Linear Programming models for 

both problems, the theoretical relationships between the two classes of models 

as well as an empiric comparison of their effectiveness to solve the proposed 

traffic engineering problems. 

� MC-41 


3.1.06 

Topics in Revenue Management 



Chair: Nishant Mishra, Management Science and Operations, 

London Business School, Regents Park, NW1 4SA, London, United 

Kingdom, nmishra.phd2003@london.edu 

1 - Allocation of Returned Products among Different Recovery 

Options through an Opportunity Cost based Dynamic 

Approach 

Oznur Ozdemir, Rotterdam School of Management Decision and 

Information Sciences, Erasmus University, Erasmus University 

Rotterdam School of Management Burgemeester Oudlaan 50, 

Room T09-44, 3062PA, Rotterdam, Netherlands, 

OOzdemir@rsm.nl, Meltem Denizel, Mark Ferguson 

In this paper we investigate the disposition decisions of a recovery firm. We 

consider a make-to-order recovery environment with two disposition options: 

refurbishing and harvesting. We formulate this problem as a multi-period LP 

model and suggest two approaches: (1) a static approach that uses the LP solution 

directly and (2) a dynamic one that employs a revenue management method 

of bid-price controls. We present a novel reformulation in the form of a transportation 

problem and develop a one-pass optimal solution procedure. Our 

numerical analysis shows the superiority of the dynamic approach. 

64 

2 - A model for competition in network revenue management 

Nishant Mishra, Management Science and Operations, London 

Business School, Regents Park, NW1 4SA, London, United 

Kingdom, nmishra.phd2003@london.edu 

We propose a model for oligopolistic capacity competition between several airlines 

offering multiple products over a network. Using a variational inequality 

approach, we show the existence and uniqueness of Nash Equilibrium under 

certain regularity conditions. We then present numerical results for some small 

and large network examples and derive insights on how competition can affect 

quantities and prices for various players. Finally, we also show how the 

model can be used to decide the value, with regards to making better revenue 

management decisions, of mergers between airlines. 

3 - Managing the Inventories of Perishable goods with 

Controllable Shelf Lives 

Qing Li, Dept. of ISMT, Hong Kong University of Science and 

Technology, Clear Water Bay, Kowloon, Hong Kong, 

imqli@ust.hk 

The shelf lives of perishable goods can be determined by purchasing and maintenance 

decisions. If the inventories are depleted under FIFO, then the retailer 

may order/maintain items of different life times to balance the cost against the 

risk of overage. Such strategy should never be used if consumer sensitivity to 

freshness is high. The retailer is less likely to order items with a longer life time 

under LIFO than under FIFO. Our numerical studies demonstrate merits of dynamic 

coordination. Consumer behavior adds a new dimension to the research 

in perishable inventory control. 

4 - Explaining Stickiness in Prices of Non-Perishable 

Goods under Dynamic Pricing Policies 

Gunnar Feldmann, Industrial and Systems Engineering, Texas 

A&M University, Zachry Engineering Center, 3131 TAMU, 

77843, College Station, TX, United States, 

gfeldmann@tamu.edu, Abhijit Deshmukh 

The use of dynamic pricing is expanding beyond its initial application for revenue 

management of perishable goods in service industries. Empirical research 

shows that the price trajectories under dynamic pricing are qualitatively different 

in perishable and non-perishable goods. Prices tend to be "sticky’ for 

non-perishable goods, leading to high-low pricing policies. In this paper we 

develop optimal price setting and price change models that consider menu costs 

and reference prices to explain why pricing policies for non-perishable goods 

converge to a high-low pattern. 

� MC-42 


3.1.07 

Solution Algorithms for Bilevel Problems 




Chair: Ayalew Getachew Mersha, Optimization and Optimal control 

group, Austrian Academy of Sciences, Johann Wilhelm Klein Strasse 

9, 4040 , Linz, Upper Austria, Austria, ayalew.mersha@oeaw.ac.at 

1 - Bilevel Stochastic Optimization Approach to Natural 

Gas Cash-Out Problem 

Vyacheslav Kalashnikov, Systems & Industrial Engineering 

(IIS), ITESM, Campus Monterrey, Ave. Eugenio Garza Sada 

2501 Sur, 64849, Monterrey, Nuevo Leon, Mexico, 

kalash@itesm.mx, Gerardo Perez, Nataliya Kalashnykova, 

Asgeir Tomasgard 

A stochastic formulation of the natural gas cash-out problem is given in form 

of a bilevel multi-stage stochastic programming model with recourse. After reducing 

the original formulation to a bilevel problem, a stochastic scenario tree 

is defined, and time series forecasting is used to produce stochastic values for 

data of gas price and demand. A new solution algorithm based upon a reduction 

of the bi-level programming problem to a single-level one, is described. 

Numerical experiments were run to compare the stochastic solution with the 

perfect information and the expected value solutions.

2 - Global optimization of quadratic bilevel problems 

Oleg Khamisov, Applied mathematics, Institute of Energy 

Systems, Lermontov str. 130, 664033, Irkutsk, Russian 

Federation, mis@isem.sei.irk.ru 

We study qudratic bilevel programming problem in optimistic sense. The leader 

problem can be nonconvex quadratic but the follower problem must be convex 

quadratic. The approach based on quadratic convex and concave support functions 

and developed earlier for ordinary (i.e. one level) nonconvex quadratic 

problem with nonconvex quadratic constraints is used. The underlying procedure 

is of the branch and bound type. Bounding, branching rules and convergence 

conditions are given together with preliminary computational results. 

3 - Computing the Pareto frontier of a bi-objective bilevel 

linear problem using a multiobjective mixed-integer 

programming algorithm 

Maria João Alves, Faculty of Economics, University of Coimbra 

/ INESC-Coimbra, Av. Dias da Silva, 165, 3004-512, Coimbra, 

Portugal, mjalves@fe.uc.pt, Stephan Dempe, Joaquim Judice 

In this work we study the bilevel linear programming problem with multiple 

objective functions at the upper level and a single objective at the lower level. 

We examine some properties of the problem, in particular the bi-objective case, 

and propose a methodological approach based on its reformulation as a multiobjective 

mixed 0-1 linear programming problem. A multiobjective reference 

point algorithm is used, which enables to characterize the whole Pareto frontier 

in the bi-objective case. Illustrative numerical examples are presented to show 

the viability of the proposed methodology. 

� MC-43 


8.2.02 


Stream: Algorithmic Decision Theory [c] 


Chair: Federico Della Croce, Automatica e Informatica, Politecnico 

di Torino, Corso Duca degli Abruzzi 24, 10129, Torino, Italy, 

federico.dellacroce@polito.it 

1 - Pareto-Search in Discrete Vector Optimization Problems. 

Walter Habenicht, Management, University of Hohenheim, Lst. 

fuer Industriebetriebslehre(510A), D-70593, Stuttgart, Germany, 

habenich@uni-hohenheim.de 

In this paper we deal with discrete vector optimization problems with large sets 

of efficient solutions. We assume that the efficient set has been identified and 

the non dominated set has been stored in a special data structure called quad 

tree. In order to organize efficiently a searching process in outcome space we 

discuss different neighborhood definitions. These neighborhoods differ in the 

computational complexity of identifying the neighbors and in their ability to 

avoid suboptimal solutions. 

2 - An empirical study of the core concept for the Multidimensional 

Knapsack Problem 

Andre Amaral, Instituto Superior Técnico, CEG-IST, Technical 

University of Lisbon, 1049-001, Lisbon, Portugal, 

andre.r.s.amaral@ist.utl.pt, José Rui Figueira 

The core concept of the 0-1 knapsack problem is well known. An 

interesting research question is how to suitably extend this concept 

to the Multidimensional Knapsack Problem (MKP) because, for the 

MKP, it is not obvious how to characterize the efficiency measure of 

the items. In a paper by Puchinger et al. (available from the 

link: http://www.ads.tuwien.ac.at/publications/bib/pdf/puchinger-06.pdf) several 

tentative definitions of efficiency measures for the MKP are computationally 

tested. In our work, new definitions of efficiency measures are presented 

for the MKP, and we will discuss the results of our extensive computational 

tests using the new definitions. 

3 - A strong branching approach for the multi-dimensional 

knapsack problem 

Federico Della Croce, Automatica e Informatica, Politecnico di 

Torino, Corso Duca degli Abruzzi 24, 10129, Torino, Italy, 

federico.dellacroce@polito.it, Andrea Grosso 


We consider the 0/1 multi-dimensional knapsack problem and discuss the performances 

of a heuristic procedure particularly suitable for a parallel computing 

environnment embedding variable fixing, core problem approach and 

a strong branching scheme based on the reduced costs of the corresponding LP 

relaxation solution value. The proposed approach is compared to the literature 

results on the well known OR-Library multi-dimensional knapsack problem 

benchmarks instances improving several best known lower bounds. 

4 - Finding Nadir Points for Multi-criteria Integer Programming 

Problems 

Banu Lokman, Industrial Engineering Department, METU, 

Middle East Technical University, Industrial Engineering 

Department, 06531, Ankara, Turkey, banutuna@ie.metu.edu.tr, 

Murat Koksalan 

Finding the worst possible value for each criterion among the set of efficient 

solutions has important uses in multi-criteria problems. Such points are called 

nadir points. It is not straightforward to find the nadir points, especially for 

large problems with more than two criteria. We develop an exact algorithm to 

find the nadir values for multi-criteria integer programming problems. We also 

find bounds with performance guarantees. We demonstrate that our algorithms 

work well in our experiments on Multi-criteria Knapsack Problems. 

� MC-44 


8.2.03 

Portfolio Decision Analysis III 



Chair: Jeffrey Keisler, Management Science & Information Systems, 

University of Massachusetts Boston, 100 Morrissey Blvd, M/5-249, 

02125, Boston, MA, United States, jeff.keisler@umb.edu 

1 - Does the decision analysis match the portfolio? 

Jeffrey Keisler, Management Science & Information Systems, 

University of Massachusetts Boston, 100 Morrissey Blvd, 

M/5-249, 02125, Boston, MA, United States, 

jeff.keisler@umb.edu, Jeffrey Stonebraker 

Decision quality says the most detailed analysis should occur where it adds the 

most value. This case study of a DA—using pharmaceutical company analyzes 

data obtained on a portfolio. Measurements of the characteristics of the portfolio 

and its partitions reveal patterns and relationships. These patterns suggest 

certain sub-portfolios would benefit more from greater analytic detail. Using 

the number of scenarios included as a measure of analytic detail, we considered 

whether the analytic detail correlated with the predicted benefit. Results 

support the hypothesis. 

2 - Hierarchical portfolio decision analysis for developing 

corporate strategies 

Kevin Bossley, Catalyze Ltd, SO21 2LL, Winchester, United 

Kingdom, kevinbossley@catalyze.co.uk 

Strategy planning is a classic portfolio decision analysis problem with organisations 

trading off competing objectives with hundreds of projects and initiatives. 

We present a case study where a bottom-up hierarchical approach is used to 

develop a balanced strategy for a multinational corporation. MCDA decision 

conferences are held with individual businesses and the results of these are 

combined in a final workshop to agree the corporate strategy which balances 

short and long term objectives with full support of the businesses. 

3 - Valuing business benefits more consistently 

Santiago Castro, Research, Cogentus, Soane Point„ 6-8 Market 

Place, RG1 2EG, Reading, Berkshire, United Kingdom, 

santiagocastro@yahoo.com 

A popular prioritisation methodology is to score the benefits of each option 

and divide by the costs to generate a value for money (VFM) ratio. However 

a scope insensitivity (SI) bias often occurs: lower cost options have a higher 

VFM than more costly ones, and receive a higher ranking in the prioritised 

list. We investigate whether SI can be alleviated by applying transformations 

to the benefit values. We have applied transformations to two resource allocation 

models. The benefit scores are raised to several different exponents with 

the intention of achieving a wider spread of the scores. 

65


� MC-46 


8.2.14 

OR on Environmental Networks and 

Management 









1 - Species specific connectivity in reserve-network design 

Leonor S.Pinto, Matematica, ISEG/UTL - cemapre, Rua do 

Quelhas, 6, 1200-781 , Lisboa, Portugal, lspinto@iseg.utl.pt, J. 

Orestes Cerdeira, Raul Brás, Mar Cabeza, Kevin Gaston 

Connectivity is a critical feature to ensure sustainability of species needs in 

reserve-network design. We define the following model to incorporate speciesspecific 

connectivity. For each species s, a representation target t_s and a graph 

G_s describing its dispersal pattern are given. Solutions are sets of sites that induce, 

for each species s, a connected component of G_s of at least t_s sites. We 

give an integer cutting algorithm to achieve minimum size solutions and lower 

bounds on their sizes. Heuristic methods are also presented, and computational 

experiments are reported. 

2 - Gene-Environment Networks under Ellipsoidal Uncertainty 

Erik Kropat, Department of Computer Science, Universität der 


Neubiberg, Germany, erik.kropat@unibw.de, Gerhard-Wilhelm 

Weber, Selma Belen 

We consider gene-environment networks under ellipsoidal uncertainty and the 

corresponding set-theoretic regression. Functionally related groups of genes 

are identified by clustering techniques. The regulatory system of clusters is 

determined by affine-linear coupling rules and explicit representations of the 

uncertain multivariate states are obtained by ellipsoidal calculus. Various settheoretic 

models are introduced for an estimation of the unknown parameters. 

We discuss the solvability by semidefinite programming and conclude with research 

challenges. 

� MC-47 


8.2.16 

OR in Oil Sector II 

Stream: OR in Oil Sector 


Chair: Irina Dolgopolova, Economics and Administrative Sciences, 

Middle East Technical University, Odtu Kent, Konuk Evi 1, B Block, 

108, Ankara, Turkey, 06420, Ankara, Turkey, 

irina.dolgopolova@gmail.com 

1 - The Development of a Decision Support System for 

Management of Petroleum Hydrocarbon Contamination 

in Soil 

Aisha Bello-Dambatta, School of Engineering, Computing and 

Mathematics, University of Exeter, Harrison Building, New Park 

Road, University of Exeter, EX4 4QF, Exeter, Devon, United 

Kingdom, ab353@exeter.ac.uk, Akbar Javadi 

Land contamination is a big environmental and infrastructural problem, the 

most common of which is petroleum hydrocarbon contamination. This could 

be exposed to humans via different routes, potentially causing harm on human 

health. We present a Web-based knowledge based decision support system for 

the risk assessment and management based on the UK framework for evaluating 

these risks. The risk assessment component is developed as an expert 

system, and the decision-support component is developed multi criteria decision 

analysis, for the site specific appraisal of remedial options. 

66 

� MC-48 


8.2.04 

Green vehicle routing and scheduling 

Stream: Optimization for Sustainable Development 


Chair: Nora Touati Moungla, LIX, Ecole Polytechnique, École 

polytechnique„ Laboratoire d’informatique (LIX), 91128, Palaiseau, 

Cedex, France, France, touati@lix.polytechnique.fr 

1 - Strategic Planning for Disaster Recovery with Stochastic 

Last Mile Distribution 

Pascal Van Hentenryck, Department of Computer Science, 

Brown University, Box 1910, 02912, Providence, RI, United 

States, pvh@cs.brown.edu, Carleton Coffrin 

We consider the single commodity allocation problem (SCAP) for disaster recovery. 

SCAPs are complex stochastic optimization problems that combine 

resource allocation and warehouse routing. This work formalizes the specification 

of SCAPs and introduces a novel multi-stage hybrid-optimization algorithm 

that utilizes the strengths of mixed integer programming, constraint 

programming, and large neighborhood search. The algorithm is validated on 

hurricane disaster scenarios generated using state-of-the-art disaster simulation 

tools and is deployed to aid federal organizations in the United States. 

2 - A Bi-objective Vehicle Routing and Scheduling Problem 

for Planning Hazardous Materials Distribution 

Konstantinos Androutsopoulos, Management Science and 

Technology, Athens University of Economics and Business, 

Evelpidon 47A, 11362, Athens, -, Greece, kandro@aueb.gr, 

Konstantinos Zografos 

This paper presents the formulation of the hazardous materials distribution 

problem as a bi-objective vehicle routing and scheduling problem on a network 

with time dependent travel time and risk attributes. The travel speed on any link 

of the underlying network is assumed piecewise linear function of the time and 

an efficient procedure is proposed for calculating the corresponding link travel 

time. A heuristic algorithm is proposed based on the weighted sum method, 

where the bi-objective problem is decomposed to a series of single-objective 

instances. Results regarding the performance of the solution algorithm on a set 

of test problems are also reported. 

3 - About green vehicle routing and scheduling problem 

Nora Touati Moungla, LIX, Ecole Polytechnique, École 

polytechnique„ Laboratoire d’informatique (LIX), 91128, 

Palaiseau, Cedex, France, France, touati@lix.polytechnique.fr, 

Vincent Jost 

The vehicle routing and scheduling problem has been studied with much interest 

within the last four decades. We explore the existing literature dealing with 

the vehicle routing and scheduling problem with environmental issues in order 

to provide an overview of the characteristics of problems that have been investigated 

within this field and how these problems are treated using combinatorial 

optimization tools. We focus particularly to the transportation of hazardous materials 

problems, time dependent and dynamic vehicle routing and multi-modal 

routing problems. 

4 - Modeling Sustainable Traffic Assignment Policies Using 


Semih Yalcindag, Industrial Engineering, Sabanci University, 


semihyal@su.sabanciuniv.edu, S. Ilker Birbil, Orhan Feyzioglu, 

Orhan Ilker Kolak, Nilay Noyan 

We develop bilevel programming models with elastic demand to determine optimal 

sustainable transportation policies. The upper level problem represents 

the decisions of policy makers based on sustainability measures, while the 

lower level represents the route choice decisions of travelers. In particular, 

we determine the emission functions in terms of traffic flow to incorporate the 

gas emission amounts into the proposed models and consider several policies to 

reduce the gas emission; toll pricing, zone charging and capacity enhancement. 

We present numerical results on a test network.

Monday, 14:00-15:20 

� MD-01 


Aula Magna 




Chair: Stefan Nickel, Institut fuer Operations Research, Universitaet 

Karlsruhe, Geb 11.40, 76128, Karlsruhe, Baden-Wuertemberg, 

Germany, Stefan.Nickel@kit.edu 

1 - Location problems on networks with routing 

Elena Fernandez, Statistics and Operations Research, Technical 

University of Catalonia, Campus Nord, C5-208, Jordi Girona, 

1-3, 08034, Barcelona, Spain, e.fernandez@upc.edu 

Various location problems on networks involve routing decisions related to 

sending flows between pairs of nodes, using the located facilities as final or 

intermediate nodes. In addition to the location of the facilities, these problems 

involve essential aspects of network design, as the arcs to be used must also 

be decided. This issue is one extra source of difficulty which, in practice, is 

reflected in formulations with a huge number of variables, highly inefficient 

even for relatively small size instances. In this talk we discuss some problems 

of this type, including some hub location models, and their relation to network 

design. Alternative formulations are presented and recent advances leading to 

successful solution methods are introduced. 

� MD-02 


3.2.14 

Algorithms for planning and scheduling 

problems 



Chair: Eleni Hadjiconstantinou, Business School, Imperial College 

London, South Kensington Campus, SW7 2AZ, London, United 

Kingdom, e.hconstantinou@imperial.ac.uk 

1 - A Multi-load AGV dispatching problem in seaport container 

terminals 

Evelina Klerides, Tanaka Business School, Imperial College 

London, South Kensington, SW7 2AZ, London, United 

Kingdom, e.klerides@imperial.ac.uk, Eleni Hadjiconstantinou 

We consider the Automated Guided Vehicle (AGV) dispatching problem in 

seaport container terminals where vehicles are allowed to carry more than one 

container at a time. The main focus of the study is on dual-load AGVs which 

may carry either two 20-ft containers or one 40/45-ft container. We propose 

a mathematical formulation for the corresponding optimization problem which 

minimizes a defined dispatching "cost’ under a rolling horizon approach. Preliminary 

computational experimentation is performed using different scenarios 

found in the literature. 

2 - A Branch-and-Price Algorithm for The Bin Packing 

Problem with Conflicts 

Joe Naoum-Sawaya, Management Sciences, University of 

Waterloo, 200 University Avenue W, N2L 3G1, Waterloo, 

Ontario, Canada, jnaoumsa@uwaterloo.ca, Samir Elhedhli 

We provide a branch-and-price algorithm for the Bin Packing Problem with 

Conflicts, a variant of the classical bin packing problem that has major applications 

in scheduling and resource allocation. First, we use a branching rule that 

matches the conflicting constraints, preserving the structure of the subproblems 

after branching. Second, maximal clique valid inequalities are generated based 

on the conflicting constraints and are added to the subproblems. The algorithm 

is tested on a standard set of problems. Numerical results indicate its efficiency 

and stability. 

EURO 24 - Lisbon 2010 MD-03 

3 - Project scheduling under multiple resource constraints 

- A case study in the Portuguese Navy 

Humberto Duarte Afonso, DAGI-DEIO, Marinha de Guerra 

Portuguesa, Praça do Municipio (Edificio de Marinha), 

1149-001, Lisboa, Portugal, hdafonso@hotmail.com, Joao 

Telhada 

Project scheduling with resource constraints has been a field widely studied. 

However, constraints typically refer to some uniform pattern. Problems where 

the requirements on the resources vary along the execution of a task have to be 

tackled differently. Some linear programming approaches are used and compared 

with those obtained by adapting the problem. Computational results are 

presented, to evaluate the relative quality of the lower bounding techniques introduced. 

This work is based on a real-life project scheduling in the Portuguese 

Navy that is used in the maintenance of the fleet. 

4 - Multi-run heuristic algorithm for conflict-free agv 

scheduling in container terminals 

Fabio Furini, DEIS, Universita’ di Bologna, Viale Risorgimento 

2, 40136, Bologna, Italy, Italy, fabio.furini@unibo.it, Eleni 

Hadjiconstantinou, Evelina Klerides 

The Automated Guided Vehicle (AGV) Scheduling algorithm aims at constructing 

conflict-free schedules for AGVs, subject to operational constraints, 

so that the deviation from an ideal timetable is minimised. We adapt a well 

known Train Timetabling approach to the specific case of AGV Scheduling 

in ports. Special attention is given to the occupancy of the cranes in order 

to obtain feasible loading and unloading sequences. Computational results on 

benchmark test instances, taken from the literature, show the efficiency of the 

proposed approach. 

� MD-03 


3.2.15 

Dynamic routing problems 



Chair: Philippe Lacomme, Laboratoire LIMOS, Université Blaise 

Pascal, BP 10125, 63173, Aubière Cedex, France, 

lacomme@sp.isima.fr 



1 - Parameter tuning for dynamic optimization. Application: 

The Dynamic Pickup and Delivery Problem with 

Time Windows 

Jawad Omari, Programmes, KADDB, Amman, Amman, Jordan, 

jawad82@batelco.jo 

A hybrid metaheuristic based on Variable Neighborhood Search, Tabu Search, 

and Guided Local Search is used to solve the Dynamic Pickup and Delivery 

Problem with Time Windows. The study focuses on how to change the search 

parameters in order to adapt to the dynamic nature of the problem; meaning, at 

the beginning of the working day, parameters are set to minimize the number of 

vehicles and the total distance traveled, later on towards the end of the working 

day, parameters are set to accommodate for the largest number of new requests. 

Li & Lim benchmark data sets will be used for comparison. 

2 - Dynamic Vehicle Routing for Demand Responsive 


Jorge Pinho de Sousa, Faculdade de Engenharia da Universidade 

do Porto / INESC Porto, Campus da FEUP, Rua Dr. Roberto 

Frias, 4200-465, Porto, Portugal, jsousa@inescporto.pt, Rui 

Gomes, Teresa Galvão Dias 

Providing quality public transportation may be very expensive in low demand 

scenarios. Demand Responsive Transportation (DRT) systems address this 

problem by providing routes and frequencies that vary according to actual demand. 

DRTs extend the "classical’ Vehicle Routing Problems in a number of 

ways, and involve multiple, conflicting objectives. To obtain an approximation 

of the Pareto solution set, we have designed metaheuristic approach that starts 

with a greedy, random construction of a feasible route. Preliminary computational 

results on randomly generated instances look quite promising. 

67

MD-04 EURO 24 - Lisbon 2010 

3 - Solving the Vehicle Routing Problem with Stochastic 

Demands using the SimuRial Procedure 

Javier Faulin, Department of Statistics and OR, Public University 

of Navarre, Los Magnolios Builing. First floor, Campus 

Arrosadia, 31006, Pamplona, Navarra, Spain, 

javier.faulin@unavarra.es, Angel A. Juan, Daniel Riera, Scott 

Grasman, Ricardo Solar 

The VRP with Stochastic Demand is a NP-hard problem in which an expected 

feasible solution could become infeasible. Our approach SimuRial is focused 

on reducing the probability of occurrence of such undesirable situations using 

Reliability and Availability concepts in order to build solutions with a low 

probability of suffering a route failure. This is basically attained by constructing 

routes in which expected demand will be somewhat lower than the vehicle 

capacity. Thus, the idea is to keep a certain amount of vehicle capacity surplus 

while designing the routes, so that if final routes’ demands exceed their expected 

values up to a certain limit, they can be satisfied without having a route 

failure. 

� MD-04 


3.2.13 

Job Shop scheduling with metaheuristics 



Chair: Ibrahim Osman, Operations & Information Management, 

American University of Beirut, Business School, Bliss Street„ 

11-0236, Beirut, Lebanon, Ibrahim.osman@aub.edu.lb 

Chair: Maurizio Faccio, Department of Innovation in Mechanics and 

Management, University of Padova, Italy, maurizio.faccio@unipd.it 

1 - Local search with a job-insertion based neighborhood 

for the job shop with setup times 

Reinhard Bürgy, Dept of Informatics, University of Fribourg, Bd 

de Pérolles 90, 1700, Fribourg, Switzerland, 

reinhard.buergy@unifr.ch, Heinz Gröflin 

We consider the job shop problem with sequence-dependent setup times. The 

objective is to schedule the operations in order to minimize the makespan. 

A feasible neighborhood, based on a job-insertion framework, is constructed 

where a critical operation is moved together with some other operations whose 

moves are "implied’. The neighborhood is shown to be opt-connected and is 

implemented in a tabu search method. Numerical results are compared to recent 

benchmarks (Balas et al. 2008, Artigues and Feillet 2008) and show that 

the method is competitive. 

2 - Scheduling in a Job Shop Production System with multiple 

parallel Machines and multiple Products in case of 

Unknown Process Cycle 

Maurizio Faccio, Department of Innovation in Mechanics and 

Management(DIMEG-Padova), University of Padova, Via 

Venezia 1, Padova, Italy, maurizio.faccio@unipd.it, Anna Azzi, 

Alessandro Persona, Fabio Sgarbossa 

A job shop production system processes after set up batches of products in 

some specific sequence. In the modern production the product’s process cycle 

is not always known from the beginning. The paper faces scheduling in a 

flexible production system with unknown production cycle, multiple productsmachines,with 

a multi-objective approach, to minimize the makespan and to 

improve the machines utilization. A metaheuritic optimization model is described 

and the obtained results from an industrial application compared with 

the classical scheduling priority rules and with Johnson’s heuristics. 

3 - A Hybrid Shifting Bottleneck-Tabu Search Heuristic for 

the Job Shop Total Weighted Tardiness Problem 

68 

Kerem Bulbul, Manufacturing Sys. & Industrial Eng., Sabanci 

University, Faculty of Engineering and Natural Sciences, 

Orhanli, Tuzla, 34956, Istanbul, Turkey, bulbul@sabanciuniv.edu 

We study the job shop scheduling problem with the objective of minimizing the 

total weighted tardiness and propose a hybrid shifting bottleneck (SB) - tabu 

search (TS) algorithm. In terms of the SB, the proposed TS replaces the reoptimization 

step by optimizing the total weighted tardiness for partial schedules 

in which some machines are currently assumed to have infinite capacity. In the 

context of the TS, the SB features a long-term memory which helps to diversify 

the local search. We exploit this synergy and demonstrate the effectiveness of 

the algorithm on standard benchmark instances. 

� MD-05 


3.2.16 

Statistics 



Chair: Leonidas Pitsoulis, Mathematical and Physical Sciences, 

Aristotle University of Thessaloniki, Engineering School, 

Department of Mathematical and Physical Sciences, 54124, 

Thessaloniki, pitsouli@gen.auth.gr 

Chair: Francisco Aparisi, Estadistica e Investigacion Operativa 

Aplicadas y Calidad, Universidad Politecnica de Valencia, Camino de 

Vera s/n, 46022, Valencia, Valencia, Spain, faparisi@eio.upv.es 

1 - A Fast Algorithm for Robust Regression 

Leonidas Pitsoulis, Mathematical and Physical Sciences, 

Aristotle University of Thessaloniki, 51124, Thessaloniki, 

Greece, pitsouli@gen.auth.gr 

The presence of groups of outliers makes linear regression a difficult problem. 

A new estimator for robust regression is introduced, called Penalised Trimmed 

Squares (PTS) estimator, which is defined by a Quadratic Mixed Integer programming 

(QMIP) problem. Due to the high computational complexity of the 

resulting QMIP problem, exact solutions for moderately large regression problems 

is infeasible. In this work we establish the theoretical properties of the 

PTS estimator and propose an approximate algorithm called Fast-PTS to compute 

the PTS estimator for large data sets efficiently. 

2 - Optimization of a Set of Multiple Poisson Statistical 

Quality Control Charts 

Francisco Aparisi, Estadistica e Investigacion Operativa 

Aplicadas y Calidad, Universidad Politecnica de Valencia, 

Camino de Vera s/n, 46022, Valencia, Valencia, Spain, 

faparisi@eio.upv.es, Eugenio Epprecht 

Some industrial applications require controlling simultaneously several Poisson 

variables. Each of the variables is based on the sum of a common Poisson 

variable and an independent variable. Each chart has a probability of false 

alarm (in-control ARL) and a probability of detecting the shift in the variable 

(out-of-control ARL). An optimization is carried out using Genetic Algorithms 

and consists of obtaining the control limits for the multiple charts given an incontrol 

ARL, minimizing the time to detect the shift in the process, minimizing 

the out-of-control ARL of the set of Poisson charts 

3 - An Efficient Optimization Method for Revealing Local 

Optima of Projection Pursuit Indices 

Larabi Marie-Sainte Souad, computer science, Toulouse1 

Capitole University, 2 rue du doyen Gabriel Marty, Manufacture 

des Tabacs ME310, 31042, toulouse, France, 

larabi.souad@yahoo.fr, Alain Berro, Anne Ruiz-Gazen 

In order to represent graphically multidimensional data in statistics, we can use 

projection pursuit methods which look for structures (groups or outliers), by 

optimizing a function called projection index (PI). To determine these possible 

structures, we must choose optimization method capable to find the global 

optimum of PI and the local optima susceptible to reveal these structures. We 

suggest using a metaheuristic called Tribes, a hybrid PSO method, which does 

not ask for parameters to settle and provoke premature convergence to local 

optima.

� MD-06 


8.2.30 

DEA Methodology IV 



Chair: Armando Milioni, Departamento de Organizacao, Instituto 

Tecnologico de Aeronautica, CTA ITA IEMB, 12228-900, Sao Jose 

dos Campos, Sao Paulo, Brazil, milioni@ita.br 

1 - Weight restrictions in Data Envelopment Analysis using 

Multicriteria Tools 

Lidia Angulo-Meza, Materials Science, Universidade Federal 

Fluminense, Av.Dos Trabalhadores 420 Vila Santa Cecilia, 

27255125, Volta Redonda, Rio de Janeiro, Brazil, 

lidia@metal.eeimvr.uff.br, João Carlos Soares de Mello, Lucas 

Oliveira 

This work illustrates the use of weights obtained with a multicriteria method 

for decision support as a tool for defining weights restrictions in Data Envelopment 

Analysis. We prove that the interval weights obtained for an additive 

multicriteria method is equivalent to the assurance region method for Data Envelopment 

Analysis weight restrictions. The technique is illustrated with an 

application in the real state business, respecting limits previously established 

for each criterion by the decision-maker. 

2 - An interactive approach to produce common weights 

and estimation of value efficiency in data envelopment 

analysis 

Majid Zohrehbandian, Mathematics, Islamic Azad University, 

karaj, Tehran, Iran, Islamic Republic Of, 

zohrebandian@yahoo.com 

Halme et al. present value efficiency for incorporating preference information 

in DEA. Here, we present a refinement to their works. We use an MOLP model 

which its objective functions are input/output variables. Then, by using Zionts- 

Wallenius method, we aid the DM in searching for the Most Preferred Solution 

(set on a supporting hyperplane which approximate indifference contour of a 

DM’s value function) and generating common weights of objective functions. 

Finally, we calculate value efficiency scores by comparing the inefficient units 

to units having the same value as the MPS. 

3 - Parametric Dea models with weight constraints 

Armando Milioni, Departamento de Organizacao, Instituto 

Tecnologico de Aeronautica, CTA ITA IEMB, 12228-900, Sao 

Jose dos Campos, Sao Paulo, Brazil, milioni@ita.br, Rodrigo 

Cesar Silva 

This paper presents the Adjusted Spherical Frontier Model (ASFM-lp) model, 

a parametric Data Envelopment Analysis (DEA) model for allocating new inputs 

to a group of Decision Making Units (DMU’s). This model considers that 

a fair allocation of the input is one that maximizes the DEA-CCR efficiencies 

of the DMUs. The formulation that we provide is a generalization of the standard 

ASFM formulation and allows the introduction of weights restrictions, 

an important development in DEA, which was disregarded in the latter model. 

Numeric examples are presented to show the application of the method. 

4 - Estimating Malmquist Productivity Index by Using 

Stochastic Programming Models 

Mahnaz Mirbolooki, Mathematics, Science and Research 

Branch, Islamic Azad University, Tehran, Iran, Tehran, Iran, 

Islamic Republic Of, mirbolouki.mahnaz@gmail.com, Gholam 

Reza Jahanshahloo, Farhad Hosseinzadeh Lotfi, Mohammad 

Hassan Behzadi 

DEA-based Malmquist productivity index measures the productivity change 

over time. This index can be decomposed into two components: measuring 

the technical efficiency change and measuring the frontier shift. We investigate 

the progress and regress of decision making units(DMUs) in the future and 

forecast the Malmquist index. For this purpose with the use of sampling from 

inputs and outputs of DMUs in past successive times, the distribution of each 

input and output of DMUs have been estimated and by introducing stochastic 

programming models, stochastic Malmquist index has been proposed. 


� MD-08 


6.1.36 

Scheduling Problems - Approaches and 

Complexity 



Chair: Frank Werner, Faculty of Mathematics, Otto-von-Guericke 

University, FMA,I, nstitute of Mathematical Optimization, PSF 4120, 

39016, Magdeburg, Germany, 

frank.werner@mathematik.uni-magdeburg.de 

1 - Single Machine Scheduling with a Non-renewable Financial 

Resource 

Evgeny Gafarov, Institute of Control Sciences of the Russian 

Academy of Sciences, Profsoyuznaya 65, 117997, Moscow, 

Russian Federation, axel73@mail.ru, Alexander Lazarev, Frank 

Werner 

We consider single machine scheduling problems with a non-renewable resource. 

This type of problems has not been intensively investigated in the 

literature so far. For several problems of this type with standard objective functions 

(namely the minimization of makespan, total tardiness, number of tardy 

jobs,total completion time and maximum lateness), we present some complexity 

results. Particular attention is given to the problem of minimizing total tardiness. 

In addition,for the so-called budget scheduling problem with minimizing 

the makespan,we present some properties of feasible schedules. 

2 - Solving a scheduling problem at cross docking terminals 

Ruslan Sadykov, INRIA Bordeaux - Sud-Ouest, 351 cours de la 

Libération, 33405, Talence, France, Ruslan.Sadykov@inria.fr 

At cross docking terminals, products from incoming trucks are sorted according 

to their destinations and transferred to outgoing trucks using a small temporary 

storage. Recently, we proposed an algorithm of high but polynomial complexity 

for a trucks scheduling problem with the objective to minimize the storage 

usage during the product transfer. We will present results of numerical tests of 

this algorithm and a comparison with other approaches from the literature. 

3 - Dual and reverse problems to minimize maximum 

penalty scheduling problems. 

Alexander Lazarev, Institute of Control Sciences, Profsoyuznaya 

st. 65, 117997 Moscow, Russia, 117997, Moscow, Russian 

Federation, jobmath@mail.ru 

We study the classical NP-hard non-preemptive single machine scheduling 

problem to minimize the maximum penalty with release dates. Penalty of 

each job is a monotone and non-decreasing function of its completion time. 

We propose polynomial algorithms for the dual and inverse problems where 

the minimum penalty is maximized. Both algorithms run in quadratic time. 

Note that the solution values of the inverse and dual problems are valid lower 

bounds for the original problem. Therefore, our algorithms can be used within 

the branch-and-bound scheme when solving the original NP-hard problem. 

4 - Optimization and Performance Analysis of Parallel Dynamic 

Programming Algorithm for Knapsack Problem 

Alexander Lazarev, Institute of Control Sciences, Profsoyuznaya 

st. 65, 117997 Moscow, Russia, 117997, Moscow, Russian 

Federation, jobmath@mail.ru, Mikhail Posypkin 

We study a parallel dynamic programming algorithm for Knapsack problem 

based on block decomposition. Two cases are considered: processors can store 

intermediate results of previous computations and intermediate results should 

be communicated to each processor. The former corresponds to classical cluster 

architecture while the latter is tailored to grid computing. For both cases we 

state a scheduling problem of minimizing the total complete time. We suggest 

a fast algorithm to solve this problem and present results of simulation and real 

experiments confirming the algorithm’s efficiency. 

69


� MD-09 


6.2.53 

Nonlinear optimization for industrial 

applications 



Chair: Thomas Lehmann, Optimization, Konrad-Zuse-Zentrum, 

Takustr. 7, 14195, Berlin, Berlin, Germany, lehmann@zib.de 

1 - Nonlinear Mixed-Integer Programs Arising in Petroleum 

Industry 

Thomas Lehmann, Optimization, Konrad-Zuse-Zentrum, 

Takustr. 7, 14195, Berlin, Berlin, Germany, lehmann@zib.de, 

Klaus Schittkowski, Oliver Exler 

In petroleum industry many mixed-integer nonlinear optimization problems 

arise. Since model functions are determined by expensive simulation runs, the 

efficiency in terms of the number of function calls is crucial. 

We introduce a toolbox of SQP-type algorithms that successively solve mixedinteger 

quadratic programs. Linear outer approximations guarantee global optimality 

for convex programs. 

Comparative numerical results for 55 test cases arising from applications in 

petroleum industry are presented. In addition 100 nonlinear and often nonconvex 

academic problems are considered. 

2 - Fast and Safe Container Cranes as Bilevel Optimal Control 

Problems 

Matthias Knauer, Zentrum für Technomathematik, Universität 

Bremen, Bibliothekstraße 1, 28359, Bremen, Germany, 

knauer@math.uni-bremen.de, Christof Büskens 

In order to use a new concept of crane systems in high rack warehouses, reference 

trajectories have to ensure that the swing of the crane is under control 

during the fast movement and disappears at the final point. These trajectories 

can be obtained solving optimal control problems. Further, on user demand, the 

crane has to be brought safely to a rest position from any state within a fixed 

time. While solving the optimal control problem for the trajectories, additional 

constraints have to be considered now, depending on the optimal solution of 

other optimal control problems. 

3 - DAE Optimal Control Problems with Applications in 

Electrical Engineering 

Martin Kunkel, Institut für Mathematik, Universität Würzburg, 

Lehrstuhl für Angewandte Mathematik II, Am Hubland, 97074, 

Würzburg, Germany, 

martin.kunkel@mathematik.uni-wuerzburg.de, Matthias Gerdts 

We consider transient partial differential algebraic equations (PDAEs) which 

arise in various applications, e.g. in simulation of integrated circuits containing 

semiconductors. Herein, the eletrical network is modelled by DAEs and 

the semiconductors are modelled using the drift-diffusion equations. Proper 

finite-element space discretization yields a system of DAEs. We discuss optimal 

control problems subject to DAEs and present possible solution strategies 

such as function space nonsmooth Newton methods. 

� MD-10 


6.2.56 

Graphs and Networks IV 



Chair: Cédric Bentz, LRI, Univ. Paris-Sud, 91405, Orsay, France, 

cedric.bentz@lri.fr 

1 - Minimizing the number of critical stages for the on-line 

steiner tree problem 

70 

Nicolas Thibault, ERMES - Paris 2, Paris, France, 

nicolasvincentpierre@gmail.com 

We study the following on-line problem. Our goal is to connect one by one 

i new member to the current steiner tree, while satisfying a quality constraint: 

the weight of the tree must be at most c (a given constant) times the weight of an 

optimal steiner tree. Then, we want to minimize the number of "critical stages" 

(stages with heavy changes in the tree). We propose an on-line strategy leading 

to at most i/(2(c-((ln 3)/2)-2)-1) critical stages. We also prove that there exists 

situations where any algorithm needs at least (i/(2(4c)+1))-1 critical stages to 

maintain the quality constraint. 

2 - Minimum d-blockers and d-transversals for the maximum 

stable set problem 

Christophe Picouleau, CNAM, Laboratoire Cedric, 292 rue saint 

Martin, 75003, Paris, France, chp@cnam.fr, Dominique de 

Werra, Marie-Christine Costa 

Let G=(V,E) be a simple undirected graph. A d-transversal is a subset of V 

which intersects any maximum stable set while a d-blocker is a subset of V 

whose removal decreases the size of a maximum stable set by at least d. A 

minimum d-transversal (resp. d-blocker) is a d-transversal (resp. d-blocker) 

with minimum cardinality. We study minimum d-transversals and d-blockers 

of stable sets in bipartite and in split graphs. Whereas both problems are polynomial 

in bipartite graphs, for split graphs the situation is different: the search 

for a minimum d-transversal is polynomial and it is NP-hard for a minimum 

d-blocker. Our results on bipartite graphs are strengthened to the minimum 

d-transversal and minimum d-blocker of minimum set covers problems. 

3 - Dense graph partition for community detection 

Julien Darlay, G-SCOP, 46, avenue Félix Viallet, 38031, 

Grenoble, France, julien.darlay@g-scop.inpg.fr, Nadia Brauner, 

Julien Moncel 

The density of a subgraph is defined as the ratio between the number of edges 

and the number of vertices of this subgraph. Finding the subgraph of maximum 

density can be done using flow technics or linear programming. We define the 

density of a vertex partition as the sum of the densities of each subgraph induced 

by a class of the partition. Main applications of this problem lie in the 

field of community detection. We show that this problem is NP-Complete and 

we propose a polynomial time algorithm when the graph is a tree. 

4 - Power Domination in Graphs 

Paul Dorbec, LaBRI, Université de Bordeaux - CNRS, 351 cours 

de la Libération, 33400, Talence, France, 

paul.dorbec@u-bordeaux1.fr 

Supervision of a power system may be performed by the adequate placement of 

phase measurement units. The minimum number of these units that need to be 

placed to entirely monitor a system corresponds to the power domination number 

of the corresponding graph (Haynes, Hedetniemi, Hedetniemi, Henning, 

2002). In contrast with other domination parameters, the power domination 

holds some spreading rules made possible by the use of Ohm’s and Kirchhoff’s 

laws. This parameter thus has a life-game like behavior and requires original 

proof techniques to be studied. In this talk, we shall define precisely the parameter, 

review some known results and discuss some current works on the 

problem. 

� MD-11 


8.2.38 

Stochastic Models in Manpower Planning 



Chair: Andreas Georgiou, Department of Business Administration, 

University of Macedonia, 156 Egnatia Street, 54006, Thessaloniki, 

Greece, acg@uom.gr 

1 - A Model Based Study on The Career Prospects of Individuals 

in An Indian University 

Arindam Gupta, Statistics, University of Burdwan, Department 

of Statistics, The University of Burdwan, Golapbag, Burdwan, 

W.B., India, 700104, Burdwan, West Bengal, India, 

arindamdeep@yahoo.com, Asis Kumar Chattopadhyay

To study the movement of individuals over occupational categories it is natural 

to start by looking at the movement of people between different categories. 

Since such moves are unpredictable at the individual level it is necessary to 

find a model to describe the mechanism of movement in probabilistic terms. 

We have used a markov model to describe this in a half open and half closed 

system. A measure of career pattern based on above model has been developed 

and the simulated distribution of this measure has also been studied. A 

numerical study has been done on the University of Calcutta, India. 

2 - Dealing with observable and latent heterogeneity in 

Markov manpower systems 

Marie-Anne Guerry, MOSI, Vrije Universiteit Brussel, Pleinlaan 

2, 1050 Brussel, Brussels, Belgium, maguerry@vub.ac.be 

In modeling manpower systems, it is of crucial importance to deal with heterogeneity. 

Most of the manpower models are taking into account heterogeneity 

due to observable sources, neglecting heterogeneity due to latent sources. In 

this paper a multinomial Markov-switching model is introduced to deal with 

heterogeneity due to latent sources for the internal flows of the personnel as 

well as for the wastage flows. A re-estimation algorithm is presented to estimate 

the parameters of the Markov-switching manpower model. 

3 - The dynamic behavior of a mixed push-pull manpower 

model 

Tim De Feyter, Centre for Corporate Sustainability, 

Hogeschool-Universiteit Brussel, Stormstraat 2, 1000, Brussels, 

Belgium, tim.defeyter@hubrussel.be, Marie-Anne Guerry 

In the mixed push-pull model, the internal mobility of a manpower system is 

regulated by push as well as pull transitions. The model is a generalization 

of the Markov and the Renewal manpower model. In this paper we study the 

dynamic and asymptotic behavior of the mixed push-pull model under the assumptions 

of time-homogeneity and a known constant recruitment policy and 

investigate the mechanisms underlying the difference with the dynamic behavior 

of the traditional push and pull model. We show that under certain conditions, 

the system evolves towards a limiting personnel distribution. 

4 - Investigating Aspirations, Priorities and Optimization 

Opportunities in Markov Manpower Planning Models 

Andreas Georgiou, Department of Business Administration, 

University of Macedonia, 156 Egnatia Street, 54006, 

Thessaloniki, Greece, acg@uom.gr 

Integrating various optimization aspects in hierarchical manpower planning 

models has been of interest since the early introduction of these structures. 

Decision makers’ aspiration levels and priorities are often contradicted by rigid 

constraints that can not be easily circumvented. This work presents ideas which 

can be employed in manpower nonhomogeneous markov systems which evolve 

in time, in an effort to reach satisfactory structures. Optimization approaches 

based on a general goal programming framework are considered and possible 

other variations that can be employed in this direction are investigated 

5 - Control aspects in an enhanced manpower planning 

model 

Nikolas Tsantas, Department of Mathematics, University of 

Patras, Department of Mathematics, University of Patras, 26500, 

Patras, tsantas@upatras.gr, Vasileios Dimitriou 

This work deals with the exercise of recruitment control to a time dependent, 

hierarchical system which incorporates training classes as well as two streams 

of recruitment; one coming from the outside environment and another from an 

auxiliary external system. The motivation for this model lies in the need to take 

into account not only the tendency of the employees to attend seminar courses 

so as to improve their career prospects, but also the organizations’ intention 

to avoid situations associated with the unavailability of skilled individuals for 

hiring. 

� MD-12 


8.2.39 

AHP 04 


Processes 


Chair: Y. Ilker Topcu, Industrial Engineering, Istanbul Technical 

University, Istanbul Teknik Universitesi, Isletme Fakultesi, Macka, 

34367, Istanbul, Turkey, ilker.topcu@itu.edu.tr 


1 - Finding Effective Strategies for Improving Textile and 

Clothing Supply Chain in Pakistan using SWOT Analysis 

and AHP 

Deedar Hussain, Production and Systems Department, 

University of Minho, Portugal and NED University, Pakistan, 

University of Minho, Campus de Gualtar, 4710-057, Braga, 

Minho, Portugal, deedar_agha@hotmail.com, Manuel 

Figueiredo, Anabela Pereira Tereso, Fernando Ferreira 

The global shift of textile and clothing manufacturing to low cost environments 

has created a strong competition in Asia and Far East. Old and new players 

are developing the missing links in the chain. Although they are at different 

development stages, they share the advantage of being suitable sources for lowvalue 

and standard products. Similar is the context for Pakistan which supplies 

mainly standard products of low added value. This study gets inputs from our 

previous work, "SWOT Analysis of Pakistan Textile Supply Chain" and aims 

to develop competitive strategies using Saatys AHP. 

2 - Risk management in the process of foreign supplier selection: 

Case Study 

Olga Fedotova, DEGEI, University of Aveiro, Campus 

Universitário de Santiago, 3810-193, Aveiro, Aveiro, Portugal, 

olgafedotova@ua.pt, Luis Ferreira 

Managing risk in inbound supply chain operations has become increasingly 

important in today’s competitive and globally environment. This research aims 

to reinforce inbound supply chain risk management by proposing a methodology, 

based on the analytic hierarchy process (AHP), for evaluating and ranking 

potential suppliers. 

A realistic case study is presented in which a Portuguese industrial manufacturer 

evaluates and ranks its current foreign suppliers of stainless steel against 

two other potential foreign suppliers. 

3 - Evaluating health-care waste disposal alternatives using 

analytical hierarchy process 

Melis Almula Karadayi, Systems Eng., Yeditepe University, 

34726, Istanbul, mkaradayi@yeditepe.edu.tr 

Today, as in all other organizations, the amount of waste generated in the 

health-care institutions is rising due to their extent of service. Disposal of 

health-care waste management, including Turkey, is one of the most common 

problems of developing countries. This paper presents the application of analytical 

hierarchy process for evaluating health-care waste disposal alternatives 

for Istanbul, including "incineration’, "steam sterilization ","microwave" and 

"landfill’. Economic, environmental, technical and social criteria and their related 

sub-criteria are employed to evaluate health-care waste dipsosal alternatives. 

4 - An AHP model to evaluate brand equity of sports clubs 

Mine Isik, Industrial Engineering Department, Dogus University, 

Zeamet S. No:21, Acibadem Kadikoy, 34722, Istanbul, -, Turkey, 

misik@dogus.edu.tr, Ozay Ozaydin, Y. Ilker Topcu, Sebnem 

Burnaz 

Brand equity is defined as the sum of all values that is attributed to a brand, 

which makes it a crucial element while directly influencing the market value. 

If the mentioned brand belongs to a product, brand value can easily be attached 

to the qualifications of that product, but if it is a service, measurement of this 

value is more complex. Although there has been numerous studies on this topics, 

there is a gap in sports sector. In this study a comprehensive evaluation 

has been done, outputting the criteria. Then, a pairwise comparison is done in 

order to prioritize these criteria. 

� MD-13 


2.2.21 

Discrete Location I 



Chair: Antonio Manuel Rodríguez-Chía, Estadística e Investigación 

Operativa, Universidad de Cadiz, Facultad de Ciencias. Pol. Río San 

Pedro., 11510, Puerto Real, Cádiz, Spain, 

antonio.rodriguezchia@uca.es 

71


1 - Discrete location for bundled demand points 

Alfredo Marín, Departamento de Estadística e Investigación 

Operativa, University of Murcia, Facultad de Matemáticas, 

Campus de Espinardo, 30100, Murcia, Spain, amarin@um.es 

We consider a discrete location problem where the demand points are grouped 

and propose a formulation, an enforcement for it and an associated agrangian 

relaxation, and then we build feasible solutions to the problem from the optimal 

solutions to the relaxed subproblems. Valid inequalities for the formulation are 

also identified and added to the set of relaxed constraints. This method produces 

good feasible solutions and enables us to address large instances of the 

problem. Computational experiments have been performed with benchmark 

instances from the literature on related problems. 

2 - Lower Bounds for a Capacitated Facility Location Problem 

with Penalties and Revenues 

Maria João Lopes, Departamento de Métodos Quantitativos, 

Lisbon University Institute(ISCTE - IUL); CIO, Av. das Forcas 

Armadas, 1649-026, Liboa, Portugal, mjfl@iscte.pt 

We consider a variant of the Capacitated Facility Location Problem in which 

the demand of each costumer may not be entirely supplied. In this variant, the 

available supply is not enough to satisfy the total demand. A penalty is associated 

with each unit left out. On the other hand, each costumer can be supplied 

by more than one facility. Nevertheless, satisfying the demand of a costumer 

entirely by exactly one facility is aimed. If this condition is satisfied revenue is 

considered. We propose a formulation for this problem and valid inequalities 

to improve the lower bounds. 

3 - Assessing the accuracy for estimation of Origin- 

Destination matrices in the railways system context 

Eva Barrena, Applied Mathematics I, University of Sevilla, Avda. 

Reina Mercedes s/n, 41012, Sevilla, Spain, ebarrena@us.es, M a 

Teresa Cáceres, Francisco A. Ortega, Miguel Angel Pozo 

The estimation of a travel matrix Origin-Destination (O-D), updated with traffic 

counts from the transport network and from a previous O-D matrix, is a 

problem that is of interest for researchers, due to the challenge linked to the 

complexity involved, as well as for transport companies interested in knowing 

the behavior of their users via a more economical and efficient procedure than 

the ones based on surveys. The most representative estimation methods are 

based on entropy maximization, statistical inference models and optimization 

methods, whose effective resolution applies iterative algorithms, heuristics and 

metaheuristics. 

The objectives of this paper consist in the evaluation of the prospects of this 

problem for a specific company such as RENFE-Cercanías Madrid, as well as 

the evaluation of possible methodologies to apply depending on their adequacy 

to the real context. This work has been partially supported by the the Andalusian 

Government project Ref. P09-TEP-5022 and by the Spanish research 

Project PT-2007-003-08CCPP (CEDEX). 

4 - On a common structure of the discrete optimization 

with ordering 

Antonio Manuel Rodriguez-Chia, Estadistica e IO, Universidad 

de Cádiz, Facultad de Ciencias, Pol. Rio San Pedro, 11510, 

Puerto Real (Cadiz), Cadiz, Spain, 

antonio.rodriguezchia@uca.es, Elena Fernandez, Justo Puerto 

This paper studies discrete optimization problems with ordering requirements. 

These problems are formulated on general discrete sets in which there exists an 

implicit ordering on their elements together with a cost function that evaluates 

each element of a given subset depending on its ordering relative to the remaining 

elements in the set. It is proven that ordered sequences over the original 

set define an independence system. The simplest and its restriction to sets of 

a fixed cardinality are studied. Ordering problems on the intersection of two 

independence systems are addressed. 

� MD-14 


2.2.15 

Pricing Issues 

Stream: Supply Chain Planning [c] 


Chair: Carla A. S. Geraldes, Department of Industrial Management, 

Polytechnic Institute of Bragança, Campus de Santa Apolónia, 

Apartado 134, 5301-857, Bragança, Portugal, carlag@ipb.pt 

72 

1 - A New Multi-Objective Approach for Pricing and Distribution 

in a Supply Chain 

D. Shishebori, Industrial Engineering, Iran University of Science 

and Technology, Narmak, 1684613114, Tehran, Tehran, Iran, 

Islamic Republic Of, shishebori@in.iut.ac.ir, S.g.r. Jalali-Naini, 

M. Karimi-Nasab 

In this research a new multi-objective approach is proposed to determine distribution 

policy for a wholesaler of supplementary nutrition to a set of local 

distribution centers positioned around the wholesaler, geographically. Each o 

local distribution center has its own stochastic demand over planning periods. 

Also the demand of each local distribution center in a planning period is considered 

as a normal distribution function. There is a time series relation between 

the mean of demand of each local distribution center in a period and previous 

periods. The model illustrates the conflict of two objectives for the main partners 

of the problem: manager of the wholesaler aims at maximizing the total 

profit of the wholesaler, and local distribution centers desire to maximize the 

minimum service level received from the wholesaler. The approach optimizes 

selling price, carrying cost, batch size and services level of multi items for each 

local distribution center in every planning period. Some of the theorems are 

proven about different characteristics of the proposed model and a new solution 

method is proposed for obtaining a set of global Pareto-optimal solutions 

of the problem. The solution method is proven to be able to find all global 

Pareto-optimal solutions of the problem in certain conditions. Finally the computational 

experiences of running the proposed approach in a real case study 

are analyzed. 

2 - Determination of pre-order and after-order prices for a 

capacitated supply system 

Bisheng Du, Department of Business Studies, Aarhus School of 

Business, Aarhus University, Fuglesangs Alle 4, DK-8210 

Aarhus V, 8210, Aarhus V, Aarhus, Denmark, bisd@asb.dk, 

Christian Larsen 

In order to transfer some of the risk of having developed inadequate capacity 

to his buyers a supplier may be willing to offer his buyers to make pre-orders, 

issued before the capacity decision, in exchange for reduced prices compared 

to the prices the buyers pay after having observed their demand. By use of a 

news-vendor model we develop the optimization problem for the single buyer 

case and investigate the optimal prices. In addition, we also investigate the case 

of demand probability updating given the advance demand information. 

3 - Competitive Price-Matching Guarantees under Demand 

Uncertainty and Customer Heterogeneity: Effects of 

Product Availability and Its Verification 

Arcan Nalca, School of Business, Queen’s University, Canada, 

arcan.nalca@business.queensu.ca, Tamer Boyaci, Saibal Ray 

Price-matching-guarantees (PMGs) are offers where firms promise to match 

any lower price offered by competitors. Firms nowadays reserve the right 

to verify the availability of the product at competitors’ and decline to match 

the price unless it is available. Focusing on these elements, we investigate 

the effects of demand uncertainty and verification of availability in the context 

of PMGs and show that verifying availability is a significant profit-enhancing 

mechanism for innovative products. 

4 - Integrated approaches to warehouse planning and operations 

Carla A. S. Geraldes, Department of Industrial Management, 

Polytechnic Institute of Bragança, Campus de Santa Apolónia, 

Apartado 134, 5301-857, Bragança, Portugal, carlag@ipb.pt, 

Sameiro Carvalho, Guilherme Pereira 

In this talk we discuss a tactical model recently available in warehouse literature. 

The model integrates the replenishment decision in inventory management, 

the allocation of products to warehousing systems and the assignment 

of products to storage locations in warehousing management. Our aim is to 

show the models’ potentialities and weaknesses when applied to a wide variety 

of problems and to identify challenging research opportunities for developing 

global warehouse decision support models that fill the gap between researchers 

and warehouse practitioners.

� MD-15 


2.2.12 

Exact Algorithms for Vehicle Routing 



Chair: Roberto Baldacci, DEIS, University of Bologna, Via Venezia, 

52, 47023, Cesena, Italy, r.baldacci@unibo.it 

1 - The vehicle routing problem with time windows and 

edge set costs 

Line Blander Reinhardt, DTU Management, The Technical 

University of Denmark, Produktionstorvet 424, 2800 , 2800 Kgs. 

Lyngby, Denmark, lbre@man.dtu.dk, David Pisinger, Mads 

Kehlet Jepsen 

In real life applications the vehicle routing problem takes many forms. The 

cost of access to a set of edges is investigated together with the well known 

time window restriction. The cost of accessing a set of edges is a single payment 

to allow all the vehicles to access all of the edges in the set payed for. 

A real life example of this is when a company makes a single payment so that 

all vehicles obtain unlimited access to a specified number of ferry routes in a 

given time period. A mathematical model, solution methods and test results are 

presented. 

2 - Experiments with new cuts on the VRP 

Eduardo Uchoa, Engenharia de Produção, Universidade Federal 

Fluminense, Rua Passo da Pátria 156, São Domingos, 

22430-210, Niterói, Rio de Janeiro, Brazil, 

uchoa@producao.uff.br, Diego Pecin, Artur Pessoa, Marcus 

Poggi de Aragão 

The most successful current exact algorithms for the VRP combine a setpartitioning 

formulation (where columns may be elementary routes or q-routes) 

with cuts. It is possible to cut: (i) in the SP variables, or (ii) in the variables 

of the original formulation. The SP cuts are potentially stronger, but have the 

side effect of making the pricing intractable on instances with many clients 

per route. We present recent developments on the second kind of cuts, over 

arc-capacity indexed variables. Computational experiments over classical VRP 

(and scheduling) instances show significant gap reductions 

3 - Column Generation and Branch-and-Price for the m- 

PVRP 

Sandra Ulrich Ngueveu, LOSI (Laboratory of Industrial Systems 

Optimization), University of Technology of Troyes, 12, rue 

Marie Curie, 10010, Troyes, France, ngueveus@utt.fr, Christian 

Prins, Roberto Wolfler-Calvo 

The m-Peripatetic Vehicle Routing Problem models regular money transportation 

with peripatetic and capacity constraints ensuring that no sequence of 

clients is repeated during m periods and that the amount of money allowed 

per vehicle is limited. The total cost is to be minimized. We present new lower 

bounding procedures and a branch-and-price algorithm based on dual heuristics 

that compute dual feasible solutions for the set partitioning relaxation. Their 

efficiency lies upon the approximation of routes with q-routes, the dual ascent 

that estimates the best dual values and column generation. 

4 - New Benchmark Results for the Capacitated Vehicle 

Routing Problem 

Roberto Baldacci, DEIS, University of Bologna, Via Venezia, 52, 

47023, Cesena, Italy, r.baldacci@unibo.it, Aristide Mingozzi, 

Roberto Roberti 

The Capacitated Vehicle Routing Problem (CVRP) is the problem of designing 

optimal delivery routes for a fleet of vehicles in order to supply a set of 

customers with given demands at minimum cost. In this paper, we improve 

the exact method recently proposed by Baldacci, Christofides, and Mingozzi 

(2008) using Subset-Row (SR) inequalities and a new route relaxation, called 

ng-route. Computational results show that the proposed method outperforms 

the best methods presented in the literature. 


� MD-16 


2.2.14 

Robust planning and rescheduling 



Chair: Natalia Kliewer, Information Systems, Freie Universitaet 

Berlin, Garystr. 21, 14195, Berlin, Germany, 

natalia.kliewer@fu-berlin.de 

1 - GRASP algorithms for the Robust Railway Network design 

problem 

Antonio J. Lozano, Department of Mathematics, University of 

Huelva, Dpto. de Matemáticas, Facultad de Ciencias 

Experimentales, 21071, Avda de las Fuerza Armadas s/n, 21007, 

Huelva, Spain, antonio.lozano@dmat.uhu.es, Juan A. Mesa, 

Federico Perea 

Designing a railway network aiming at maximizing trip coverage assuming an 

alternative transportation mode (for instance a bus) competing with the railway 

system is a NP-hard problem. If we add a robustness component to the problem, 

which in our case consists of allowing links to fail, the problem becomes 

even more complex. In this talk we will introduce a class of GRASP algorithms 

to give a feasible near-optimal solution in a reasonable amount of time. 

2 - Strategic Planning in the Airline Industry under uncertainty 

Marc Naumann, Decision Support & Operations Research Lab, 

University of Paderborn, Warburger Str 100, 33098, Paderborn, 

NRW, Germany, naumann@dsor.de, Leena Suhl, Achim 

Koberstein 

Today, fuel costs are a growing part in airlines’ expenditures and their fluctuations 

become higher. We connect schedule development and fleet assignment 

and develop a new strategic planning model for airlines under fuel price uncertainty. 

We present a stochastic model that determines the optimal offered 

flights, passenger routes and aircraft types. To counteract fuel price uncertainty, 

financial hedging is considered. We show that the optimal offered flights and 

the aircraft types depend on the fuel price development. Finally, the integration 

of hedging improves profit at given risk levels. 

3 - An Applicable Recovery Framework for Airline Operations 

Lucian Ionescu, Information Systems, Freie Universität Berlin, 

Garystr. 21, 14195, Berlin, Berlin, Germany, 

Lucian.Ionescu@fu-berlin.de, Viktor Dück, Natalia Kliewer 

The airline transportation frequently has to deal with occuring delays, which 

may lead to infeasible schedules during the day of operations. This forces the 

Operations Control Center to recover these schedules through mostly expensive 

actions. For the offline scheduling problem there exists a lot of indicators 

for robustness with possibly negative correlated impacts on each other. To 

objectively measure these impacts a realistic evaluation framework including 

real-time applicable recovery actions is needed. In this context, we present 

an online approach to cost-efficiently deal with disruptions by an extendable 

recovery framework with regard to the estimated impact for all following tasks. 

4 - Approaches to increase delay-tolerance of vehicle and 

crew schedules in public bus transport 

Bastian Amberg, Information Systems, Freie Universitaet Berlin, 

Garystr. 21, 14195, Berlin, Germany, 

bastian.amberg@fu-berlin.de, Natalia Kliewer, Stefan 

Kramkowski 

In public bus transport delays occur frequently during the transportation process. 

Delayed buses not only affect the vehicle schedule but also the associated 

crew schedule. Delayed drivers cause similar effects the other way round. 

Thus planned schedules can become infeasible and the operations control has 

to manage expensive recovery actions. We present offline approaches to increase 

delay-tolerance of both schedules. Focus is on using buffer time to cope 

with minor disruptions and to control delay propagation. The approaches are 

compared with regard to planned costs and delay-tolerance. 

73


� MD-17 


1.3.14 

Breaks in Vehicle Routing and Scheduling 



Chair: Christoph Manuel Meyer, Department of Economics, Chair of 

Logistics, University of Bremen, Wilhelm-Herbst-Straße 5, 28359, 

Bremen, Germany, meyer@wiwi.uni-bremen.de 

1 - Vehicle Routing under Consideration of the European 

Social Legislation 

Christoph Manuel Meyer, Department of Economics, Chair of 

Logistics, University of Bremen, Wilhelm-Herbst-Straße 5, 

28359, Bremen, Germany, meyer@wiwi.uni-bremen.de, Herbert 

Kopfer 

In practice, planners have to consider legal restrictions on driving and working 

hours in transportation planning. In this presentation, the impact of the 

European social legislation on vehicle routing and scheduling is investigated. 

Special attention is given to the integration of the optional rules of the legislation 

which have often been neglected in vehicle routing methods so far. These 

optional rules allow for an increased solution space. We will show that their 

inclusion in models and solution approaches results in improved vehicle schedules 

both for centralized and distributed planning. 

2 - Vehicle Routing with Breaks and Rests, a Practitioner’s 

Point of View 

Robert Scheffermann, Logistics Systems Engineering, 

Forschungszentrum Informatik, Haid-und-Neu-Straße 10-14, 

76131, Karlsruhe, Germany, scheffer@fzi.de, Andreas Cardeneo, 

Werner Heid 

Recently published methods to schedule breaks and rests in VRPs share the 

same basic mechanisms: branching, domination and slack-time. We argue 

that when scheduling rests in VRPs it is inevitable to consider multiple timewindows. 

We will present their impact on these mechanisms, showing that the 

number of branches grows dramatically and the effect of domination becomes 

very weak. We will also show that the naïve utilization of slack-time leads to 

volatile solutions. We present methods for branching based on multi-objective 

genetic programming and domination based on a S-metric selection. 

3 - A pickup and delivery model considering EU driving 

time regulations 

Alexandra Hartmann, Business School, University of Applied 

Sciences, Saarland, 66115, Saarbrücken, Germany, 

alexandra.hartmann@htw-saarland.de 

Motivated by a real-life application in long-haul freight transportation, a new 

pickup and delivery model is proposed for routing and scheduling a group of 

requests assigned to a vehicle. In addition, rest period planning is integrated in 

the model by taking EU driving and working regulations into account. Multiple 

time windows are specified at the origin and destination of each request. The 

objective is to minimize the total route duration and the violation of time windows. 

Computational results for the new mixed-integer linear program based 

on real-life data are presented. 

� MD-18 


1.3.15 

Queueing Systems 











74 

1 - The Effect of the Number of Servers in Quasi-Randominput 

Queues 

Moshe Eben-Chaime, Industrial Eng. & mgt., Ben Gurion 

University, P.O. Box 653, 84105, Be’er-Sheav, Israel, 

even@bgu.ac.il 

The finite source population of quasi-random input (QRI) queues creates state 

dependent arrival rates. Thus, probabilities and performance measure admits 

no closed form expressions, a barrier to the development of analytical results. 

Yet, we present new such results. First, monotone effect of the number of 

servers on most performance measure is shown. QRI queuing models are applied 

to the design of human-machine systems. Establishing the convexity of 

the mean queue length in the number of servers implies diminishing marginal 

productivity of both the servers and the sources — the machines. 

2 - Queuing system simulation and efficiency evaluation 

by Petri nets and Data Envelopment Analysis 

Dragana Makajic-Nikolic, Laboratory for Operational Research, 

Faculty of Organization Sciences, Serbia, gis@fon.rs, Gordana 

Savic, Mirko Vujosevic, Novak Novokmet 

Real postal services queuing system with two fixed channels is modelled by 

Petri Nets and several scenarios for distinct rules of additional channel activation 

are simulated in CPN Tools software. In order to decide the best rule, 

DEA is applied for scenarios efficiency evaluation. Total time all users spend 

in the system and expected number of users in all queues are considered as inputs, 

since the total occupation time of the fixed and the additional channels are 

used as DEA outputs. Described procedure is repeated several times in cases 

of different arrival rates queuing system. 

3 - Queues with boundary assistance and the many effects 

of truncation 

Giang Nguyen, d’informatique, ULB, Belgium, 

giang.nguyen@ulb.ac.be, Peter Taylor, Guy Latouche 

Consider a simple system with two queues, each with its own Poisson stream 

of customers and its own server, of which the service time is exponentially distributed. 

If a server is free and its associated queue is empty, then it serves 

a neighboring customer, provided there is one. The service rate of a server 

does not change regardless of whether it is serving its own or a neighbouring 

customer. 

This system is an example of queues with boundary assistance and may be 

modeled by a Quasi-Birth-and-Death process, with the level being the minimum 

queue length and the phase the difference between the two queue lengths. 

We investigate the asymptotic behavior of the level in the original infinitelymany-phase 

system and in its finite approximations where the queue length 

difference is bounded. In the infinite system, we determine the convergence 

norm of the rate operator of the QBD, and consequently the interval in which 

the decay rate of the minimum queue length lies. 

We consider four finite approximations: one is the infinitely-many-phase system 

truncated without augmentation, and three are obtained from different augmenting 

schemes such that stochasticity is preserved. We show that the first has 

a monotonically increasing decay rate that approaches the convergence norm of 

the rate operator in the infinite system, as the truncation size tends to infinity. 

All three truncated and augmented systems have decay rates that are independent 

on the truncation size. Finally, we observe that the stability regions for the 

original system and the aforementioned finite approximations are nonequivalent. 

4 - Analytical study of a queueing system with nongeometric 

tail behaviour 

Mark Van Lokeren, MOSI, Vrije Universiteit Brussel, Pleinlaan 

2, 1050, Brussel, mvlokere@vub.ac.be, Bart Steyaert, Herwig 

Bruneel 

We consider a queueing system with two types of customers. Arriving customers 

of type A enter the system surely, whereas the probability that an arriving 

customer of type B enters the system depends on the total number of 

customers already present in the system (discouraged arrivals). Service times 

do not depend on the type of customer. We obtain analytical expressions for 

several performance measures of the system. The non-geometric tail behaviour 

of the total number of customers in the system (and the queue length) is discussed 

in detail.

� MD-19 



Industrial Organization 



Chair: Bruno M.P. M. Oliveira, FCNAUP, R. Dr. Roberto Frias, 

4200-465, Porto, Portugal, bmpmo@fcna.up.pt 

Chair: Luis Ferreira, Matemática, Escola Superior de Estudos 

Industriais e de Gestão, IPP, ESEIG, Rua D Sancho I, número 981, 

4480-876, Vila do Conde, Porto, Portugal, migferreira2@gmail.com 

1 - Edgeworthian economies with selfish participants 

Bruno M.P. M. Oliveira, FCNAUP, R. Dr. Roberto Frias, 

4200-465, Porto, Portugal, bmpmo@fcna.up.pt, Luis Ferreira, 

Alberto A. Pinto, Athanassios Yannacopoulos, Barbel 

Finkenstadt 

A key problem of economic theory is the convergence of the prices in a market 

to their equilibrium values. We show that under some fairly general and 

easy to check symmetry conditions, depending on the initial distribution of endowments 

and the agents preferences, the sequence of Edgeworthian prices in 

a random matching economy converges to the Walrasian prices for this economy. 

We associate a selfishness factor to each participant which brings up a 

game alike the prisoner’s dilemma. We observe that the fraction of selfish participants 

is related to their increase in utility. 

2 - Patents in new technologies 

Luis Ferreira, Matemática, Escola Superior de Estudos 

Industriais e de Gestão, IPP, ESEIG, Rua D Sancho I, número 

981, 4480-876, Vila do Conde, Porto, Portugal, 

migferreira2@gmail.com 

We present a new R&D investment function in a Cournot competition model 

inspired in the logistic equation. We do a full characterization of the associated 

game and study the short and long term economical eects derived from using 

this new R&D investment function. For low production costs, that can correspond 

to the production of old technologies, the long term economical eects 

are not very sensitive to small changes in the eciency of the R&D programs. 

However, for high production costs, that can correspond to the production of 

new technologies, the long term economical eects are very sensitive to small 

changes in the eciency of the R&D programs. 

3 - A StochasticC Competitive R&D Race Where Winner- 

Takes-All 

Boaz Golany, Industrial Engineering & Management, Technion - 

Israel Institute of Technology, Technion City, 32000, Haifa, 

Israel, golany@ie.technion.ac.il, Inbal Mund, Uriel G. Rothblum 

The paper considers an environment in which multiple firms compete over the 

development of a ceratin product. Each firm needs to decide whether to launch 

the required R&D project and how much to invest in it and these decisions are 

known to all the competing firms. We prove the existence of a unique Nash 

Equilibrium solution, define the special features of that solution, provide an explicit 

formula that computes it and draw some managerial insights that emerge 

from the structure of the Nash equilibrium. Under certain assumptions, we 

also derive an explicit expression for the (unique) globally optimal solution and 

compare it with the Nash equilibrium solution. 

4 - A two-ends model for the competitive firm under price 

uncertainty 

Alberto A. Álvarez-López, Quantitative Applied Economics II, 

UNED (National University of Distance Education), Paseo 

Senda del Rey, 11, 28040, Madrid, Spain, aalvarez@cee.uned.es, 

Inmaculada RodrÍguez-puerta 

A competitive risk-averse firm has just produced a known amount of a perishable 

output. The firm can intend it for two possible ends (one with certain price, 

the other with uncertain price), so that the total amount is fully distributed between 

both of them. We find a "frontier price’ for the certain end, beyond which 

the firm prefers to intend a positive amount for the uncertain end. We also give 

comparative static results (which are proved with analytic methods scarcely 

used in the literature) and examine a dual approach of the model which lets us 

enhance the scope of its applicability. 

EURO 24 - Lisbon 2010 MD-20 

� MD-20 


1.3.33A 




Chair: Yuri Stoyan, Department of Mathematical Modeling and 

Optimal Design, Institute for Mechanical Engineering Problems of 

the National Academy of Sciences of Ukraine, 2/10 Pozharsky St., 

61046, Kharkov, Ukraine, stoyan@ipmach.kharkov.ua 

1 - A strategic neighborhood search approach for the 

cylinder packing problem 

Shigeyuki Takahara, Kagawa Prefectural Industrial Technology 

Center, 587-1 Goto-cho, 761-8031, Takamatsu, Japan, 

takahara@itc.pref.kagawa.jp 

This paper is concerned with the problem of packing cylinders of different sizes 

into a rectangular container. In order to solve this problem, a strategic neighborhood 

approach is proposed. This approach is based on a greedy cylinder 

allocation algorithm and several improvement heuristics. The strategic procedure 

determines the arrangement of each cylinder with reference to search 

process. The effectiveness of the proposed approach is shown by comparing 

the results obtained with the approaches presented in literature by using benchmark 

problems. 

2 - A formulation space search heuristic for the circle packing 

problem 

Claudia Lopez, Brunel University, United Kingdom, 

claudia.lopez@brunel.ac.uk, J. E. Beasley 

We present a heuristic algorithm based on the formulation space search (FSS) 

method to solve the circle packing problem (CPP). The CPP considered is of 

finding the maximum radius of a specified number of identical circles without 

overlaps into a two-dimensional container. The problem is formulated as 

a nonlinear optimisation problem involving both Cartesian and polar systems. 

FSS consists of moving thorough different formulations of the same problem. 

Our heuristic improves on previous results based on same method, is a computationally 

effective approach when compared with other work. 

3 - Covering a non-convex polytope by minimal number of 

congruent parallelepipeds 

Yuri Stoyan, Department of Mathematical Modeling and Optimal 

Design, Institute for Mechanical Engineering Problems of the 

National Academy of Sciences of Ukraine, 2/10 Pozharsky St., 

61046, Kharkov, Ukraine, stoyan@ipmach.kharkov.ua, E.s. 

Sosurcka 

The problem of covering a non-convex polytope by a minimal number of congruent 

parallelepipeds is considered. A mathematical model of the problem 

on the ground of phi-functions and special covering functions is constructed 

and its characteristics are investigated. These characteristics allow the search 

of approximate solutions of the problem to be reduced to solving a sequence 

of linear programming problems. These sequences are chosen either randomly 

or by a special way including a branch and bound algorithm. A number of 

numerical results are given. 

4 - An Efficient Mechanism for Solving Two Phase Packing 

Problems 

Arik Sadeh, Management of Technology, Holon Institute of 

Technology, 52 Golomb Street, P.O. Box 305, 58102, Holon, 

Israel, sadeh@hit.ac.il, David Raz 

Packing problems are NP complete even for a small scale problem. A procedure 

for finding a solution for a two phase packing problem is proposed. First, 

the mechanism conducts a pseudo efficient frontier of solutions, while in the 

second phase a scaled down packing problem is solved. There is a considerable 

reduction in the number of feasible solution to be calculated. The concept 

is applied to transporting a daily portfolio of goods, characterized by volume, 

weight and predetermined allowable weight. Transporting bins are associated 

with set up cost and extra weight cost, if any. 

75


� MD-21 


6.2.47 

Optimization Modeling II 



Chair: Bjarni Kristjansson, Maximal Software, Ltd., Boundary 

House, Boston Road, W7 2QE, London, United Kingdom, 

bjarni@maximalsoftware.com 

1 - Recent enhancements in GAMS 

Jan-Hendrik Jagla, GAMS Software GmbH, 50933, Cologne, 

Germany, jhjagla@gams.com, Lutz Westermann 

We will demonstrate new features and improvements recently added to the General 

Algebraic Modeling System (GAMS). Continuously improving the system 

while maintaining our commitment to backward compatibility makes the development 

of optimization-based decision support applications efficient and 

productive. Among others we will present recently added solvers, enhanced 

interfacing with other applications through our APIs, and more hidden features 

that users might not know about, yet! 

2 - MPL for Python - Introducing new scripting and library 

interfaces for the MPL Modeling Language 

Sandip Pindoria, Maximal Software Ltd, Boundary House, 

Boston Road, W7 2QE, London, United Kingdom, 

sandip@maximalsoftware.co.uk, Bjarni Kristjansson 

MPL has been distributed with a standard GUI interface for development and 

object-oriented library for deployment for many years now. With the advent of 

scripting language frameworks such as Python, that are becoming increasingly 

popular, there are now new opportunities for integrating optimization into realworld 

applications. With "MPL for Python" we are introducing a new scripting 

and component library interfaces for MPL, that takes full advantage of the 

many powerful features of Python. 

3 - Fast & flexible modeling with AIMMS 

Frans de Rooij, AIMMS, Paragon Decision Technology B.V., 

Schipholweg 1, 2034 LS, Haarlem, Netherlands, 

frans.de.rooij@aimms.com 

AIMMS is software for developing and deploying optimization models. 

We will demonstrate how AIMMS enables fast model formulation, without 

having to learn a programming language. The integrated GUI allows you to 

visualize and interpret results. AIMMS offers flexible, advanced algorithmic 

capabilities: column generation, outer approximation for MINLP, multi-start 

for NLP, Stochastic Programming, Robust Optimization, etc. 

Industry cases will show how AIMMS models are deployed, either as complete 

optimization application, or as optimization engine integrated with existing applications. 

� MD-22 



OR in Education II 



Chair: Ruy Costa, Mathematics Dept, FCT-UNL, Campus de 

Caparica, 2829-516 , Caparica, Portugal, rcosta@fct.unl.pt 

1 - OR education at European level: a case of cooperation 

Portugal-Romania 

76 

Joao Miranda, Technologies and Design, ESTG/IPPortalegre, 

Lugar da Abadessa, Apt 148, 7301-901, Portalegre, Portugal, 

jlmiranda@estgp.pt, Mariana Nagy 

European cooperation is a relevant subject nowadays, when there is the common 

purpose to build a competitive and knowledge-based economy in the European 

Union. In this communication, it is presented a case of Erasmus mobility 

between the most significant institutions of higher education in Portalegre 

(Portugal) and Arad (Romania), which addresses OR topics. It is proposed a 

module of 8 lecturing hours based on 4 sessions of 2 hours each: i) an Introductory 

Session, to focus the basics of computational Linear Algebra, Linear 

Programming, Integer Programming, with computational support (Excel and 

MATLAB); ii) an Interim Session, to address modelling subjects on a dropand-by 

session; iii) an Advanced Session, on the sequence of first one, to treat 

uncertainty and also using multi-criteria decision making methods; iv) an Ending 

Session, to perform the evaluation of learning outcomes. It is targeted the 

exploitation of this cooperation at European Level, which also includes curricula 

normalization and adjustments, cultural exchanges, and research lines 

sharing in the sense to promote the mobility of students and faculty. This mobility 

case thus contributes for the growth and sustainability of international 

cooperation that is necessary onto the foreseen network of higher education 

institutions at European level. 

2 - The challenge of teaching OR 

Gertjan de Lange, Sales & Marketing Director, AIMMS 

(Paragon), PO Box 3277, 2001 DG, Haarlem, NH, Netherlands, 

g.de.lange@aimms.com 

Teaching an OR is a challenge for both universities and companies acting in 

OR. Our educational program and related activities are in place to promote easy 

access and co-operation. This includes free access to AIMMS for students, development 

of OR games in co-operation with universities, optimization modeling 

competitions, and supporting open source initiatives such as COIN-OR. In 

this session, we want to share our experiences and challenges. 

3 - Forecasting the overall academic performance of university 

undergraduates using neural networks 

Ruy Costa, Mathematics Dept, FCT-UNL, Campus de Caparica, 

2829-516 , Caparica, Portugal, rcosta@fct.unl.pt 

We used data from FCT-UNL and considered neural networks with one hidden 

layer with a number of neurons varying from 1 to 4 with four input variables 

(MAT — the summation of passing grades in Mathematics courses in the first 

year (or first two years) in the University; PHY — sum for Physics courses ; 

CHE — sum for Chemistry courses and OTH — sum for other area courses) 

and only one output variable (DUR, the total number of years an undergraduate 

spent in university). We will present results and discuss further improvements 

in the models. 

� MD-23 


6.2.49 

MOO: Multi-Objective Combinatorial 

Optimization 



Chair: Matthias Ehrgott, Engineering Science, University of 

Auckland, Private Bag 92019, 1001, Auckland, New Zealand, 


1 - On the computation of all supported efficient solutions 

in multi-objective integer network flow problems 

Augusto Eusébio, Escola Superior de Tecnologia e Gestão, 

Instituto Politécnico de Leiria, Morro do Lena — Alto Vieiro, 

2411-901, Leiria, Portugal, aeusebio@estg.ipleiria.pt, José Rui 

Figueira 

We present a new algorithm for identifying all supported non-dominated vectors 

in the objective space, as well as the corresponding efficient solutions in 

the decision space, for multiobjective integer network flow problems. Our approach 

is based on a negative-cycle algorithm used in single objective minimum 

cost flow problems, applied to a sequence of parametric problems. The 

proposed approach uses the connectedness property of the set of supported nondominated 

vectors/efficient solutions to find all integer solutions in maximal 

non-dominated/efficient facets. 

2 - A multi-objective branch-and-cut algorithm and its application 

to the multi-modal traveling salesman problem 

Nicolas Jozefowiez, LAAS-CNRS, 7 avenue du Colonel Roche, 

31077, Toulouse, France, nicolas.jozefowiez@laas.fr, Gilbert 

Laporte, Frédéric Semet

We will present a generic branch-and-cut algorithm applicable to a wide class 

of multi-objective optimization problems for which a lower bound can be 

defined as a polynomially (or pseudo-polynomially) solvable multi-objective 

problem. It is applied to and implemented for the Multi-Modal Traveling Salesman 

problem. Computational experiments are conducted and the method is 

compared to the classical epsilon-constraint algorithm. 

3 - Analyzing optimal paths in coloured-edge graphs with 

Euclidean weights 

Felipe Lillo, School of Computing and Mathematical Sciences, 

AUT university, AUT Tower, WT406, cnr Wakelfield & Rutland 

Street, Auckland CBD, 1142, Auckland, Auckland, New 

Zealand, flillovi@aut.ac.nz, Andrew Ensor 

The weighted coloured–edge graph is a tool that allows the modelling of multimodal 

networks by assigning to each edge both a weight for optimization 

criteria and a colour for transport mode. The shortest path between two points 

is computed by imposing a partial order relation on the path weights in each 

colour resulting in a Pareto set of optimal paths. This work experimentally 

studies the order of this set for coloured–edge graphs whose edge weights are 

determined by Euclidean distance. Computational results indicate the tractability 

of the model is linked to the level of randomness as well as the dimension 

of the Euclidean space. 

4 - Tight upper bound on the number of optimal paths in 

weighted coloured-edge graphs 

Andrew Ensor, School of Computing and Mathematical 

Sciences, AUT University, AUT University. AUT Tower, Level 

1, 2-14 Wakefield St, 1010, Auckland, Auckland, New Zealand, 

aensor@aut.ac.nz, Felipe Lillo 

A weighted coloured-edge is a graph for which each edge is assigned both a 

positive weight and a discrete colour, and can be used to model transportation 

and computer networks in which there are multiple transportation modes. In 

such a graph paths are compared by their weight in each colour, resulting in a 

Pareto set of optimal paths from one vertex to another. This paper will give a 

tight upper bound on the cardinality of the Pareto set and explain some results 

toward establishing the average case cardinality. 

� MD-24 


6.2.50 

Bioinformatics IV 


Medicine 


Chair: Ceyda Oguz, Department of Industrial Engineering, Koc 

University, Rumeli Feneri Yolu, Sariyer, 34450, Istanbul, Turkey, 

coguz@ku.edu.hk 

1 - Manhattan Hamiltonian Path in NMR study 

Marta Szachniuk, Institute of Bioorganic Chemistry, PAS, 

Noskowskiego 12/14, 61-704, Poznan, Poland, 

Marta.Szachniuk@cs.put.poznan.pl 

Defining a good graph model is often a very important aspect of dealing with 

a novel problem. It allows for the theoretical analysis of a problem and introducing 

new methods to solve it. Structural bioinformatics provides many tasks 

which require modeling via graph tools. Here, we present the problem of transfer 

pathway reconstruction within NMR spectrum, recorded for RNA molecule 

during its tertiary structure determination. The characteristics of the problem 

made us transform it to a version of Hamiltonian path, which we named Manhattan 

Hamiltonian path. 

2 - A study of different hyper-heuristics for sequencing by 

hybridization problem 

Aleksandra Swiercz, Institute of Computer Science, Poznan 


aswiercz@cs.put.poznan.pl, Wojciech Mruczkiewicz, Jacek 

Blazewicz, Graham Kendall, Edmund Burke 


Hyper-heuristic is a new searching technology, which goal is to raise the level 

of generality at which optimization system can operate. A Hyper-heuristic 

works with a set low-level heuristics and does not have a knowledge about the 

problem it is solving. The low-level heuristics are dedicated to a given problem 

and search the solution space. A hyper-heuristic chooses in intelligent way 

which heuristic should be called at each moment. A hyper-heuristic framework 

has been constructed for the DNA sequencing by hybridization and traveling 

salesman problems. 

3 - Petri net based model of atherosclerosis 

Piotr Formanowicz, Institute of Computing Science, Poznan 


piotr@cs.put.poznan.pl, Dorota Formanowicz, Tomasz Nowak 

Atherosclerosis is one of the leading death cause in developed countries. It is 

a very complex process whose nature is not well understood. In this paper a 

Petri net based model of the main part of this process is presented. An application 

of the Petri net theory to the description of biological processes allows for 

a very precise analysis of the resulting models. Here, such an analysis of the 

atherosclerosis model from a mathematical point of view is given. The model 

allows also for a simulation of the process, since Petri net theory provides a lot 

of analysis techniques and tools. 

4 - Mathematical Model of Blood Flow and Nutrients Delivery 

Trishna Fadjrir, Medicine, University of Sumatera Utara, Jl. 

Ampera, Jl. Dr Mansur, 20155, Medan, Indonesia, 

tfadjrir@yahoo.com, Herman Mawengkang 

Abstract The placenta is a round, flat organ that forms during pregnancy to give 

the baby food and oxygen from the mother. During a normal pregnancy, the 

placenta stays firmly attached to the inside wall of the uterus until the baby 

has been born. The placenta provides the baby with nutrients and oxygen from 

the mother and carries away fetal waste. This paper proposes a mathematical 

model which could describe blood flows to the exchange of oxygen, nutrients 

and waste products between the fetus and the mother. The structure of placenta 

was explored, particularly the flow patterns of fetal arteries. The basic 

equation was developed using a compartment model in order to investigate the 

effect blood jetting out of a maternal spiral artery. Analysis the model could 

lead estimating nutrient delivery to the fetus. 

� MD-25 


6.2.48 

Financial Dynamics and Bubbles 






1 - New Unified Approaches on Identification and Optimization 

of Financial Processes 

Gerhard-Wilhelm Weber, Institute of Applied Mathematics, 

Middle East Technical University, ODTÜ, 06531, Ankara, 

Turkey, gweber@metu.edu.tr 

This presentation contributes to an improved modeling, risk management and 

optimization of financial processes, with environmental aspects included. We 

model stochastic differential equations and estimate parameters with the help 

of (generalized) additive models, splines, our new CMARS and nonlinear regression. 

We embed the related goals on accuracy and stability with the minimization 

of risk and the maximization of utility, by generalized approaches in 

portfolio optimization. With a discussion of structure, frontiers and an outlook 

we conclude. 

2 - Stability investigation of key Czech banks by means of 

the financial modelling 

Petr Gurný, Department of Finance, VSB -TU Ostrava, Sokolska 

tr. 33, 70121, Ostrava, Czech Republic, petr.gurny@vsb.cz 

The paper is devoted to the investigation of the Czech banks health, which we 

can regard as one of the most important tasks in time of the financial crisis. The 

main goal of the paper is an estimation of the future probability of default (PD) 

for three key Czech banks. At first the revised model (built on the basis of the 

linear discriminant analysis) for prediction of bank failure will be presented. 

Afterwards the relevant financial indicators needed for estimation of the future 

PD will be simulated via Lévy processes (VG, NIG) and their dependencies 

will be captured via copula functions. 

77


3 - Modelling Financial Bubbles by Stochastic and Ellipsoidal 

Calculus and Copulas 

Efsun Kürüm, Financial Mathematics, Institute of Applied 

Mathematics, METU, Institute of Applied Mathematics, 06531, 

Ankara, Turkey, efsun.kurum@gmail.com, Gerhard-Wilhelm 

Weber 

To diagnose bubbles is a very important issue in finance, where copulas have 

been considered the last decade. By copulas we model bubbles which are 

known as upward price movements over extended range that implode. In 

copula-based modelling we deal with bubbles by stochastic differential equations. 

We also model the dynamics of bubbles by ellipsoidal calculus and topology. 

We test the model by a large database from a central bank. We conclude 

with a structural features, discussions of future applications and research challenges. 

4 - Determination of Trade Based Manipulation in Istanbul 

Stock Exchange 

Melik Kamisli, Bozuyuk Vocational School, Bilecik University, 

Bilecik University Vocational School, Yeni Mahalle Cerrahlar 

Sokak No:11, 11300 , Bozuyuk/Bilecik, Turkey, 

melik.kamisli@bilecik.edu.tr, Nuray Girginer 

The trade based manipulation has negative effects on investors, stock market 

and so, depending on them on whole economy. Consequently, a study based 

on determination of manipulation will provide information to related individuals. 

The aim of this study is evaluating the usability of financial ratios in trade 

based manipulation as an indicator when the investors make the stock selection 

decision. The results show that, Return on Assets and Book Value per Share are 

important financial ratios to determine the trade based manipulation in Istanbul 

Stock Exchange. 

� MD-26 


3.1.11 

Machine Learning for Multiple Sources 



Chair: Sureyya Ozogur-Akyuz, Department of Mathematics and 

Computer Science, Bahcesehir University, Bahcesehir University, 

Dept of Mathematics and Computer Science, Cıragan cad. Besiktas, 

34353, Istanbul, Turkey, sureyya.akyuz@bahcesehir.edu.tr 

Chair: Zakria Hussain, Computer Science, Unvertsity College 

London, Gower Street, WC1E 6BT, London, United Kingdom, 

Z.Hussain@cs.ucl.ac.uk 

1 - Exploring relations of entities via common space representation 

Sandor Szedmak, ISIS, Electronics and Computer Science, 

University of Southampton, Highfield, Building 1, SO17 1BJ, 

Southampton, United Kingdom, szedmak777@gmail.com 

We are facing the following problem, there are given classes of entities with 

different distributions, e.g. enzymes, chemical reactions controlled by the enzymes, 

chemical molecules participating in the reactions, and medical diagnostic 

data. The question is that: how can one explore the highly complex 

interdependencies of these entities and predict the relations of new unseen entities? 

Directly attacking problems with several data sources by the known 

machine learning methods is mostly intractable. Several alternatives have been 

developed to visualise the potential relations between the entities by embedding 

them into a common space, e.g. recommendation systems, but the learning capabilities 

of these approaches are rarely addressed . Our question is that: can 

we learn the common space and its structure to provide a learner with good predicting 

power and with acceptable computational complexity? Simple examples 

can show that the known learning methods, e.g. Support Vector Machine, 

can be reinterpreted as common space problem of the input and output entities. 

Similarly the structural learning can be described in this way too. These 

methods can then be reduced into a nearest neighbour predictor working in a 

properly chosen common space. Going further on this road an approach which 

is able to handle incomplete data sources is presented as well. 

2 - Selecting Rows and Columns of Large Datasets for 

Training Support Vector Regression Models 

78 

Kübra Yaman, Graduate School of Science and Engineering, Koc 

University, Rumelifeneri Yolu, 34450 Sarıyer, Istanbul, Turkey, 

kyaman@ku.edu.tr, Ozden Gur Ali 

We propose the Row and Column Selection Algorithm to identify the most informative 

points and variable subsets from large datasets to train SVR models. 

This algorithm consists of two steps. First we identify few support vectors by 

penalizing the support vector weights, next we select a subset of features by penalizing 

the feature weights in linear models. We show that the test set accuracy 

of RBF-SVR model trained on the set selected by the algorithm is significantly 

better than the accuracy of the same model trained on the benchmark random 

sampling model. 

3 - Adapted Infinite Kernel Learning by Multi-Local Algorithm 

Sureyya Ozogur-Akyuz, Department of Mathematics and 

Computer Science, Bahcesehir University, Bahcesehir 

University, Dept of Mathematics and Computer Science, Cıragan 

cad. Besiktas, 34353, Istanbul, Turkey, 

sureyya.akyuz@bahcesehir.edu.tr, Gürkan Üstünkar, 

Gerhard-Wilhelm Weber 

The interplay of machine learning (ML) and optimization methods is an emerging 

field of artificial intelligence. Kernel selection for classification models is 

an important task for ML algorithms. One of the recent kernel learning methods 

is developed for multiple source data called infinite kernel learning (IKL) modelled 

by semi-infinite optimization.In this study, we improved IKL by multi 

local algorithm and adaptive simulated annealing (ASA) algorithm. Experiments 

on high dimensional biological data show significant improvement both 

in running time and error percentage. 

4 - Sparse Convex One-Class Kernel Fisher Discriminant 

Analysis 

Tom Diethe, Department of Computer Science, University 

College London, Gower Street, WC1E 6BT, London, United 

Kingdom, t.diethe@cs.ucl.ac.uk, Janaina Mourão-Miranda, John 

Shawe-Taylor 

Outlier detection is a classical topic in robust statistics. This paper describes 

a convex formulation of One-Class Fisher Discriminant Analysis (OC-FDA), 

which can be solved using off-the-shelf optimisers. Sparsity is enforced 

through an l1-norm constraint on the weight vector. The size of the enclosing 

hypersphere is varied using quantile values of the estimated posterior, adjusted 

by a single parameter. We compare with the One-Class SVM (OC-SVM) on 

a toy example and a real-world neuroimaging dataset. The extension to the 

multiview setting is also given, with some empirical results. 

� MD-27 


8.2.06 




Chair: Christina Erlwein, Department of Financial Mathematics, 

Fraunhofer ITWM, Fraunhofer Platz 1, 67663, Kaiserslautern, 

Germany, christina.erlwein@itwm.fraunhofer.de 

1 - Sharpe Ratios and Implied Risk Free Rates 

Michael Best, C&O, University of Waterloo, N2L 2H43G1, 

Waterloo, ON, Canada, mjbest@uwaterloo.ca 

For a mean-variance Markowitz portfolio optimization model having just a budget 

constraint, the choice of a market portfolio implies a unique risk free rate. 

However, practical portfolio optimization problems have a variety of additional 

linear inequality constraints. These constraints may cause kinks or points of 

non-differentiability on the resulting efficient frontier. In terms of linear programming 

terminology, these kinks are principally caused by extreme points. 

We show that choosing a kink point as a market portfolio results in a continuum 

of implied risk free rates and give a formula for them. 

2 - Optimal Trading of Algorithmic Orders in a Liquidity 

Fragmented Market Place 

Natasa Krejic, Department of Mathematics and Informatics, 

University of Novi Sad, Trg Dositeja Obradovica 4, 21000, Novi 

Sad, Serbia, natasak@uns.ac.rs, Miles Kumaresan

An optimization model for execution of atomic orders at multiple trading 

venues is proposed and analyzed. The optimal trajectory consists of both market 

and limit orders, and takes advantage of any price or liquidity improvement 

at a particular market. The complexity of multi-market environment yields a 

bi-level nonlinear optimization problem. The lower level problem admits a 

unique solution and hence the second order conditions are satisfied under a set 

of reasonable assumptions. The model is computationally affordable and the 

simulation results show its effectiveness using real trade data. 

3 - HMM-based investment strategies for asset allocation 

Christina Erlwein, Department of Financial Mathematics, 

Fraunhofer ITWM, Fraunhofer Platz 1, 67663, Kaiserslautern, 

Germany, christina.erlwein@itwm.fraunhofer.de, Rogemar 

Mamon, Matt Davison 

We develop and analyse investment strategies based on hidden Markov models. 

Filtering techniques are utilised to decide whether to invest in growth or 

value stocks. We develop two investment strategies based on filtered information 

within the multi-dimensional HMM, one switching strategy and one mixed 

strategy. Using datasets on Russell 3000 growth and value indices from 1995 

to 2008, the switching strategy yields high Sharpe ratios compared to those 

obtained from pure index strategies and the mixed strategy. The performance 

of the mixed strategy is compared to the classical mean-variance approach. A 

simulation analysis further shows a higher performance stability of the HMM 

strategies. 

� MD-28 



Stochastic Programming Algorithms 



Chair: Katarina Vla, Department of Mathemtics and Informatics, 

University of Novi Sad, Trg Dositeja Obradovica 3, 21000, Novi Sad, 

Serbia, katarinav@dmi.uns.ac.rs 

1 - Scenarios and events in multi-stage stochastic linear 

programming 

Cesar Beltran-Royo, Estadística e Investigación Operativa, 

Universidad Rey Juan Carlos, Calle Tulipán, s/n, 28933 , 

Móstoles, Madrid, Spain, cesar.beltran@urjc.es, Laureano 

Fernando Escudero, Romy-Elena Rodriguez-Ravines 

To solve the multi-stage linear programming problem, one may use a deterministic 

or a stochastic approach. The drawbacks of the two techniques are 

well known: the deterministic approach is unrealistic under uncertainty and the 

stochastic approach suffers from scenario explosion. We introduce a new technique 

whose objective is to overcome both drawbacks. The focus of this technique 

is on events instead of scenarios and for this reason we call it multi-stage 

Event Linear Programming (MELP). As we show in our results, the MELP 

approach represents a promising compromise between the stochastic and the 

deterministic approach, regarding capacity to deal with uncertainty and computational 

tractability. 

2 - Solution Procedures for Probabilistically Constrained 

Problems 

Ebru Mevlude Angun, Industrial Engineering Department, 

Galatasaray University, Ciragan Cad. Ortakoy, 34357, Istanbul, 

Turkey, eangun@gsu.edu.tr 

In this talk, we consider stochastic programming problems with probabilistic 

constraints. These problems are usually very hard to solve, because the feasible 

set defined by even a single probabilistic constraint may happen to be nonconvex. 

We consider the probabilistically constrained problem and its convexified 

version, and apply scenario approximation and sample average approximation 

methods combined with deterministic optimization procedures. We apply these 

approaches to a disaster management problem, and obtain interesting numerical 

results. 

3 - Integer linear stochastic programming with multiple objective 

Amrouche Salima, Mathématiques, université Saad Dahlab 

Blida, Université SAAD DAHLAB Blida Route De Soumaa BP 


270 BLIDA, amrouchesalima@gmail.com, Algeria, 

amrouchesalima@gmail.com 

The real life decision problems have three main properties. The first one is 

to have conflicting objectives in the problem structure, the second one is the 

stochasticity in the description of problem parameters in contexts where the 

probability distribution of random parameters is known. and the last one is due 

to involvement of integer decision variables which increased dimension of the 

problem. Multiobjective nature with discrete variables and imprecise parameters 

make the mathematical expression of the problems harder to solve with 

the traditional approache. An efficient algorithm is developed, via extending 

the well-known L-shaped method using generalized benders decomposition to 

efficiently handle the integer variables in the first stage and the integer recourse 

in the second stage of the model formulation. The proposed modeling and solution 

methods able to identify all the integer feasible solutions that are efficient 

for the problem in a finite number of iterations. 

4 - Sample-Path Approach to Constrained Optimization 

Katarina Vla, Department of Mathemtics and Informatics, 


Sad, Serbia, katarinav@dmi.uns.ac.rs, Natasa Krejic, Natasa 

Krklec 

We consider simulated response problem with constraints where the objective 

function is computed by taking expectation over the sample response function 

and has no explicit form. Therefore the gradient information is also unavailable. 

The constraints are given as simple box constraints. We use the samplepath 

method within derivative free approach. A sequence of local quadratic 

models that approximate the objective function is constructed by interpolation. 

The sample sizes varies and is determined using estimations of different kind. 

Convergence is analyzed and numerical results are discussed 

� MD-29 


8.2.11 

Boolean Optimization in Graph Theory 



Chair: Martin Milanic, FAMNIT, University of Primorska, 

Glagoljaska 8, 6000, Koper, Slovenia, martin.milanic@upr.si 

1 - Optimizing Boolean functions for cryptographic applications 

Enes Pasalic, FAMNIT, University of Primorska, Glagoljaska 8, 

6000, Koper, Slovenia, enespasalic@yahoo.se 

In this talk certain optimization problems related to the design of Boolean functions 

used in cryptography are discussed. Boolean functions are basic cryptographic 

primitives in the design of symmetric cryptographic systems, such as 

stream and block ciphers, and many other cryptographic algorithms. Its design 

is closely related to certain combinatorial problems, among others to integer 

optimization methods and some problems related to graph theory. We briefly 

overview these representations, with the emphasis on the translating a Boolean 

optimization problem into a graph theoretical problem. 

2 - On some graph classes related to perfect graphs: a survey 

Flavia Bonomo, Computer Science, University of Buenos Aires, 

1428, Buenos Aires, Argentina, fbonomo@dc.uba.ar, Guillermo 

Duran, Martin Safe, Annegret Katrin Wagler 

Perfect graphs form a well-known class of graphs introduced by Berge in the 

1960s in terms of a min-max type equality involving two famous graph parameters. 

In this work, we study variations and subclasses of perfect graphs 

defined by means of min-max relations of other graph parameters. Our focus 

is on clique-perfect and coordinated graphs. We show the connection between 

these graph classes and both hypergraph theory and the clique graph operator. 

In this talk, we will present previous results, some new contributions and the 

main open problems. 

3 - Scheduling the Argentine volleyball league: A realworld 

application of the Traveling Tournament Problem 

with couples of teams 

Guillermo Duran, Ingenieria Industrial, University of Chile, 

Republica 701, 1000, Santiago, Chile, gduran@dii.uchile.cl, 

Flavia Bonomo, Javier Marenco, Daniela Saban 

79


We describe the process for designing the fixture of the Argentine first division 

volleyball league. This league is composed by 12 teams grouped into couples 

of teams. The minimization of the travel distances is an important task, since 

the teams are located throughout the country, hence this problem is a variation 

of the well-known Traveling Tournament Problem. We have applied integer 

programming techniques and a tabu search heuristic to tackle these issues, and 

the resulting fixtures have been successfully used in the 2007/2008, 2008/2009, 

and 2009/2010 leagues. 

4 - Complexity results for equistable graphs and related 

classes 

Martin Milanic, FAMNIT, University of Primorska, Glagoljaska 

8, 6000, Koper, Slovenia, martin.milanic@upr.si, Jim Orlin, 

Gabor Rudolf 

The class of equistable graphs is defined by the existence of a cost structure on 

the vertices such that the maximal stable sets are characterized by their costs. 

In this talk, we will discuss some complexity results for equistable graphs and 

related classes. A simple pseudo-polynomial-time dynamic programming algorithm 

is given that solves the maximum weight stable set problem along with 

the weighted independent domination problem in equistable graphs. These results 

are obtained within the wider context of Boolean optimization; corresponding 

hardness results are also provided. 

� MD-30 


8.2.13 

MCDA II: Sorting Models, theoretical 

aspects and other issues. 


other Issues 


Chair: Constantin Zopounidis, Dept. of Production Engineering and 

Management, Technical University of Crete, University Campus, 

73100, Chania, Greece, kostas@dpem.tuc.gr 

Chair: Michael Doumpos, Dept. of Production Engineering and 


73100, Chania, Greece, mdoumpos@dpem.tuc.gr 

1 - Multi-criteria relational clustering: An empirical analysis 

of the relation’s properties 

Julien Roland, CoDE-SMG, Université Libre de Bruxelles, 1050, 

Bruxelles, Belgium, julien.roland@ulb.ac.be, Stefan Eppe, Yves 

De Smet 

Recently, the authors have proposed a method to address the problem of relational 

multi-criteria clustering. In this presentation, we will focus on the 

properties of the relations between the clusters. Based on artificial and real 

data sets, we will highlight some pathological results that will help us define 

desirable features. Properties such as transitivity and totality are helpful to simplify 

the structure of a relational partition. Finally, this discussion will lead to 

identify ways to improve the current method. 

2 - Learning Non-monotone Additive Value Functions for 

Multicriteria Decision Making 

Michael Doumpos, Dept. of Production Engineering and 


73100, Chania, Greece, mdoumpos@dpem.tuc.gr 

Multiattribute additive value functions constitute an important class of models 

for multicriteria decision making. Such models are often used to rank a set of 

alternatives or to classify them into pre-defined groups. Disaggregation techniques 

have been used to construct such models using linear programming techniques 

based on the assumption of monotone preferences. This paper presents 

a methodology to construct non-monotone models for classification purposes, 

using an evolutionary optimization approach. 

3 - A group decision making procedure and its incorporation 

in the MMASSI/TI DSS. 

80 

Teresa Pereira, Engenharia e Gestão Industrial, ESEIG-IPP 

(Escola Superior de Estudos Industriais e de Gestão do Instituto 

Politécnico do Porto, Rua D. Sancho I, 981, 4480-771, Vila do 

Conde, Porto, Portugal, teresapereira@eu.ipp.pt, Sameiro 

Carvalho 

MMASSI/TI — Multi-criteria Methodology for Selection of Information Technologies 

— is a DSS that incorporates a multi-criteria model to support the 

decision making process. The DSS was tested in different organizations contexts. 

Although the DSS has shown effectiveness in the support of decision 

making, one of the main difficulties encountered in its application was related 

with group decision support. We present a group decision making procedure 

and its incorporation in the MMASSI/TI DSS. 

4 - Negotiation template analysis with ELECTRE-TRI based 

scoring system 

Tomasz Wachowicz, Operations Research, Karol Adamiecki 

University of Economics in Katowice, Katowice, Poland, 

tomasz.wachowicz@ae.katowice.pl 

Electronic negotiation experiments show that negotiators quite often reject the 

notion of compensation that is fundamental to additive scoring system most frequently 

used for evaluation of negotiation template. We present a novel method 

rejecting compensation that derives from ELECTRE-TRI model, which parameters 

are determined basing on the examples evaluation. To increase the 

precision of the scoring system the limiting profiles are scored in terms of satisfaction 

units and the whole negotiation space is calibrated and divided into 

sub-categories. Simple supporting software is also presented. 

� MD-31 


8.2.15 

Societal Complexity and Education 



Chair: Dorien DeTombe, Methodology of Societal Complexity, Chair 

Euro Working Group, P.O.Box 3286, 1001 AB , Amsterdam, 

Netherlands, detombe@nosmo.nl 

1 - Religious-Political Conflict Resolution 

Cathal Brugha, Management Information Systems, University 



This paper uses Nomology, the science of the laws of the mind, to explain 

Religious-Political Conflict Resolution, in terms of an Adapting system, aspects 

of which include Body, Mind, Soul and Spirit. It suggests that people of 

all religions and none will have to do a lot of "soul-searching’ before progress 

can be made to reduce the causes of religious-political conflict. The paper 

endorses the direction of the discussion about the separation of church and 

state that took a path from Hegel to Rawls, to Habermas, and uses the metaframework 

to propose where the discussion should go in the future. It suggests 

that this will involve cooperative inter-religious examination and challenging 

of religious texts. 

2 - Basic Social Math 

Jared Hanson, Saïd Business School, Oxford University, Al 

Yamamah University, PO Box 45180, 11512, Riyadh, Saudi 

Arabia, jared.hanson@sbs.ox.ac.uk 

In order to establish a new perspective that can foster innovation in the management 

sciences, better tools are needed at the fundamental level of math that is 

used for social analysis. This project seeks to consolidate principles and techniques 

used in the more sophisticated quantitative methods of Operations Research 

and extend these tools downward to their fundamental, but still rigorous, 

level where they cross into the linguistic frameworks that are used to interpret 

and understand data in managerial, financial, and educational systems. 

� MD-32 


8.2.17 

OR in Forestry III 



Chair: Ola Eriksson, Forest Rescource Management, SLU, 

Skogsmarksgränd, SE-901 83, Umeå, Västerbotten, Sweden, 

ola.eriksson@srh.slu.se

1 - A stochastic dynamic programming approach to optimize 

short-rotation coppice systems management 

scheduling under fire risk. An application to eucalypt 

plantations. 

Liliana Ferreira, Rua do Convívio, n o 156 Telheiro - Barreira, 

2410-333 , Leiria, liliana.ferreira@estg.ipleiria.pt, Miguel 

Constantino, Jose Borges, Jordi Garcia_Gonzalo 

We present a management scheduling model for short rotation coppice systems 

that may take into account the risk of wildfire. A stochastic dynamic programming 

(SDP) is proposed to determine the policy (e.g. fuel treatment, stool 

thinning, coppice cycles and rotation length) that maximizes expected financial 

revenues, considering alternative wildfire occurrence and damage scenarios. 

SDP stages are defined by the number of harvests and state variables include 

the number of years since the plantation. A typical Eucalyptus globulus Labill 

stand in Central Portugal was used as a test case. 

2 - Integrating fire risk considerations in landscape level 

forest planning 

José Gonzaléz-Olabarria, Centre Tecnològic Forestal de 

Catalunya, 25280, Solsona, Spain, jr.gonzalez@ctfc.es, Timo 

Pukkala 

In this study, economic revenue and the overall fire resistance of a landscape, 

were combined to generate optimal forest plans for a forest landscape in Catalonia. 

The risk of fire was integrated into the economic objective by considering 

potential fire, and a fire resistance index was calculated for each stand from the 

properties of the stand and its adjacent neighbors. The optimal forest plans for 

each combination of objectives were generated using simulated annealing as 

optimization method, and the resistance to fire of the resulting landscapes were 

tested using a fire spread simulator. 

3 - Accountability on forest fires: Evidence of social responsibility 

Rute Abreu, Business and Economics, Guarda Polythecnic 

Institute, 6300-559, Guarda, ra@ipg.pt, Fátima David 

This research studies the influence of social responsibility to forest management 

and explores, direct and indirect, consequences of forest fires in Portugal 

that promotes the development of pioneering practices and bring greater accountability, 

improve transparency and increase sustainability. This research 

presents dual theoretical framework: organisational and accountability and assures 

the link of SR with forest protection and environment preservation. The 

trilogy: prevention, detection and surveillance provides explanations for economic 

and social decisions of forest management. 

4 - European forest decision support systems - what do we 

have and what is missing? 

Ola Eriksson, Forest Rescource Management, SLU, 

Skogsmarksgränd, SE-901 83, Umeå, Västerbotten, Sweden, 

ola.eriksson@srh.slu.se 

The FORSYS COST Action has set out to define requirements for forest decision 

support system (DSS) implementation and use. This involves, among 

other things, to give a comprehensive account of existing forest DSSs. At the 

moment more than 50 systems are registered. This forms the basis for an analysis, 

with a European focus, of what kind of planning problems the DSSs deal 

with, what planning phases they support, with what methods they do it, and 

what seems to be missing capabilities. The presentation also includes a short 

oversight of FORSYS. 

� MD-33 


8.2.19 

Environmental Management I 



Chair: Marta Castilho Gomes, CESUR, Instituto Superior Técnico, 

Av. Rovisco Pais, 1049-001, Lisboa, Portugal, 

marta.gomes@ist.utl.pt 

1 - Study of Electronic Ballast for Lighting Systems from 

the viewpoint of Electromagnetic Compatibility and 


Power Quality issues: A contribution to Efficiency Analysis 

William Vianna, Production Engineering, Federal University at 

Santa Catarina (UFSC) - BRAZIL, Rua Cônego Bernardo, 100. 

ap. 202, Trindade, 88036570, Florianópolis, Santa Catarina, 

Brazil, wpwilliam@hotmail.com, Raul Eduardo Fernandez 

The study aims to reduce the impact of electronic ballasts for lighting systems 

used in large-scale business, commercial and residential buildings with a view 

to optimizing the technical and economic such equipment from the viewpoint 

of electromagnetic compatibility and power quality. The results evidence noncompliance 

of applicants conducted electromagnetic emissions, the need for 

improvement in product designs and increased legal and industrial regulation 

and monitoring. 

2 - A cost optimization model for hazardous medical waste 

management 

Marta Castilho Gomes, CESUR, Instituto Superior Técnico, Av. 

Rovisco Pais, 1049-001, Lisboa, Portugal, 

marta.gomes@ist.utl.pt, João Nunes de Almeida, João 

Quinhones Levy 

This work presents a Mixed Integer Linear Programming model for managing 

hazardous medical waste (HMW). In Portugal, HMW is divided into two 

groups depending on whether it must be incinerated or decontaminated before 

being transported to landfills. Both groups are considered in the model, which 

optimizes construction and operating costs regarding the location of facilities 

(transfer stations and disposal sites) as well as the waste flows between the 

nodes. Real data were used and the results may support decision regarding the 

urgent need to expand the HMW incineration capacity in the country. 

3 - Integral model of entrepreneurship for regional waste 

management 

Zoran Rakicevic, Faculty of Organization scinece, Belgrade 

University, Jove Ilica 154, 11000, Belgrade, Serbia, 

zoran.rakicevic@gmail.com, Jasmina Omerbegovic-Bijelovic 

Waste management (WM) is becoming increasingly important — because of 

draining of the natural resources and of polluting the environment, as well as 

the opportunities for creating new value/jobs and for networking. This paper 

points out three levels of networking models for WM: global, regional and local, 

each one of which has its efficiency and effectiveness, „its" stakeholders 

(interests); this paper shows also how it is possible to incorporate aims and 

limitations of the global and local in regional WM-models. Also, it points to 

the concept of entrepreneurial value creation. 

� MD-34 


8.2.23 

Model Formulations and Real World 

Applications of Lot Sizing and Scheduling I 


Quantity 


Chair: Alistair Clark, Dept of Mathematics and Statistics, University 

of the West of England, Frenchay Campus, Coldharbour Lane, BS16 

1QY, Bristol, United Kingdom, Alistair.Clark@uwe.ac.uk 

1 - A multi-objective mixed integer linear programming approach 

to shelf-life integrated lot-sizing and scheduling 

in yogurt production 

Pedro Amorim, Industrial Engineering and Management, Faculty 

of Engineering of Porto University, Rua Doutor Roberto Frias, 

4200-465, Porto, Portugal, amorim.pedro@fe.up.pt, Bernardo 

Almada-Lobo 

The lot sizing and scheduling problem with parallel lines and shared buffers, 

arising in a fresh diary plant producing yogurt, is addressed in this work. Yogurt 

production is very complex with perishable intermediate products and final 

products pushed towards just-in-time due to freshness. To target this problem 

a multi-objective mixed integer linear programming model is proposed. The 

model takes into account the main concerns in yogurt production: integration 

of shelf-life, sequence-dependent times and costs, limitations of the fermentation 

phase, inventory costs and operational efficiency. 

81


2 - Solving the stochastic single-level capacitated lot sizing 

problem with a scenario approach 

Florian Sahling, Department of Production Management, Leibniz 

University Hannover, Koenigsworther Platz 1, 30167, Hannover, 

Germany, sahling@prod.uni-hannover.de, Stefan Helber, Katja 

Schimmelpfeng 

We present a non-linear model formulation for the stochastic single-level, 

multi-product dynamic lot sizing problem with a gamma service level constraint. 

Based on demand forecasts, the subject is to determine an efficient, 

robust and stable production schedule which minimizes the expected setup and 

holding costs. The proposed model formulation is approximated by a mixedinteger 

linear program using a scenario approach. A numerical investigation 

of synthetic problem instances shows under which conditions precise demand 

forecasts are particularly useful. 

3 - A Multi-Level Capacitated Lot Sizing Problem considering 

discounts, set-ups and line assignment using a MIP 

based heuristic and Column Generation 

Philipp Thurnher, V-Research Center for Tribotronics and 

Technical Logistics, Stadtstraße 33, 6850, Dornbirn, Vorarlberg, 

Austria, Philipp.Thurnher@v-research.at, Hubert Missbauer 

We develop a MLCLSP based on a case study of a metalworking company. The 

model has two levels where raw materials are ordered regarding discounts and 

products are made at high set-up costs on five partly exchangeable lines constrained 

by production times and stock capacity. The decomposition is done using 

column generation. Columns are generated applying both a Dantzig-Wolfe 

and a Lagrangean decomposition as dual coordination techniques. Additionally 

an LP-Relaxation as primal coordination is used to allocate capacities to 

the sub-problems. The sub-problems are solved with CPLEX. 

4 - Integrated production planning of a portuguese pulp 

and paper mill 

Maristela Santos, Department of Applied Mathematics and 

Statistics, University of Sao Paulo, Av. Trabalhador 

São-carlense, 400 - Centro, Caixa Postal: 668, 13560-970, São 

Carlos, São Paulo, Brazil, mari@icmc.usp.br, Andre Alves, 

Bernardo Almada-Lobo 

We deal with a production planning problem that arises in an integrated Portuguese 

Pulp and Paper mill. We propose a new model considering several critical 

production units such as the digester, paper machine, tanks and the units 

designed to produce energy. Simple Mathheuristics are developed to obtain 

feasible solutions that enable the analysis of the production process as a whole. 

The proposed approaches are tested on real-world data. The expected cost savings 

and the performance of the generated plans are benchmarked against real 

ones. 

� MD-35 


6.2.46 

Soft OR and Problem Structuring Methods 

III 



Chair: Vesna Cancer, Faculty of Economics and Business, University 

of Maribor, Razlagova 14, 2000, Maribor, Slovenia, 

vesna.cancer@uni-mb.si 

Chair: Victor Vidal, IMM, Technical University of Denmark, 2800, 

Lyngby, Denmark, vvv@imm.dtu.dk 

1 - Creative Thinking Techniques in Multi-Criteria 

Decision-Making 

Vesna Cancer, Faculty of Economics and Business, University of 

Maribor, Razlagova 14, 2000, Maribor, Slovenia, 

vesna.cancer@uni-mb.si 

The paper presents several example cases of business Creative Problem Solving 

showing the mutual assistance of the creative thinking tools and approaches for 

probortunity finding and definition, for the generating, decomposing and analyzing 

of ideas, and the multi-criteria decision-making methods for their evaluation, 

selection and verification. Creative approaches are not limited to problem 

definition and problem structuring only. This paper introduces the use of the 

techniques based on questions, with an emphasis on the 5Ws&H Technique, in 

the establishing of the criteria weights. 

82 

2 - Innovation, Creativity and Soft Systems Methodology 

Giles Hindle, Operational Research, Warwick Business School, 

University of Warwick, CV4 7AL, Coventry, United Kingdom, 

giles.hindle@wbs.ac.uk 

Soft Systems Methodology is characterised as an approach to tackling messy, 

ill-structured problems. However, a recent set of organisational innovation 

projects, plus my own sympathy with Ackoff’s arguments for the utility of 

Idealized Design, has led me to explore the use of design thinking within SSM. 

I argue the expression stage using Rich Picturing is vital for helping participants 

"step back’ and look holistically at the situation. This secures a platform 

for creativity. The modelling stage of SSM is then perfectly suited to support 

organisational and process innovation through design thinking. Case vignettes 

are presented based upon projects from both the public and private sectors. 

3 - Systemic strategy development and implementation 

using the example of a lead brokerage financial service 

provider 

Paul Flachskampf, Institut for Management Cybernetics, 

Schurzelterstr. 25, 52074, Aachen, NRW, Germany, 

paul.flachskampf@ifu.rwth-aachen.de 

This paper examines systemic strategy development using the example of a 

lead brokerage financial service provider. To do so, different elements of several 

systemic and cybernetic induced theories are combined to a three step approach: 

1) system diagnosis to derive redesign actions, 2) identification of critical 

system variables, 3) creation of a modular software prototype with the help 

of a control loop model. This approach is induced by action research building 

on an interactive inquiry process that balances problem solving and actions 

implemented in a collaborative context. 

4 - On the dynamics of Value Networks 

Bent Erik Bakken, St. Georgs vei 4, 0280 OSLO, 0280 OSLO, 

Norway, Oslo, beerikba@online.no 

Corporate strategic research increasingly question the ubiquity of value chains 

as drivers for strategic design. Value networks has been proposed as a an alternative 

useful configuration. This paper shows that value chains have been a 

mainstay of system dynamics applications, from "Industrial Dynamics" (Forrester, 

1961) onwards, yet value chains show even more complex dynamics. 

Implications for reserach in strategy and systems dynamics as well as for practice 

is discussed. 

� MD-36 


3.1.05 

Fuzzy Optimization 


Intelligence 


Chair: Joao Miguel da Costa Sousa, Dept. of Mechanical 

Engineering, Technical University of Lisbon, 1049-001 , Lisboa, 

Portugal, jmsousa@ist.utl.pt 

Chair: Susana Vieira, Center of Intelligent Systems - IDMEC, 

Instituto Superior Técnico, Pav. Eng. Mec. III - sala 3.8, Av. Rovisco 

Pais, 1049-001, Lisbon, Portugal, susana@dem.ist.utl.pt 

1 - Multi-objective Optimization of Feature Selection Problems 

using Fuzzy Models 

Ozlem Turksen, Dept. of Mechanical Engineering, 

CIS/IDMEC-LAETA, TU Lisbon, Instituto Superior Técnico, 

Av. Rovisco Pais, 1049-001, Lisbon, Portugal, 

turksen@science.ankara.edu.tr, Susana Vieira, Joao Miguel da 

Costa Sousa, Aysen Apaydin 

Feature selection is one of the most important techniques in data preprocessing 

for data mining. In this paper, the feature selection optimization problem 

is defined as a multi-objective problem addressing simultaneously two goals: 

minimize classification error and reduce the number of features. The problem 

is addressed by using an NSGA II approach to find the multiple Pareto optimal 

solutions. Therefore, in this paper the solutions are evaluated using fuzzy models, 

as they are universal approximators and can be interpretable under certain 

conditions.

2 - Fuzzy Criteria for Multi-objective Ant Feature Selection 

Susana Vieira, Center of Intelligent Systems - IDMEC, Instituto 

Superior Técnico, Pav. Eng. Mec. III - sala 3.8, Av. Rovisco 

Pais, 1049-001, Lisbon, Portugal, susana@dem.ist.utl.pt, Joao 

Miguel da Costa Sousa 

Feature selection has been an active research area in data mining, pattern recognition 

and statistics communities. In this paper, a multi objective ant feature 

selection algorithm is proposed to cope with the difficulty of finding optimal 

solutions for the feature selection optimization problem, and is compared with 

the previous aggregated multi-criteria approach. The objectives are described 

using fuzzy optimization, since it allows for an easier and more transparent 

description of the criteria used in the feature selection process. 

3 - A New Filter-wrapper Method for Large Databases by 

Optimizing Fuzzy Classification Models 

Joao Miguel da Costa Sousa, Dept. of Mechanical Engineering, 

Technical University of Lisbon, 1049-001 , Lisboa, Portugal, 

jmsousa@ist.utl.pt, Federico Cismondi, Andre Fialho, Susana 

Vieira, Shane Reti, Michael Howell, Stan Finkelstein 

Large databases require significant and frequently impractical computer processing 

times while selecting the relevant subset of features. We propose using 

first a filter method to reduce the number of features, and then applying a 

wrapper method to develop an accurate classification model. Tested filter methods 

were Fisher’s rank, variance and entropy, while used wrappers were a tree 

search bottom-up and an ant colony feature selection. The results for a fuzzy 

classification algorithm show better computational times and classification performance 

when filters are used preceding wrappers. 

4 - An Integrated Evaluation Framework for Customization 

Strategies 

Gulcin Buyukozkan, Industrial Engineering, Galatasaray 

University, Ciragan cad. no.36, 34357, Istanbul, Turkey, 

gulcin.buyukozkan@gmail.com 

The need for product customization during product development (PD) processes 

continues to increase. Mass customization (MC) relates to the ability 

to provide customized products or services through flexible processes in high 

volumes and at reasonably low costs. There exist different strategies for the degree 

of customization in PD. Determining the right strategy of MC is essential 

to the competitiveness of a company. To support managerial decision making, 

this study proposes an integrated strategic evaluation framework to assess 

effectively MC strategies in PD process with a real case study. 

� MD-37 


3.1.09 

OR for Development and Developing 

Countries I 






1 - Passengers air transport system with mini and major 

hubs: a case study applied to Brazil 

Nelio D Pizzolato, Industrial Engineering, Catholic University of 

Rio de Janeiro, R. Marques de Sao Vicente, 225, 22453-900, Rio 

de Janeiro, RJ, Brazil, ndp@puc-rio.br, Rafael M. A. de 

Figueiredo, Morton E. O‘Kelly, Madiagne Diallo 

This research deals with a hub-and-spoke model that includes mini hubs. In 

an extended country such as Brazil, in which the population is mostly concentrated 

around a few areas, while large extensions generate low volumes but 

long distances, regional mini hubs might be a way to make the operation more 

economical and attractive for the passengers. This paper has two purposes: one 

is to identify regional clusters in order to locate mini hubs, and second is to 

articulate these with countrywide major hubs. 


2 - Engineering research in developing countries, some reflections 

based on scientific publications 

Víctor Bucheli Guerrero, Industrial Engineering, Universidad de 

los Andes, Carrera 1 N 18A - 12, Bobotá, D.C., Colombia, 

vbucheli@uniandes.edu.co, Roberto Zarama Urdaneta 

In developing countries universities have important challenges. A key challenge 

is to increase their indexed publications. In this paper we study the dynamics 

of the knowledge production in engineering and their relations with 

context. With that in mind, we observed the Academic Ranking of World Universities 

in Engineering/Technology and Computer Sciences (ARWU-ENG). 

For this purpose, we construct a Bayesian Network Model to explore some 

strategies to increase the scientific performance. Based on the model we propose 

a journal selection strategy. We illustrate the strategy with a case study. 

3 - Dynamic modelling of the usage of a rural community 

health centre in a developing country: the growth of 

trust 

Honora Smith, School of Mathematics, University of 


United Kingdom, honora.smith@soton.ac.uk, Paul Harper 

Community health centres are being used in the developing world as a means 

of delivering accessible, low-cost service. A major management issue is level 

of uptake of services and its effect on sustainability. We present a model for 

spatio-temporal spread of usage of a community health centre, in a rural developing 

region. We include a trust factor in the modelling. Trust in the provider 

is built through word-of-mouth contacts and development activities. As a case 

study, we analyse data collected a community health centre in a rural region of 

northern India. 

� MD-38 


6.2.44 

Recent Advances in the Economics 

Supported by OR II 









1 - Sustainability screw: role of relative time scales 

Jerzy Filar, Mathematics and Statistics, University of South 

Australia, Mawson Lakes Blvd, 5095, Mawson Lakes, SA, 

Australia, j.filar@unisa.edu.au, Jacek Krawczyk, Manju Agarwal 

We postulate a time scales’ conjecture stating that most reasonable notions of 

sustainability must include a suitable synchronisation of time scales of both the 

processes of human development and those of the natural environment. We 

analyse a coarse, five variable, model of man nature interactions expressed as a 

system of differential equations where production and human capital are coupled 

with both renewable and non-renewable natural resources. A "sustainability 

screw’ phenomenon is demonstrated describing a spiral like trajectory of 

three key variables. 

2 - The Gambler’s Ruin Problem: Simulation 

Silvana Ligia Vincenzi Bortolotti, Matemática/Estatística, 

UTFPR/UFSC, Rua Jornalista Tito de Carvalho, 155, 88040480, 

Florianopolis, Santa Catarina, Brazil, sligie@globo.com, 

Fernando Moreira Junior, Afonso Farias Sousa Junior 

The gambler’s ruin is a classical problem of stochastic processes used to calculate 

the probability of a gambler who has a certain amount of financial resources, 

fall into ruin. In this paper, it were simulated and analyzed various 

situations and verified that it is worthwhile the company enter the market. We 

considered the three possible scenarios. The results also allow the company 

to know what proportion of the market that it should be to have a desirable 

probability of survival, as well as what should be the initial capital needed. 

83


� MD-39 


6.2.45 

Dynamic Programming Approach to 

Optimal Control Problems 



Chair: Fausto Gozzi, Dipartimento di Scienze Economiche e 

Aziendali, Luiss University - Roma - Italy, viale Romania 32, 00197, 

Roma, RM, Italy, fgozzi@luiss.it 

1 - Optimal mix in the interconnection of drinking water 

sources 

Alessandra Buratto, Department of Pure and Applied 

Mathematics, University of Padova, Via Trieste, 63, 35121, 

Padova, Italy, buratto@math.unipd.it, Chiara D’Alpaos 

We formulate and solve a stochastic optimal control model in order to determine 

the optimal feedback abstraction policy for a provider of water services 

who has invested in the interconnection of two different sources (e.g. groundwater 

vs river abstraction). The interconnection of water abstraction plants 

gives, de facto, the provider the option to strategically decide the optimal mix 

of different water sources to be used in supplying water to a community. Our 

aim is to show that this operational and technical flexibility is economically 

relevant if optimally exercised. 

2 - Optimal population problem: the role of finite life 

Giorgio Fabbri, Department of Economic Studies, University of 

Naples "Parthenope", Via Medina 40, 80133, Napoli, Italy, 

giorgio.fabbri@uniparthenope.it, Raouf Boucekkine, Fausto 

Gozzi 

Most of the literature on optimal population size does not mention the finite 

characteristic of human lifetime. The fact that our lives are finite changes the 

nature of the optimal population problem. We study the optimal dynamics of a 

finite life optimal population size model with procreation and education costs. 

We show how the optimal policy is influenced by the preferences parameters 

of the planner and the duration of the lifetime. The finite-lifetime makes gives 

a infinite dimensional structure to the problem. We study it using the dynamic 

programming in a suitable Hilbert space setting 

3 - On the sub-optimality cost of immediate annuitization 

in DC pension funds 

Elena Vigna, Dipartimento di Statistica e Matematica Applicata, 

Università di Torino and Collegio Carlo Alberto, corso Unione 

Sovietica 218 bis, 10135, Torino, Italy, 

elena.vigna@econ.unito.it, Marina Di Giacinto 

We investigate the optimal annuitization time for a retiree of a DC pension fund 

who takes programmed withdrawals. We exploit the model of GHV (2010), 

who formulate the problem as combined stochastic control and optimal stopping 

problem with a quadratic loss function. We prove a theorem that assesses 

whether immediate annuitization is optimal or not. With numerical simulations 

we investigate optimal annuitization time, ruin frequency, comparison between 

optimal and immediate annuitization. The cost of sub-optimality of immediate 

annuitization is measured with financial and actuarial criteria. 

� MD-40 


6.2.52 







1 - A Cutting Plane Algorithm for the RSAP 

84 

Paula Carroll, Management Information Systems, UCD, Quinn 

Business School, Belfield, 4, Dublin, paula.carroll@ucd.ie, 

Bernard Fortz, Martine Labbé, Seán McGarraghy 

We describe the Ring Spur Assignment Problem (RSAP) which arises in the 

design of hierarchical two layer ring/spur telecommunications networks. We 

describe this NP Hard problem and relate it to problems described in the literature. 

We present a complete Integer Programming (IP) formulation for the full 

two layer network topology. We present a cutting plane algorithm and describe 

the inequalities added and their separation. We present initial computational 

results obtained using XpressMP on benchmark problems. Finally, we present 

some conclusions and recommendations. 

2 - Optimal streaming network with relay nodes 

Eric Gourdin, CORE/TPN/TRM, Orange Labs, 38 rue du 

General Leclerc, 92794, Issy-les-Moulineaux, France, 

eric.gourdin@orange-ftgroup.com, Fabien Mathieu 

In a streaming network, a source node sends a multimedia file to several identified 

customers and the file is played in real-time. The capacity of such a 

network relies heavily on the upstream capacity of the server (the maximum 

bandwidth it can send into the network). If this capacity is limited, then the 

server can use some relay nodes to increase its streaming capacity. Indeed, a 

relay node can duplicate the received data before sending them back into the 

network. In this talk, we will propose several mixed integer models related to 

the design of an optimal streaming network. 

3 - Transmission Expansion Planning with Re-design 

Michael Poss, Computer Science Department, Université Libre 

de Bruxelles, Boulevard du Triomphe CP 210/01, 1050, Brussels, 

Belgium, mposs@ulb.ac.be, Claudia Sagastizabal 

Expanding an electrical transmission network requires heavy investments that 

need to be carefully planned, often at a regional or national level. We show that 

the problem is NP-hard and that, unlike the so-called Network Design Problem, 

a transmission network may become more efficient after cutting-off some of its 

circuits. For this reason, we introduce a new model allowing for the network 

to be re-designed when it is expanded. We then turn into different reformulations 

of the problem, that replace the bilinear constraints by using a "big-M’ 

approach. Computational results are presented. 

4 - A New Demand Uncertainty Set Definition and The Robust 

Network Loading Problem 

Aysegul Altin Kayhan, Industrial Engineering, TOBB University 

of Economics and Technology, TOBB Ekonomi ve Teknoloji 

Universitesi Endustri Muhendisligi Bolumu, Sogutozu Caddesi 

No=43 Sogutozu, 06560, Ankara, Turkey, aaltin@etu.edu.tr, 

Hande Yaman, Mustafa Pinar 

We consider the Robust Network Loading Problem, which includes the private 

network design problem as a special case. We study the problem from 

service provider’s perspective, who wants to lease private lines in the most 

cost-efficient manner. Rather than assuming that the communication demands 

are known, we introduce a genuine polyhedral demand model, which inherits 

the strengths of two known polyhedral demand definitions. We develop two 

compact mixed integer programming formulations of the problem and provide 

some experimental results for several well-known instances to discuss the economic 

and performance implications of incorporating the robustness dimension 

in our design efforts. 

� MD-41 


3.1.06 

Advances in Revenue Management 



Chair: Ayse Kocabiyikoglu, Department of Business Administration, 

Bilkent University, Bilkent, 06800, Ankara, Turkey, 

aysekoca@bilkent.edu.tr 

1 - Decision Making in Revenue Management: An Experimental 

Study 

Mert Hakan Hekimo˘glu, Bilkent University, 06800, Ankara, 

mhekim@bilkent.edu.tr, Ayse Kocabiyikoglu, Itir Gogus

In the standard two-class revenue management model, the firm allocates a fixed 

capacity between two market segments with hierarchical willingness to pays. 

If the firm allocates too much to the higher priced segment, it is left with unused 

capacity at the end of the selling period; if it allocates too little, demand 

from the more valuable segment is unmet. The revenue maximizing allocation 

is given by Littlewood’s Rule, but little is known about how managers actually 

make these decisions. We describe results from a study that investigates 

revenue management decisions across different conditions. We also investigate 

the impact of varying problem parameters on the decision maker’s allocation 

choice. 

2 - Performance-Based Contracts for Outpatient Medical 

Services 

Houyuan Jiang, Judge Business School, University of 

Cambridge, Trumpington Street, CB2 1AG, Cambridge, United 

Kingdom, h.jiang@jbs.cam.ac.uk, James Z. Pang, Sergei Savin 

Under the principal-agent framework, we analyze Payment by Results (PbR) 

and Performance-based Contract (PBC) approaches for the study of contract 

design in the context of outpatient medical services. We gain important insights 

by comparing PbR and PBC under different information structures: complete 

information, asymmetric information, and private actions. We show that when 

the agent’s capacity allocation decisions are observable and contractible, PbR 

and PBC approaches produce the same outcomes. However, if agent’s decisions 

are not observable and contractible, PBC outperforms PbR. 

3 - Dynamic Pricing of High Speed Rail with Transport 

Competition, Substitutable Schedule and Overbooking 

Kimitoshi Sato, Graduate School of Business Administration, 

Nanzan University, ksato@nanzan-u.ac.jp, 466-8673, Nagoya, 

Japan, ksato@nanzan-u.ac.jp, Katsushige Sawaki 

In this paper, we present a revenue management model of dynamic pricing of 

HSR(high speed rails)’s fares in which the passengers are allowed to choose 

among other transport modes by taking account of Air-HSR competition. Each 

transport mode offers the multiple substitutable schedules. We use the NMNL 

model to describe the customer’s discrete choice. The passengers choose an alternative 

transport mode based on the comfort, total trip time, the total price and 

the frequency. Furthermore, we allow cancellation, no-shows and overbooking, 

and show the existence of Nash equilibrium. 

� MD-42 


3.1.07 

Theory of Bilevel Programming 




Chair: Joydeep Dutta, Math and Stat, Indian Institute of Technology, 

Office : Room No 575, Faculty Building, Academic Area, 208016, 

Kanpur, Uttar Pradesh, India, jdutta@iitk.ac.in 

1 - Optimality conditions for bilevel programming problems 

Stephan Dempe, Mathematics and Computer Sciences, 

Technische Universitaet Freiberg, 09596, Freiberg, Germany, 

dempe@math.tu-freiberg.de, Alain B. Zemkoho 

In the talk, an overview of different transformations of the bilevel programming 

problem into a one-level programming problem and related necessary 

conditions for local optimal solutions will be given. Special emphasize is on 

the question which constraint qualifications are suitable in relation to which 

transformations. 

2 - Bilevel programming in Hilbert space 

Ayalew Getachew Mersha, Optimization and Optimal control 

group, Austrian Academy of Sciences, Johann Wilhelm Klein 

Strasse 9, 4040 , Linz, Upper Austria, Austria, 

ayalew.mersha@oeaw.ac.at 

In this talk the bilevel programming problems in a Hilbert space is considered. 

After reformulations of classes of these problems we use the idea of optimal 

control theory. We give conditions for the existence and uniqueness of the 

solution. A numerical algorithm based on the idea of semi-smooth Newton 

method is proposed. We also discuss the associated difficulty in solving such 

problems in functional spaces. 


3 - Optimality conditions for a simple bilevel programmimg 

problem 

Joydeep Dutta, Math and Stat, Indian Institute of Technology, 

Office : Room No 575, Faculty Building, Academic Area, 

208016, Kanpur, Uttar Pradesh, India, jdutta@iitk.ac.in 

In this talk we consider a simple bilevel programming problem of a single decsion 

variable. The upper level problem consists of a convex function which 

is to be minimized over a convex set which in turn is the solution of another 

convex optimization problem which we call the lower-level problem. We study 

the optimality conditions for this class of problems. 

� MD-43 


8.2.02 


Stream: Algorithmic Decision Theory [c] 


Chair: Sasa Pekec, Fuqua School of Business, Duke University, 1 



1 - Management and development of profiles by learning 

for Multicriteria Decision Support 

Arnaud Martin, Institut de Recherche en Informatique de 

Toulouse, Toulouse University, 118 Route de Narbonne, IRIT, 

Université Paul Sabatier, F-31062 , Toulouse CEDEX 9, France, 

amartin@irit.fr, Pascale Zaraté, Guy Camilleri 

As part of the development of Coopeartive DSS, users’ profiles, which are defined 

by several criteria whose quantity which evolves dynamically, are modeled 

and implemented. The purpose of this work consists in bringing the best 

support to activity of the user according to his profile. These users’ profiles will 

evolve thanks to existent automated techniques in learning domain, especially 

reinforcement, in order to get the profile of a user according to his answers 

and actions. The support brought to the user by the system will also evolve 

according to the evolution of his profile. 

2 - Aggregation operations in multi-criteria simple games 

Luisa Monroy, Economia Aplicada III, Universidad de Sevilla, 

Avda. Ramon y Cajal, 1, 41018, Sevilla, lmonroy@us.es, 

Francisco Ramon Fernandez 

Multi-criteria simple games provide an appropriate general framework within 

which to analyze group decision problems, especially voting systems and collective 

choice procedures where more than one alternative can be simultaneously 

chosen and where numeric utility values are either inappropriate or impossible 

to derive. We study extended multi-criteria simple games obtained 

from different aggregation operations, such as union, intersection, marginalization, 

and composition, all defined for multi-criteria simple games. Real voting 

systems are modelled as extended multi-criteria simple games. We show that a 

voting system can be established when these operations are applied to complex 

voting systems. 

3 - Preconditions for Information Aggregation in Prediction 

Markets 

Sasa Pekec, Fuqua School of Business, Duke University, 1 



We study necessary and sufficient conditions for information to aggregate in 

a highly stylized one-shot unit-demand uniform-price prediction market. We 

show that the market price properly aggregates information if and only if (i) 

the number of realized trades is non-negligible compared to number of market 

participants and (ii) proportion of buyer offers is asymptotically matching the 

proportion of seller offers. These conditions prevail in more general prediction 

market models that include the model studied here as a special case. 

4 - Bidding Strategies for Real-Life Small Loan Auctions 

Hannele Wallenius, Industrial Engineering and Management, 

Helsinki University of Technology, P.O. Box 9500, 02015 HUT, 

85


02015 HUT, Espoo, Finland, hannele.wallenius@tkk.fi, Lauri 

Puro, Jyrki Wallenius, Jeffrey Teich 

We define and identify bidding strategies in real-life small loan auctions (Prosper.com). 

In this auction, lenders bid for borrowers’ loan listings and the winners 

get to fund the loan at an interest rate determined by the auction. The exceptionally 

large empirical database provided by Prosper.com offers a unique 

opportunity to test and further develop the theory of online auctions. This study 

shows that bidding behavior is not homogeneous among bidders, as the traditional 

auction theory suggests. Instead, bidders use many different bidding 

strategies. Moreover, learning and bidders’ consistency over time in different 

auctions is studied. 

� MD-44 


8.2.03 

Modelling complex systems 

Stream: SD Modeling in Sustainable Development 


Chair: Pierre Kunsch, MOSI, Vrije Universiteit Brussel, Pleinlaan 2, 

1050, Brussels, Belgium, pkunsch@vub.ac.be 

1 - A systemic dynamic approach of the financial crisis 

Pierre Kunsch, MOSI, Vrije Universiteit Brussel, Pleinlaan 2, 

1050, Brussels, Belgium, pkunsch@vub.ac.be 

This paper presents a system-dynamics modelling of the world crisis, its origin, 

consequences, and macro-economic remedies proposed by governments. The 

specific case of Belgium is analysed in more details 

2 - Use of Management Science Systemic Methodologies 

in Environmental Management and Sustainability 

Alberto Paucar-Caceres, Business School, Manchester 

Metropolitan University, Aytoun Building, Aytoun Street, M1 

3GH, Manchester, United Kingdom, a.paucar@mmu.ac.uk 

This paper investigates and discusses the use of systemic methodologies (SM) 

developed in management science/operational research (MS/OR), in particular, 

those SM that have been informing the complexity inherent in environmental 

management and sustainable (EM/S) practices. By surveying a sample of 

the top MS/OR and systems journals, we assess the extent to which systemic 

management science methodologies developed recently have been used in tackling 

EM/S problems. By critically reviewing applications in EM/S the paper 

hopes to raise awareness amongst environmentalists, operational researchers 

and management scientists of the benefits of using systemic approaches developed 

in MS/OR. 

3 - Closed Loop Supply Chains: A Systems Dynamics 

Model for Analyzing Sustainable Business Policies for 

Shared Partners in a Chain 

Jose Cruz, Operations and Information Management, University 

of Connecticut, School of Business, 2100 Hillside Road, 

06269-1041, Storrs, CT, United States, 

jcruz@business.uconn.edu 

We propose and implement a system dynamics approach to study the problem 

space and generate policy insights for managing the overall closed loop chain 

partners. Our work is unique in addressing a holistic approach that includes 

initial pricing and sales of products, designing products for manufacturer and 

specified recycling rates, and looking at a holistic revenue stream of initial retailers, 

recyclers and remanufacturers. We address the following questions: Is 

the closed loop economically sustainable? What is the impact of product design 

in closing the loop? 

4 - Collective animal manure management simulation and 

environmental impact 

Francois Guerrin, CA, Inra/Cirad, Station de la Bretagne - BP 

20, 97408, Saint-Denis, France, francois.guerrin@cirad.fr, 

Jean-Marie Paillat 

This contribution uses Systems Dynamics for simulating pig slurry spreading 

plans in Brittany (France), and their harmful environmental impacts on groundwater 

and coastal waters. Individual pig farmers use part of their slurry to fertilise 

their own crops; they export their excess slurry to remote crop farms. The 

model dynamically simulates the slurry stock evolutions at the pig farms and 

the spreading fluxes on crops, both at the pig farms and the remote crop farms. 

It also simulates several indicators for assessing the environmental impact of 

this spreading. 

86 

� MD-45 


8.2.12 

Novel opportunities of DC programming 

and DCA for Industry and Finance 

Stream: Nonconvex Programming: Local and Global 

Approaches 


Chair: Hoai An Le Thi, Computer Science, University Paul Verlaine - 

Metz, Ile du Saulcy„ 57 045 , Metz, France, lethi@univ-metz.fr 

1 - Solving the Min m-Dominating Set Problem by a Continuous 

Optimization Approach based on DC Programming 

and DCA 

Julien Schleich, Computer Science, University of Luxembourg, 

5, rempart Saint-Thiébault, 57000, METZ, France, 

julien.schleich@uni.lu, Hoai An Le Thi, Pascal Bouvry 

We propose a new optimisation approach based on DC (Difference of Convex 

functions) programming and DCA (DC Algorithm) to the graph problem minimum 

m-Dominating Set. This problem is beforehand recast as a polyhedral DC 

program with the help of exact penalty in DC programming. The related DCA 

is original and computer efficient because it consists of solving a few linear 

programs and converges after a finite number of iterations to an integer solution 

while working in a continuous domain. Numerical simulations show the 

efficiency and robustness of DCA and its superiority with respect to standard 

methods. 

2 - A DC programming approach for constrained twodimensional 

non-guillotine cutting 

Mahdi Moeini, Computer Science, University Paul Verlaine - 

Metz, 57045, METZ, France, moeini@univ-metz.fr, Hoai An Le 

Thi, Tao Pham Dinh 

We investigate a new application of DC (Difference of Convex functions) 

programming and DCA (DC Algorithm) in solving the constrained twodimensional 

non-guillotine cutting problem. This problem consists of cutting 

a number of rectangular pieces from a large rectangular object. The cuts are 

done under some constraints and the objective is to maximize the total value of 

the pieces cut. We reformulate this problem as a DC program and solve it by 

DCA. The performance of the approach is compared with the standard solver 

CPLEX. 

3 - Power Control in Cellular Networks via DC Programming 

and DC Algorithms (DCA) 

Anh Son Ta, LMI, INSA de Rouen, LMI, INSA de Rouen, +33, 

Rouen, France, taanh_son@yahoo.com, Hoai An Le Thi, Tao 

Pham Dinh 

Power control is typically used in wireless cellular networks in order to optimize 

the transmission subject to quality of service (QoS) constraints. It has 

been shown early that this problem can be efficiently solved by using the socalled 

geometric programming. In this paper, we investigate a new and efficient 

algorithm based on DC (Difference of Convex functions) programming 

and DCA (DC Algorithms) for solving it. Preliminary numerical simulations 

demonstrate the efficiency of DCA and its superiority compared to the geometric 

programming algorithm. 

4 - Scheduling of lifting vehicle and Quay Crane in automated 

port container terminals 

Le Hoai Minh, LMAH, Université Le Havre, France, France, 

mlehoai@yahoo.fr 

Container terminals, are continuously facing the challenge of strong competition 

between ports. In this study, we present the terminal of Normandy; Le 

Havre port. We consider a mixed integer programming problem for the problem 

of assigning optimal delivery tasks to Lifting Vehicles. There are a lot of 

algorithms designed for this problem such as B&B method, cutting plane,... By 

using an exact penalty technique we treat this problem as a DC program in the 

context of continuous optimization. Further, we combine the DCA with the 

classical Branch and Bound method for finding global solutions.

� MD-46 


8.2.14 

Semi-Infinite Optimization I 

Stream: Semi-Infinite Optimization 


Chair: Vladimir Shikhman, Dept. Mathematics, RWTH Aachen 


shikhman@mathc.rwth-aachen.de 

1 - Generalized Semi-Infinite Programming: the Nonsmooth 

Symmetric Reduction Ansatz 

Vladimir Shikhman, Dept. Mathematics, RWTH Aachen 


shikhman@mathc.rwth-aachen.de, Hubertus Th. Jongen 

We introduce the Nonsmooth Symmetric Reduction Ansatz (NSRA) on the 

closure of the feasible set cl(M) in GSIP. cl(M) is given by infinitely many 

constraints of maximum-type. Under NSRA it is the feasible set in Disjunctive 

Optimization given by finitely many such constraints. The new issue is: the Lagrange 

polytope at the lower level is not a singleton. The "full-dimensionality" 

of its vertices is given. We introduce nondegenerate KKT points and GSIPindex. 

NSRA is generic and stable at KKT points. All KKT points are generically 

nondegenerate. We discuss the critical point theory for GSIP. 

2 - Robustness modulation and chance-constrained optimization 

with application to control: Making it easy 

through randomization 

Simone Garatti, Dept. of Electronics and Information, 

Politecnico di Milano, p.zza L. da Vinci 32, 20133, Milan, Italy, 

sgaratti@elet.polimi.it, Marco Campi 

Robust control can lead to overconservative designs because all emphasis is 

placed on safe-guarding the closed-loop against all possible negative occurrences. 

In many applications, 100% robustness is not necessary and accepting 

a small compromise in robustness guarantees along a chance-constrained optimization 

approach can often lead to a huge improvement in performance. Yet, 

the real stumbling-block is the lack of computationally tractable algorithms 

able to trade robustness guarantees for performance. This talk aims at opening 

new directions to solve this problem through randomization. 

3 - Multiobjective programming problems with fuzzy random 

coefficients 

Monga K Luhandjula, Decision Sciences, University of South 

Africa, Muckleneuk Campus, Pretoria, 0003, Pretoria, Gauteng, 

South Africa, luhanmk@unisa.ac.za 

Neither Stochastic Optimization nor Fuzzy Mathematical Programming is adequate 

for capturing hybrid situations where fuzziness and randomness co-occur 

in a multiobjective programming setting. In this paper, we propose an approach 

for finding a satisfying solution of a multiobjective program with fuzzy random 

coefficients. The key idea behind our approach is to explore , with good 

reasons, correspondences between fuzzy random variables and random closed 

sets. A numerical example is carried out for the sake of illustration. It shows 

the efficiency of the proposed method. 

� MD-47 


8.2.16 

OR in Oil Sector I 

Stream: OR in Oil Sector 


Chair: Irina Dolgopolova, Economics and Administrative Sciences, 

Middle East Technical University, Odtu Kent, Konuk Evi 1, B Block, 

108, Ankara, Turkey, 06420, Ankara, Turkey, 


1 - Network Design of Petroleum Supply Chains 

Leão Fernandes, Centro de Estudos de Gestão, Instituto Superior 

Técnico, Av.Rovisco Pais, 1049-001, Lisboa, Lisboa, Portugal, 


leao.fernandes@clc.pt, Ana Paula Barbósa-Póvoa, Susana 

Relvas 

Petroleum supply chain (PSC) is a strategic sector of modern economy, known 

for huge network of investments and complex infrastructures. While the former 

confers strategic business risk, the later reveals the need of optimization tools 

for decision making at design, planning and operational levels. This investigation 

focus the design level and presents a mixed integer programming model 

to optimize PSC structure while considering its characteristics and the business 

risk associated. Test results based on factual logistic data for the Portuguese 

PSC network and research directions are presented 

2 - A simulated annealing approach for scheduling 

workover rigs on onshore oil production 

Geraldo Mauri, Rural Engineering, Federal University of 

Espírito Santo - UFES, Alto Universitário s/n, 29500-000, 

Alegre, Espírito Santo, Brazil, mauri@cca.ufes.br, Glaydston 

Ribeiro, Luiz A. N. Lorena 

Onshore oil wells are dependent of maintenance services such as cleaning 

and reinstatement. Wells need maintenance services and a scheduling of the 

workover rigs (WR) must be defined. The WR scheduling problem consists of 

finding the best schedule for the limited number of WR, minimizing the production 

loss associated with the wells waiting for maintenance service. We present 

a Simulated Annealing (SA) algorithm for solving this problem. Computational 

results on real problems obtained in Brazil are reported and SA presents 

better solutions than all those approaches reported in the literature. 

3 - Multiproduct Pipeline Scheduling Systems 

Susana Relvas, DEG, IST, Av. Rovisco Pais, 1049-001, Lisbon, 

Portugal, susanaicr@ist.utl.pt, Ana Paula Barbósa-Póvoa, 

Henrique Matos 

Petroleum supply chain comprehends complex logistics operations, to ensure 

supply and demand at several nodes. Transportation is critical since large quantities 

of products are required with high reliability. Pipelines constitute an adequate 

transportation mode, but represent a complex problem, due to products’ 

competition for a unique resource. To adequately schedule the pipeline 

pumping sequence, proper inventory management policies must be met at the 

destination tank farm. This problem is addressed in a Mixed-Integer Linear 

Programming Model and tested for a real Portuguese Company setting. 

4 - Novel continuous nonlinear optimization approach for 

refinery scheduling 

Joao Lauro D. Faco’, Dept. of Computer Science, Universidade 

Federal do Rio de Janeiro, Av. do PEPE, 1100 / 203, 22620-171, 

Rio de Janeiro, RJ, Brazil, jldfaco@ufrj.br, Fabio Fagundez, 

Adilson Elias Xavier 

We present a novel continuous nonlinear optimization approach for crude oil 

operations in a refinery, from tankers to crude distillation units. The schedule is 

modeled as a sequence of refinery states, where the contents of the equipment 

(tanks, pipelines, and tankers) are mapped to state variables and crude oil flows 

from an equipment to another are mapped to control variables, which change 

the refinery states. Schedule discrete decisions are modeled with complementarity 

constraints. Recent examples from the literature are solved in reasonable 

computing times (seconds). 

� MD-48 


8.2.04 

Multi-objective optimization 

Stream: Optimization for Sustainable Development 


Chair: Nora Touati Moungla, LIX, Ecole Polytechnique, École 

polytechnique„ Laboratoire d’informatique (LIX), 91128, Palaiseau, 

Cedex, France, France, touati@lix.polytechnique.fr 

1 - Computation of best-compromise route for cycling 

Gaël Sauvanet, Laboratoire d’Informatique de l’Université de 

Tours, Polytech’Tours, 64 avenue Jean Portalis, 37200, TOURS, 

France, gael.sauvanet@univ-tours.fr, Emmanuel Néron, Hervé 

Baptiste 

87

ME-01 EURO 24 - Lisbon 2010 

Today, for environmental reasons, the use of alternative transportation such as 

cycling is expanding. Communities have initiated several programs to adapt the 

network to cycling because users are seeking to be able to move along safely. 

Indeed, for a cyclist, traveling time and distance are not the only criteria to consider: 

security, difficulty and touristic attraction may be taken into account. We 

propose methods for designing touristic, safe and short route for cycling. We 

propose methods both for determining Pareto set, as well as best compromise 

solution using user’s preferences. 

2 - New insights of preference modeling in multiobjective 

optimization 

Souhila Kaci, CRIL, 16 rue de l’Université, SP 16, 62307, Lens, 

France, kaci@cril.fr, Nora Touati Moungla 

Multi-Objective Optimization (MOO) problems consist of selecting solutions 

w.r.t. multiple objectives. It generally returns Pareto-optimal solutions, i.e., 

there is no solution that is better to these solutions w.r.t. all objectives. Paretooptimality 

based optimization assumes that all objectives have equal importance, 

however a user may express preferences over objectives. Therefore preferences 

act as a filter and select the "preferred" solutions among Pareto-optimal 

ones. We investigate different ways to incorporate and reason about preferences 

in MOO problems. 

88 

Monday, 15:40-17:00 

� ME-01 


Aula Magna 

Plenary Talk 1 



Chair: Dominique de Werra, IMA, EPFL, FSB, EPFL, CH 1015, 

Lausanne, Switzerland, dominique.dewerra@epfl.ch 

1 - The Agencies Method for Modeling Coalitions and Cooperation 

in Games 

John F. Nash, Jr., Department of Mathematics, Princeton 

University, Fine Hall, Washington Road, NJ 08544-1000, 

Princeton, United States, kjfnj@math.princeton.edu 

Our work in this research project represents the beginning of an effort to study 

the game-theoretic phenomenon of cooperation in cooperative games (that is, 

in games where it is understood that the players MAY cooperate whenever 

they would naturally desire to do that so as to realize mutually advantageous 

benefits). (This project of work could be described as work along the lines 

of the (old) "Nash program", seeking to reduce the study of "cooperative 

games" to the area of the (theoretically simpler) study of "equilibrium" in "noncooperative 

games".) The theoretical key for the inter-linking of these areas is 

the concept of the "evolution of cooperation" (in Nature) which has been studied 

both by theoretical biologists and by game theorists. Our key idea for the 

reduction of a process realized through cooperation to a process achieved by 

actions taken independently (and separately) by the Players that are involved in 

the game context lies in the introduction of moves (or actions) of "acceptance" 

together with the supposition of an indefinitely repeated game context in which 

the Players may react (in punishing fashions) against unfavorable actions on 

the part of other Players. Our model designed for this purpose has the players 

strategically making "demands" in relation to the behavior of the other players. 

And at the same time any Player also chooses, strategically, how he (or she) 

will allocate the resources available to a coalition if he (or she) has become 

accepted (through acceptance elections) to have that power. As it happens, in 

short, the studied model, for a game with three players, involved 39 strategic 

parameters to be controlled by the players, and we represented it in terms of 

42 variables. There was a substantial challenge of computation to find game 

theoretic solutions. These were sought in terms of PURE STRATEGIES. And 

the game example itself, although being described by a characteristic function 

giving rise to three numerical parameters setting the payoff benefits accessible 

by specific coalitions, was an "NTU game" rather than a game with "transferable 

utility". However we have compared solution results with corresponding 

indications for the games deriving from the Shapley value or from the nucleolus. 

The general context of computations is found to be quite challenging. We 

have used Mathematica in the connection with the work done up to now. There 

seem to be possibilities for refinements in the model structure and games of 

more than three players can also be studied.


� MF-02 


3.2.14 

Flexible shop scheduling by Metaheuristics 

and solutions for real problems 



Chair: Lorena Pradenas, Industrial, Universidad de Concepción, 

Barrio Universitario, Concepción, casilla 160 C, correo 3, 

Concepción, Concepción, Concepción, Chile, lpradena@udec.cl 

1 - A genetic algorithm for the flexible job shop problem 

Lorena Pradenas, Industrial, Universidad de Concepción, Barrio 

Universitario, Concepción, casilla 160 C, correo 3, Concepción, 

Concepción, Concepción, Chile, lpradena@udec.cl, Rosa 

Medina, Víctor Parada 

The Flexible Job Shop Problem is part of the family of scheduling problems. 

It extends the job shop problem in order to optimize the use of resources in 

a flexible production system, that is, those with machines that can process 

more than one type of operation. This problem has been studied by many 

authors, who have proposed mathematical models and heuristics approaches. 

Due to its’ combinatorial complexity, the exact methods that solve the mathematical 

models only solve small instances. Among the heuristics approaches, 

the metaheuristics of local search have demonstrated a better performance. In 

this study, a sequential genetic algorithm is presented to solve The Flexible 

Job Shop Problem. The proposed algorithm is tested using literature instances. 

The results show that the algorithm is effective for finding good solutions for 

the problem. 

2 - The inverse power index problem 

Sascha Kurz, Mathematics, Physics and Informatics, University 

of Bayreuth, Universitätsstraße 30, 95440, Bayreuth, Bavaria, 

Germany, sascha.kurz@uni-bayreuth.de 

Weighted voting games are frequently used in decision making. Each player 

has a weight and each proposal is accepted if the weight sum of the supporting 

players exceeds a quota. One line of research is the efficient computation 

of so-called power indices measuring the influence of a player. We treat the 

inverse problem: Given an influence vector and a power index, determine a 

weighted voting game such that the distribution of influence among the players 

is a close as possible to the given target value. We present exact algorithms and 

computational results for the Shapley-Shubik index. 

3 - Signal decoding in multi-antenna systems using 

second-order cone programming 

Edmund Burke, School of Computer Science & IT, University of 


Nottingham, United Kingdom, ekb@cs.nott.ac.uk, Jakub 

Marecek, Andrew J. Parkes 

Receivers in modern wireless communications (WiFi N, WiMAX, LTE, LTE 

Advanced) all implement a solver for integer least squares, a problem in integer 

non-linear programming. We first present an overview of solvers currently 

implemented, and subsequently compare them to a solver of our own, based 

on second-order cone programming cuts. Extensive computational results are 

presented. 

4 - A binary programming model to the Dynamic Search 

Problem 

Carlos Diego Rodrigues, Informatique, Université d’Avignon, 

334, Chemin des Meinajaries, Agroparc, 84411, Avignon, 

PACA, France, cdiegor@gmail.com, Boris Detienne, Dominique 

Quadri, Philippe Michelon 

Search problems are among the first problems studied by Operations Research 

in several fields (game theory, graph theory, stochastic programming, etc). 

They are associated to many practical applications, notably in security. We 

present the first binary programming model to this problem and how this model 

can be adapted to most of the cases already appearing in the literature. A statistical 

validation process, that can evaluate any given solution plan, is also 

shown and used to corroborate with our model, allowing us to establish some 

numerical results concerning its performance. 

EURO 24 - Lisbon 2010 MF-03 

� MF-03 


3.2.15 

Population-based metaheuristics for 

routing problems 





Chair: Philippe Lacomme, Université de Clermont-Ferrand, 63177, 

Clermont Ferrand, France, placomme@sp.isima.fr 

1 - Memetic algorithm with diversity enhancement: an efficient 

approach for hard multi-attribute vehicle routing 

problems 

Thibaut Vidal, Informatique et recherche opérationnelle, 

Université de Montréal / Université de Technologie de Troyes, 

5536 chemin de la Cote des Neiges, apt 4, H3T1Y9, 

MONTREAL, QC, Canada, thibaut.vidal@cirrelt.ca, Teodor 

Gabriel Crainic, Michel Gendreau, Nadia Lahrichi, Walter Rei 

We introduce a new memetic algorithm for a hard class of multi-depot periodic 

vehicle routing problems with time-windows, where very few efficient 

algorithms could be found. New genetic operators for periodic problems are 

proposed. Furthermore, in contrast with other population management methods 

operating solely during survivor selection, our algorithm favors diversity 

through the very evaluation of individuals, which is driven by fitness as well 

as contribution to the population diversity. Comparative studies underline the 

efficiency of our "diversity enhancement" method on these problems. 

2 - A new Ant Colony Optimization on Vehicle Routing 

Problem with heterogeneous fleet, mixed backhauls, 

and time windows 

Farah Belmecheri, LOSI, University of Technology of Troyes, 

12, Rue marie curie, 10010, Troyes, France, France, 

farah.belmecheri@utt.fr, Christian Prins, Farouk Yalaoui, Lionel 

Amodeo 

This paper presents a new Ant Colony Optimization to solve the Vehicle Routing 

Problem (VRP) with: Heterogeneous fleet (H), Mixed Backhauls (MB), 

Time Windows (TW) which is called HVRPMBTW. This metaheuristic consists 

to construct the routes using the probabilities of insertion of customers; 

the local searches are added to improve the solutions. In computational results, 

this method is applied on the sets of instances of HVRPMBTW and the classical 

HVRP with limited and unlimited vehicles (instances of Taillard 1996). 

The results confirm the efficiency of this new ACO. 

3 - An Artificial Bee Colony Algorithm for the Capacitated 

Vehicle Routing Problem 

Sin C. Ho, Department of Business Studies, Aarhus School of 

Business, Aarhus University, Fuglesangs Allé 4, 8210, Aarhus V, 

Denmark, sinch@asb.dk, Wai Yuen Szeto, Yongzhong Wu 

This paper introduces an artificial bee colony heuristic for the capacitated vehicle 

routing problem. The artificial bee colony heuristic is a swarm-based 

heuristic, which mimics the foraging behavior of a honey bee swarm. The performance 

of the heuristic is evaluated on two sets of benchmark instances. A 

new scheme is also developed to improve the performance of the artificial bee 

colony heuristic. Computational results show that the heuristic with the new 

scheme produces good solutions. 

4 - A Hybrid Genetic Algorithm (HGA) for the Multi-Depot 

Pickup and Delivery Problem (MDPDP) 

Pairoj Chaichiratikul, Imperial College Business School, 

Imperial College London, 18, Elvaston Place, SW7 5QF, 

London, United Kingdom, 

pairoj.chaichiratikul06@imperial.ac.uk, Eleni Hadjiconstantinou 

The problem of serving a number of pickup and delivery locations using a 

heterogeneous fleet of vehicles located at several depots is formulated as a 

mixed-integer linear programming problem. The objective is to find minimumdistance 

routes subject to precedence, capacity and maximum-route length constraints. 

This is an NP-hard problem and we use ILOG CPLEX for optimally 

solving instances of small size only. A new meta-heuristic approach (HGA) 

is proposed, implemented and computationally tested on various test instances 

from the literature. Competitive near-optimal solutions are reported. 

89

MF-04 EURO 24 - Lisbon 2010 

� MF-04 


3.2.13 

Shop scheduling with metaheuristics 






Chair: Nicolau Santos, INESC Porto and University of Porto, 

Portugal, nicolausantos@gmail.com 

1 - A solution approach for flexible job shop scheduling 

problem using artificial immune system 

Aydin Sipahioglu, Industrial Engineering, Osmangazi University, 

Meselik, 26480, Eskisehir, Turkey, asipahi@ogu.edu.tr, Alper 

Aladag 

Flexible Job Shop Scheduling Problem (FJSSP) is one of the hardest NP-hard 

class problems in combinatorial optimization. Therefore, lots of heuristic and 

meta-heuristic methods are used to solve this problem. In this study, a new 

hierarchical Artificial Immune System (AIS) solution approach inspired by biological 

immune system has been developed for FJSSP. Empirical results on 10 

benchmark problems known as BRdata in the literature indicate that the proposed 

approach can obtain better solution results than former approaches and 

it is very efficient to solve FJSSP. 

2 - Master-slave multicore metaheuristic for the permutation 

flowshop problem 

Eva Vallada, Estadística e Investigación Operativa Aplicadas y 

Calidad, Universidad Politécnica de Valencia, Cno. Vera s/n 

Edificio I-3, 46022, Valencia, Spain, evallada@eio.upv.es, Gema 

Escrivá 

In this work a cooperative metaheuristic for the permutation flowshop problem 

with the objective to minimize the total tardiness is proposed. The method 

is based on the master-slave model and uses recent technologies, where more 

than one core is available in the processor of the computer. Each core runs the 

algorithm and communications are allowed between the master and the slaves. 

A comparative evaluation against the serial counterparts and other multicore 

algorithms is carried out, using different sending strategies. Results show that 

the multicore method outperforms the remaining ones. 

3 - Permutation flowshop scheduling with makespan and 

tardiness objectives by GRASP 

Iryna Yevseyeva, UESP, INESC Porto, Portugal, Campus da 

FEUP, Rua Dr. Roberto Frias, 378, 4200-465, Porto, -, Portugal, 

irynayev@yahoo.com, Jorge Pinho de Sousa, Luis Guardao 

In this work, permutation flowshop scheduling problem is considered taking 

into account two conflicting objectives: makespan and tardiness. For this 

scheduling problem, a multi-start metaheuristic, GRASP, is extended for optimizing 

several objectives. As a result of applying construction and local 

search phases of GRASP, a Pareto front of efficient solutions is obtained at 

each GRASP iteration. An external archive of elite solutions is kept. For intensification 

and diversification goals path-relinking between elite solutions is 

performed. The approach is tested on benchmark problems. 

4 - Metaheuristics For Minimizing Total Tardiness in the 

Permutation Flowshop With Release Dates 

Nicolau Santos, INESC Porto and University of Porto, Portugal, 

nicolausantos@gmail.com, João Pedro Pedroso 

The permutation flowshop scheduling problem is a widely studied combinatorial 

problem with many aplications in real life problems. In this work we focus 

on minimizing the total tardiness, given a release and a due date for each job. 

We propose metaheuristics based on local search with the insertion neighborhood, 

and provide computational results for assessing their quality. We compare 

our methods to several well-known heuristics adapted for this problem. 

90 

� MF-05 


3.2.16 

EURO Doctoral Dissertation Award 

Stream: EURO Doctoral Dissertation Award 


Chair: Mikael Rönnqvist, Department of Finance and Management 



mikael.ronnqvist@nhh.no 

1 - Operating room planning and scheduling: solving a 

surgical case sequencing problem 

Brecht Cardoen, Vlerick Leuven Gent Management School & 

Faculty of Business and Economics, Katholieke Universiteit 

Leuven, Reep 1, B-9000, Gent, Belgium, 

brecht.cardoen@vlerick.be 

We present the main results of the author’s PhD dissertation, which was defended 

at the Katholieke Universiteit Leuven and supervised by E. Demeulemeester. 

The thesis studies the impact of planning and scheduling procedures 

on a hospital’s operating room performance. It incorporates an extensive review 

of both scientific contributions and the current practice of hospitals in 

Flanders (Belgium). The emphasis of the research, though, is directed towards 

the development, the testing and the application of exact and heuristic algorithms, 

such as dedicated branch-and-bound procedures or mixed integer linear 

programming approaches, for surgery sequencing in a day-care environment. 

2 - Application-oriented Mixed Integer Non-Linear Programming 

Claudia D’Ambrosio, DEIS, Universita‘ di Bologna, 40136, 

Bologna, Italy, c.dambrosio@unibo.it 

The main topic of the thesis is Mixed Integer Non-Linear Programming, with 

focus on non-convex problems (i.e., problems for which the feasible region of 

the continuous relaxation is a non-convex set) and real world applications. Different 

kinds of algorithms are presented: linearization methods, heuristic and 

global optimization algorithms. Also, different kinds of real-world applications 

are solved, arising, for example, from Hydraulic and Electrical Engineering 

problems. The last part of the thesis is devoted to software and tools for mixed 

integer non-linear programming problems. 

3 - A mathematical contribution of statistical learning and 

continuous optimization using infinite and semi-infinite 

programming to computational statistics 

Sureyya Ozogur-Akyuz, Department of Mathematics and 

Computer Science, Bahcesehir University, Bahcesehir 

University, Dept of Mathematics and Computer Science, Cıragan 

cad. Besiktas, 34353, Istanbul, Turkey, 

sureyya.akyuz@bahcesehir.edu.tr 

In Machine Learning algorithms, one of the crucial issues is the representation 

of the data. As the data become heterogeneous and large-scale, single kernel 

methods become insufficient to classify nonlinear data. Convex combinations 

of kernels were developed to classify this kind of data. Nevertheless, selection 

of the finite combinations of kernels is limited up to a finite choice. In order 

to overcome this discrepancy, we propose a novel method of "infinite’ kernel 

combinations by infinite and semi-infinite programming regarding all elements 

in kernel space. This provides to study variations of combinations of kernels 

when considering heterogeneous data in real-world applications. 

� MF-06 


8.2.30 

DEA Methodology V 



Chair: Marcos Estellita Lins, Production Engineering, Federal 



lins@pep.ufrj.br

1 - Ratio-Based Efficiency Analysis (REA) 

Antti Punkka, Systems Analysis Laboratory, Aalto University 

School of Science and Technology, SAL, Aalto University, 

P.O.Box 11100, 02150, Espoo, Finland, antti.punkka@tkk.fi, 

Ahti Salo 

A DMU’s efficiency can be defined as the ratio between the sum of its weighted 

outputs and that of its weighted inputs. The resulting efficiency ratio depends 

on what aggregation weights are employed. We develop results for REA which 

show what other DMUs a given DMU dominates by having a higher efficiency 

for all feasible weights; what rankings the DMU can attain among all DMUs; 

and how efficient the DMU can be relative to the most and least efficient DMU 

across the set of feasible weights. We formulate LP/MILP models from which 

these results can be computed for realistically large problems. 

2 - Measuring the balance of DEA efficiency scores: The 

balance score Beta 

Nadia Vazquez Novoa, Institute of Management Control and 

Business Accounting, TU Braunschweig, Pockelsstraße 14, 

38106, Braunschweig, Lower Saxony, Germany, 

n.vazquez-novoa@tu-bs.de, Heinz Ahn, Ludmila Neumann, Juan 

Pablo De Francesco 

The endogenous weight determination is one of the main advantages of DEA, 

but it also leads to cases where the efficiency scores result from extreme criteria 

weights. A common approach to avoid such "unbalanced’ results is to integrate 

weight constraints into the DEA models. However, weight restrictions require 

additional — more or less subjective — information. As an alternative approach, 

a balance score is presented. This measure endogenously identifies 

DMUs with unbalanced efficiency scores, quantifies their (un-)balance, and 

leads to recommendations of how to improve their balance. 

3 - Three way decomposition of the Efficiency of Andalusian 

Economy 

Antonio F. Amores, Departament of Economics, Quantitatives 

Methods and Economics History, Pablo de Olavide University, 

Ctra. Utrera Km.1, Ed. 3, 2 o , Despacho 16, 41013, Seville, 

Spain, afamoher@upo.es, Thijs ten Raa 

Ten Raa (2006) proposes a model for the efficiency evaluation of an industrial 

organization that singles out firm, industry and organization efficiency. We 

extend this approach to tackle the whole economy efficiency and its decomposition 

into: firm, industry, organization, allocation and economy efficiencies. 

This analysis is applied to the Andalusian economy (Spain). 

4 - The use of Problem Structuring Methods in health policy 

and performance assessment 

Marcos Estellita Lins, Production Engineering, Federal 



lins@pep.ufrj.br, Angela Silva, Maria Stella Castro Lobo, 

Roberto Fiszman, Leonardo Pessoa, Nilo Chagas 

Integrating soft and hard OR as seemingly conflicting paradigms provides a 

better communication among decision makers regarding the several stages of 

applied OR problem solving. This is particularly true when public policy is 

concerned. While this is well developed in OR, the same is not true in the DEA 

field, where a pioneering research was done by Mingers in 2009 using PSM 

to help selecting DEA variables. In this work we use cognitive maps to help 

understanding the causal relationship and weaknesses of DEA when reflecting 

performance and supporting decisions concerning public policy. 

� MF-07 


8.2.47 

DEA Application IX 



Chair: Meryem Duygun Fethi, School of Management, University of 

Leicester, University Road, LE1 7RH, Leicester, Leicestershire, 

United Kingdom, m.fethi@le.ac.uk 


1 - Reallocating agricultural greenhouse gas emissions 

among European countries through a Zero Sum Gains 

DEA Model 

Amílcar Serrão, Management Department, Evora University, 

Largo dos Colegiais, 7000-550, Evora, Alentejo, Portugal, 

aserrao@uevora.pt 

This research work uses a ZSG-DEA BCC model, which represents a situation 

similar to a zero sum game. This model is applied to EU15 countries, 

considering one output (agricultural greenhouse gas emissions) and two inputs 

(livestock units and utilized agricultural area). Using the smoothed frontier for 

the 3-dimensional DEA BCC scores, we determine new targets for this ZSG- 

DEA BCC model, reallocating agricultural greenhouse gas emissions among 

EU15 countries. A uniform BCC DEA frontier is built, where all DMUs are 

100% efficient. After emissions reallocation, all DMUs became efficient. 

2 - Efficiency Study of Indian Banks during the Reforms — 

A Data Envelopment Analysis Approach 

Debaprosanna Nandy, COMMERCE, A.C. COLLEGE OF 

COMMERCE, D.B.C. ROAD, 735 101, JALPAIGURI, WEST 

BENGAL, India, dp_nandy@yahoo.com 

Indian Banks are going through a transitional phase. An important objective 

of reforms measures is to increase the efficiency of the banking sector. Policy 

makers have clearly recognized that inefficiency is an important factor contributing 

to the high level of cost of banking services. Indian banks have to 

be more efficient and competitive to cope with the financial crisis as well as 

global challenges in the near future. This paper attempts to measure the efficiency 

of Indian Banks using a variety of efficiency measures computed by the 

nonparametric method of Data Envelopment Analysis. 

3 - Efficiency analysis of the MFIs in Bolivia 

Elena Sanchez, Economics, Self employee, Calle Waldo Ballvian 

1018, n.a, Cochabamba, Bolivia, elenas55@hotmail.com 

The DEA was used to measure the efficiency of 12 MFIs in Bolivia, by using 

3 types of efficiency: technical, pure technical, and scale efficiency, under 

IOM and OOM, assuming CRS and VRS. 2 variables are evaluated: gross loan 

portfolio and number of borrowers. the findings are:4 MFIs are on the efficient 

frontier when GLP is evaluated, inefficienciences are due to scale inefficiency, 

5MFIs are on the EF when No. of borrowers is evaluated, inefficiences are due 

PTE. In overall, TI is due PTI rather than SI in both aprroaches. The main 

conclusion: 64% of MFIs seems to be efficient. 

4 - Efficiency, capital structure and institutional shareholders 

Tsai Lien Yeh, International Business, Ming Chuan University, 

2F, No. 9, Lane 17, Yung-Kang St.,Taipei, Taiwan, 10650, 

R.O.C., 10650, Taipei, Taiwan, tlyeh@mail.mcu.edu.tw, 

Jhu-Ning Jhang 

By employing a sample of 37 Taiwanese banks, this study investigates the relationship 

among banks efficiency, capital structure and institutional shareholders. 

The data envelopment analysis approach is used to measure technical efficiency 

as the firm performance indicator. Empirical outcomes support that: 

first, banks can mitigate agency cost to increase firm efficiency by decreasing 

the debt ratio. Second, debt and the presence of institutional shareholders may 

be deemed as substitute disciplinary devices, the share of institutional holdings 

is negatively associated with debt ratio. 

� MF-08 


6.1.36 

Scheduling: Algorithms and practical cases 



Chair: Mario Vanhoucke, Faculty of Economics and Business 

Administration, Ghent University and Vlerick Leuven Gent 

Management School, Tweekerkenstraat 2, 9000, Ghent, Belgium, 

mario.vanhoucke@ugent.be 

1 - A meta-heuristic optimization approach for job shop 

scheduling and its practical applications 

Veronique Sels, Faculty of Economics and Business 

Administration, Ghent University, Faculteit Economie en 

Bedrijfskunde, Tweekerkenstraat 2, 9000, Gent, Oost 

91


Vlaanderen, Belgium, veronique.sels@ugent.be, Mario 

Vanhoucke 

In this presentation we present a comparison of different meta-heuristic optimization 

approaches for the well-known job shop scheduling problem. A 

detailed comparison of these procedures will be made based on benchmark 

test instances from literature and their solution quality will be compared with 

state-of-the-art results. Moreover, the practical use of these procedures will be 

illustrated in a simulation experiment used during a consultancy project in a 

Belgian production company. 

2 - Managerial Insights in Reactive Personnel Re-rostering 

Broos Maenhout, Business Informatics and Operations 

Management, Ghent University, Tweekerkenstraat 2, 9000, Gent, 

Belgium, Broos.Maenhout@Ugent.be, Mario Vanhoucke 

In case of schedule disruptions caused by unplanned personnel absences, the 

personnel scheduler must restore feasibility by reconstructing the personnel 

roster. This rescheduling is a complex task as the personnel scheduler should 

take multiple objectives and many constraints into account. In order to reduce 

the complexity, we explore the boundaries of the time horizon and the personnel 

staffing size that will be considered by the re-rostering process based on 

computational experiments in a real-life problem environment. 

3 - An Experimental Investigation of Meta-heuristics for the 

Multi-mode Resource-constrained Project Scheduling 

on new Dataset Instances 

Vincent Van Peteghem, Faculty of Economics and Business 

Administration, Ghent University, Tweekerkenstraat 2, 9000, 

Gent, Belgium, vincent.vanpeteghem@ugent.be, Mario 

Vanhoucke 

An overview is presented of the existing meta-heuristic solution procedures 

available in literature to solve the multi-mode resource-constrained project 

scheduling problem, in which multiple execution modes are available for each 

of the activities of the project. A fair comparison is made between the different 

meta-heuristic algorithms on the existing benchmark datasets and on a newly 

generated dataset. Computational results are provided and recommendations 

for future research are formulated. 

4 - Audit-staff scheduling with alternative audit teams and 

setup times 

Vincent Van Peteghem, Faculty of Economics and Business 

Administration, Ghent University, Tweekerkenstraat 2, 9000, 

Gent, Belgium, vincent.vanpeteghem@ugent.be, Mario 

Vanhoucke 

In this paper, an algorithm is presented for the medium term audit-staff scheduling 

problem, in which teams of auditors are assigned to a set of audit engagements, 

with varying auditors’ availability, alternative audit teams and variable 

setup times, dependent on the audit team efficiencies. This algorithm is applied 

on real-life data from a small Belgium audit firm, with 15 auditors and more 

than 250 audit engagements per year. 

� MF-09 


6.2.53 

Discrete and Continuous Optimization for 

Gas Networks 



Chair: Armin Fügenschuh, Optimierung, Zuse Institut Berlin, 

Takustraße 7, 14195, Berlin, Germany, fuegenschuh@zib.de 

1 - Checking Feasibility in Stationary Models of Gas Transportation 

Claudia Stangl, Mathematics, University of Duisburg-Essen, 

Butlerstrasse 10, 47058, duisburg, Germany, 

claudia.stangl@uni-due.de, Rüdiger Schultz 

Checking the feasibility of transportation requests belongs to the key tasks in 

gas pipeline operation. In its basic form, the problem is to decide whether a 

certain quantity of gas can be sent through the network from prescribed entries 

to prescribed exit points. In the stationary case, we obtain a very big (nonlinear, 

mixed-integer, finite dimensional) inequality system. We present elimination 

and approximation techniques so that the remaining system gets within the 

reach of standard NLP-solvers. 

92 

2 - High Accuracy Stationary Optimization in Gas Networks 

Martin Schmidt, Institute of Applied Mathematics, Leibniz 

Universität Hannover, Welfengarten 1, 30167, Hannover, Lower 

saxony, Germany, mschmidt@ifam.uni-hannover.de, Marc 

Steinbach, Bernhard Willert 

Stationary optimization of fuel gas consumption is an important task in current 

planning processes for gas networks. We consider detailed models of large 

networks with gas dynamics described by a system of ODEs including gas temperature 

and gas quality parameters. Additionally, we incorporate compressor 

stations with several compressor types in high detail. Discrete decisions are addressed 

at a separate MIP level and fixed in our nonlinear program. We discuss 

the NLP model and present numerical results on real life network sizes. 

3 - Topology Planning in Gas Distribution Networks 

Jesco Humpola, Zuse Institute Berlin, Germany, 

humpola@zib.de 

A procedure of topology planning within large-sized, real-world gas distribution 

networks is presented. Given a budget, we decide which combination of 

network extensions like pipelines, compressors or valves should be added to 

the gas network to increase its capacity as well as its flexibility. We present a 

nonlinear mixed-integer model for this problem. To solve this large scale instance, 

every nonlinearity undergoes a linear outer approximation. We present 

computational results obtained by a special tailored version of the MILP solver 

SCIP. 

4 - Large-scale decomposition with Fejer processes for 

convex programming 

Andrey Velichko, Institute of Automation and Control Processes, 

Radio, 5, 690041, Vladivostok, Russian Federation, 

vandre@dvo.ru 

Fejer processes with projection operator studied by Eremin and small decreasing 

disturbances approach analysed by Nurminski are used to design iterative 

and parallel algorithms for convex programming with large quantity of linear 

inequality constraints. Convex hull of projections onto disjoint subsets of a 

feasible region is used as Fejer operator. Selection of initial point, step adjustments 

and feasible set decomposition strategies are proposed to get linear 

convergence and nearly polynomial (in constraints) algorithm complexity of 

degree 4-5. MPI Toolbox package for Octave software is used. 

� MF-10 


6.2.56 

Multi-index Assignment Problems 



Chair: Frits Spieksma, of Operations Research and Business 

Statistics, Katholieke Universiteit Leuven, Naamsestraat 69, B-3000, 

Leuven, Belgium, frits.spieksma@econ.kuleuven.be 

1 - On the Completability of Incomplete Latin Squares 

Reinhardt Euler, Informatique, Université de Brest, 20 av. Le 

Gorgeu, B P 817, 29285, Brest, France, 

reinhardt.euler@univ-brest.fr 

Latin squares correspond to the solutions of the planar 3-index assignment 

problem. The question whether an incomplete latin square is completable, 

arises as a special case, and establishing necessary and sufficient conditions 

for the completability of such squares is a major research direction. In this talk 

we focus on minimally non-completable incomplete latin squares and - point 

out their relationship with polyhedral theory; - answer the completability question 

for incomplete (3,n)-latin rectangles. We also discuss some consequences 

for class-teacher time-table problems. 

2 - On the Multi-level Bottleneck Assignement Problem 

Trivikram Dokka, Operations Reseach and Business Statistics, 

KATHOLIEKE UNIVERSITEIT LEUVEN, Naamsestraat 69,

B-3000, Leuven, Belgium, trivikram.dokka@yahoo.co.uk, Frits 

Spieksma, Gautam Appa, Anastasia Kouvela 

We consider the multi-level bottleneck assignment problem (MBA). This problem 

is described in the recent book "Assignment Problems" by Burkard et al. 

(2009) on pages 188 - 189. One of the applications described there concerns 

bus driver scheduling. We view the problem as a special case of a bottleneck 

m-dimensional multi-index assignment problem. First, we note that problem 

MBA is NP-hard even for m = 3. Further, we describe a method for MBA that 

consists in solving iteratively bottleneck matching problems, and we show that 

this method is a 2-approximation algorithm. 

3 - Characterising odd-hole inequalities related to Latin 

squares 

Yiannis Mourtos, Management Science & Technology, Athens 

University of Economics & Business, Greece, mourtos@aueb.gr, 

Dimitris Magos 

The convex hull of binary vectors representing Latin squares is the polytope of 

the 3-index planar assignment problem. We study the facial structure of P by 

examining valid inequalities induced by the odd holes. We define the concept 

of the lifting set of an odd hole, present an efficient algorithm for identifying 

it and derive tight bounds on the left-hand side coefficients of an induced 

odd-hole inequality. Hence, we characterise the class of odd holes that yield 

maximal inequalities without lifting and show that they are facet-defining, thus 

unifying and generalising previous. Finally, we show how these inequalities 

can be generalised to planar multi-index assignment problems. 

4 - The Focus of Attention Problem 

Frits Spieksma, of Operations Research and Business Statistics, 

Katholieke Universiteit Leuven, Naamsestraat 69, B-3000, 

Leuven, Belgium, frits.spieksma@econ.kuleuven.be, Dries 

Goossens, Sergey Polyakovskiy, Gerhard J. Woeginger 

We consider the problem of assigning sensors to track targets so as to minimize 

the error in the resulting estimation for target locations. This so-called Focus 

of Attention problem is a special case of a three index assignment problem. 

We provide a complete complexity and approximability analysis of this Focus 

Of Attention problem: We establish strong NP-hardness, and we construct a 

polynomial time approximation scheme. 

� MF-11 


8.2.38 

Structural Equation Modelling Approach in 

User Acceptance of Information Technology 

I 



Chair: Sevgi Ozkan, Information Systems, Middle East Technical 

University, ODTU Enformatik Enstitüsü, Ismet Inönü Bulvari, 06531, 

Ankara, Turkey, sozkan@ii.metu.edu.tr 

1 - Analysis of the User Acceptance for Implementing 

ISO/IEC 27001:2005 in Turkish Public Organizations 

Tolga Mataracioglu, Department of Information Systems 

Security, TUBITAK UEKAE, Tunali Hilmi Cad. Binnaz Sok. 

2/3., Kavaklidere/Cankaya, 06700, Ankara, Turkey, 

mataracioglu@uekae.tubitak.gov.tr, Sevgi Ozkan 

This study aims to develop a model for the user acceptance for implementing 

the information security standard (i.e. ISO27001) in Turkish public organizations. 

The results of the surveys performed within four public organizations in 

Turkey reveal that the legislation on information security public which organizations 

have to obey is significantly related with the user acceptance during 

ISO27001 implementation process. The fundamental components of our user 

acceptance model are perceived usefulness, attitude towards use, social norms, 

performance expectancy and effort expectancy. 

2 - Information Technology Assessment of Organizations 

Emre Sezgin, Information System, Informatics Institute, Middle 

East Technical University, Informatics Institute, 06531, Ankara, 

Turkey, esezgin1@gmail.com, Sevgi Ozkan 


The aim of the study is proposing a model for IT assessment that can be applied 

to companies to evaluate their condition of IT and to determine their IT 

level in comparison with their rivals. For this reason, Information Technology 

Assessment Model has been developed to assess the information technology 

use, advancements and potentials of a company. The model is based on specific 

research criteria retrieved from best practices (COBIT, ITIL, CMMI) and 

standards (ISO 38500, ISO 27002) considering technology management assessment 

procedure (Gregory, 1995) as the academic framework of the study. 

3 - Investigating Technology acceptance Model in context 

of Mobile Health Care 

Can Peker, IS, METU, METU, ANKARA, Çankaya, Turkey, 

cpeker@gmail.com, Sevcen Yazarli, Sevgi Ozkan 

Healthcare industry is one of the most challenging industries considering the 

new Information technologies. Mobile healthcare systems are evolving for distance 

information. IT applications like global positioning systems and wireless 

technologies offer remote connection to Information. Mostly the systems 

cannot be adopted because of insufficient user acceptance. In this work Technology 

Acceptance Model (Davis. 1989) will be studied. The related social 

and individual factors will be served for the acceptance of Mobile healthcare 

Systems. 

4 - Analysis of the Possible Adoption Behaviors of Cloud 

Services in Turkish Public Organizations 

Bilge Karabacak, Informatics Institute, METU, METU 

Informatics Institute, Inonu Bulvari, 06531, Ankara, TR, Turkey, 

e171018@metu.edu.tr, Sevgi Ozkan 

Cloud computing is the final era of the Internet in which applications and data 

resides at remote sites of the cloud providers. Since data is processed at remote 

location, cloud computing raises some concerns on privacy of information. 

Some public organizations in United States are already using services for 

a while. Similar models of usage can be applicable for Turkish public bodies. 

Considering these models, our study aims to examine the possible adoption 

behaviors of the cloud services in public bodies of Turkey by taking the regulations, 

organizational politics and culture into account. 

� MF-12 


8.2.39 

AHP 05 


Processes 


Chair: Josef Jablonsky, Dept. of Econometrics, University of 

Economics Prague, W.Churchill sq. 4, 13067, Prague 3, Czech 

Republic, jablon@vse.cz 

1 - Influence of the Criteria in the Bayesian AHP 

Pilar Gargallo, Facultad de Económicas, Universidad de 

Zaragoza, Gran Vía 2, 50005, Zaragoza, Spain, 

pigarga@unizar.es, José María Moreno-jimenez, Alfredo 

Altuzarra, Manuel Salvador 

This work proposes different methodologies for measuring the influence of a set 

of criteria on the final priorities of the Analytic Hierarchy Process (AHP) in a 

global context (a hierarchy). The priorities have been obtained by means of the 

Bayesian prioritization procedure of Altuzarra et al. (2007). Cross-validation 

methods have been used when measuring the influence. The methodology is 

illustrated by means of an empirical example. 

2 - Multiple Criteria Decision Making in Management Accounting 

and Control — State of the Art and Research 

Perspectives 

Judith Huelle, Chair of Management Accounting and Control, 

University of Goettingen, Platz der Göttinger Sieben 3, 37073, 

Goettingen, Lower Saxony, Germany, 

Judith.Huelle@wiwi.uni-goettingen.de, Ralf Kaspar, Klaus 

Moeller, Tobias Klatt 

93


The study’s purpose is to examine to what extent MCDM methods are used 

in the context of management accounting and control. Therefore, an extensive 

bibliometric analysis covering the last three decades was conducted using Business 

Source R○ Complete database. The results indicate an increase in MCDM’s 

importance. Furthermore, the majority of publications deal with the areas of 

strategic and performance management, more specifically with strategic planning 

and performance evaluation. Hereby, AHP is the most popular tool for 

supporting the management in complex decisions. 

3 - Analysis of Investments in Alternate Energy Resources: 

AHP and DEA Approach 

Josef Jablonsky, Dept. of Econometrics, University of 

Economics Prague, W.Churchill sq. 4, 13067, Prague 3, Czech 

Republic, jablon@vse.cz 

Recoverable energy resources are of high interest in many countries as one 

of the possibility to ensure their energy needs. They are often supported in 

different ways by governments. The paper analyses possible investments in 

five energy resources (wind power, small water, photovoltaic, geothermal, and 

biomass energy power plant). The alternatives are evaluated according to five 

groups of criteria (technical, ecological, economic, social, and strategic) by 

means of AHP model. Except the AHP model we use DEA analysis for evaluation 

of efficiency of private investments in given field. 

� MF-13 


2.2.21 

Discrete Location II 



Chair: Antonio Manuel Rodríguez-Chía, Estadística e Investigación 

Operativa, Universidad de Cadiz, Facultad de Ciencias. Pol. Río San 

Pedro., 11510, Puerto Real, Cádiz, Spain, 

antonio.rodriguezchia@uca.es 

1 - Solving discrete location problems by neural networks 

Enrique Dominguez, Dept. of Computer Science, 

E.T.S.I.Informatica - University of Malaga, Campus Teatinos s/n, 

29071, Malaga, Spain, enriqued@lcc.uma.es, Jose Muñoz 

This paper presents a new bidirectional neural model for solving discrete location 

problems. The proposed neural model (NELOC) is based on a bidirectional 

architecture composed by two layers. NELOC have been successfully 

applied to diverse discrete location problems. In addition, the effectiveness 

and efficiency of NELOC have been analyzed in comparison to other heuristics 

methods. Results show that the proposed neural network generates good solutions 

for different discrete location problems with a reasonable computational 

effort. 

2 - Facility location planning for a multi-layer biofuel production 

chain 

Frank Schwaderer, Institute for Industrial Production (IIP), 

Karlsruhe Institute of Technology (KIT), Germany, 

frank.schwaderer@kit.edu, Magnus Fröhling, Frank Schultmann 

The considered biofuel production chain consists of the following process 

steps: preparation, pyrolysis, gasification and synthesis. As the pyrolysis product 

is transportable the pyrolysis unit can be operated at independent production 

sites. In order to identify the number of production sites and steps as well as 

capacities at each location, a WLP is formulated regarding economies of scale 

through special ordered sets of type 2. The model considers energy balances of 

the single process steps and regional conditions such as biomass potentials and 

is exemplarily applied to a specific region. 

3 - Solving the the uncapacitated hub location problem 

with a row generation algorithm 

Sergio García Quiles, Department of Statistics, University Carlos 

III of Madrid, Escuela Politécnica Superior, Avenida de la 

Universidad, 30, 28911, Leganés, Madrid, Spain, 

sergio.garcia@uc3m.es, Mercedes Landete, Alfredo Marín 

In hub location problems, some product must be sent between the nodes (customers) 

of a network through some special transhipment points called hubs: 

nodes which benefit from a economy of scales. It has many applications in 

fields such as Telecommunications, Air Transport or postal delivery. 

This work shows how the uncapacitated hub location problem can be solved by 

using an initial reduced formulation and adding rows dynamically as they are 

needed. A computational study shows the performance of this model. 

94 

� MF-14 


2.2.15 

Supply Chain Planning 



Chair: Shabnam Rezapour, Industrial Engineering, Urmia university 

of Technology, 14185-671, Urmia, Vest Azarbaijane, Iran, Islamic 

Republic Of, shabnam_rezapoor@yahoo.com 

1 - The effect of workload constraints in periodic order release 

models 

Michiel Jansen, Industrial Engineering, Eindhoven University of 

Technology, Postbus 513, Paviljoen E.14, 5600 MB, Eindhoven, 

Netherlands, m.jansen@tue.nl, Ivo Adan 

We study periodic capacitated models for Supply Chain Operations Planning 

(SCOP) within a hierarchical planning concept. The SCOP problem is one of 

timing the release of orders to production units (PU). An important instrument 

in the coordination of order releases is the planned lead time, i.e. the time 

between order release and the planned availability of goods. In capacitated 

models for SCOP, the planned lead time is made conditional on some (restriction 

of) the work-in-process (WIP). The choice of the planned lead time and 

WIP constraints heavily influences WIP levels and how efficient resources can 

be utilized. We explore these relations for a PU that is represented by a single 

server with generic service times. We find that there is a simple, intuitive 

relation between the planned lead time, WIP constraints, and the maximum 

utilization of a production unit. We explore this relation further using a generating 

function approach. We also develop accurate approximations for mean 

and variance of the WIP and the order flow times. Finally, we show for a simple 

case, how these results can be applied to find the efficient combinations of 

planned lead time and the capacity constraint. 

2 - Operative planning of resources and operations in 

long-term planning context 

Jasmina Omerbegovic-Bijelovic, Operations management, 

Faculty of organizational sciences, Jove Ilica 154, 11040, 

Belgrade, Serbia, omeja@fon.rs, Zoran Rakicevic 

One of the most significant problems of Operations Man-t is the breaking 

down of the long-term (a few years/months) plans into operative ones (monthly, 

weekly). It manifests because of the set of limitaions in resources, which limit 

the exit (products/services scope) from the system in question. In practice, that 

problem solves by convention: long-term plans are agreements about potential 

needs for resources, and the operative plans are realised as business agreements 

(order for suppliers). Theoretically, the solution lies in shorter time frames, 

simulations, statistical methods, fuzzy sets. 

3 - A critical analysis of optimisation models for "Clicksand-Mortar’ 

integration 

Kathrin Fischer, Institute for Operations Research and 

Information Systems, Hamburg University of Technology 

(TUHH), TU Hamburg-Harburg, Schwarzenbergstrasse 95 D, 

D-21073, Hamburg, Germany, kathrin.fischer@tu-harburg.de 

In recent publications, inventory management and delivery strategies for 

"Clicks-and-Mortar’ firms were studied. Using optimisation models, it was 

found that deliveries to online customers should either be made only from the 

stores, or exclusively from an online depot. It is shown here that these results 

hold true only under very simplifying assumptions. Under realistic assumptions, 

combined solutions can be optimal. Moreover, it is discussed how the 

respective supply chain optimisation models can be adapted in order to develop 

strategies applicable to real-life planning situations. 

4 - Strategic Design of Competing Supply Chain Network 

for Markets with Stochastic Demands 

Shabnam Rezapour, Industrial Engineering, Urmia university of 

Technology, 14185-671, Urmia, Vest Azarbaijane, Iran, Islamic

Republic Of, shabnam_rezapoor@yahoo.com, Reza Zanjirani 

Farahani 

In this paper we consider a new two-stage model for competitive supply chain 

network designing with anticipating variable prices and service levels competition 

in markets under stochastic price and service level dependant elastic demands 

and with existing external rivals presence. The objective is to design the 

new entrant chain’s network under a capacity constraint in order to maximize 

its future income in the competing markets. The structure of the new chain’s 

network is assumed to be set "once and for all’ but further price and service 

level adjustments are possible. 

� MF-15 


2.2.12 

Shortest Path Problems with Resource 

Constraints 



Chair: Simon Spoorendonk, DTU Management Engineeing, DTU 

Technical University of Denmark, Produktionstorvet, Building 426, 

2800, Kgs. Lyngby, Denmark, spoo@man.dtu.dk 

1 - Resource Constrained Shortest Path found by Parallel 

Labeling Algorithm 

Bjørn Petersen, DTU Management Engineering, Technical 

University of Denmark, Produktionstorvet 424, 2800, Kongens 

Lyngby, Denmark, Denmark, bjorn@diku.dk 

Resource Constrained Shortest Path Problems are often solved with labeling 

algorithms. Modern hardware trends go toward more and more parallelization. 

This talk shows how to parallelize a labeling algorithm. Various other 

aspects of making a modern labeling algorithm perform is also shown, including 

search structures, bounding, and dominance rules. Extensive computational 

results based on test data from Vehicle Routing Problems are presented. 

2 - A New State Space Relaxation for Computing Bounds 

to Routing Problems 

Roberto Roberti, DEIS, University of Bologna, Via Sacchi, 3, 

47521, Cesena, Italy, roberto.roberti6@unibo.it, Roberto 

Baldacci, Aristide Mingozzi 

We introduce a new state space relaxation (SSP), called ng-path relaxation, for 

computing bounds and solving Vehicle Routing Problems (VRPs). We present 

the theoretical aspects of the new SSP and show that, computationally, it outperforms 

other SSPs proposed in the literature to solve VRPs (i.e., q-routes and 

t-routes relaxations). Extensive computational results over the main instances 

from the literature of the Capacitated VRP, the VRP with Time Windows, and 

the TSP with Time Windows show the effectiveness of the new relaxation. 

3 - Pricing Non-Elementary Routes for the Capacitated 

Arc-Routing Problem 

Stefan Irnich, Logistics Management, Johannes Gutenberg 

University, Jakob-Welder-Weg 9, 55128, Mainz, Germany, 

irnich@uni-mainz.de 

Traditional exact methods for the CARP rely on branch-and-cut or the transformation 

into the CVRP. An alternative exact approach is the solution of the 

CARP by column generation. Letchford and Oukil (Comp. & OR, 2009, vol. 

36, p. 2320–2327) price out elementary routes using a MIP formulation with 

directed flow variables. In contrast, we propose solving the subproblem directly 

as an undirected postman problem. We present a new sparse formulation 

for this subproblem and a branch-and-cut algorithm for its solution. 

4 - A branch-and-cut algorithm for the elementary shortest 

path problem with resource constraints 

Simon Spoorendonk, DTU Management Engineeing, DTU 

Technical University of Denmark, Produktionstorvet, Building 

426, 2800, Kgs. Lyngby, Denmark, spoo@man.dtu.dk 

The elementary shortest path with resource constraints have commonly been 

solved with dynamic programming algorithms. Assuming an undirected graph, 

we present a compact formulation of this problem and a branch-and-cut algorithm 

to solve it. Two types of resources are discussed: a capacity and a fixed 

charge resource. The former is the subproblem of the capacitated vehicle routing 

problem and the latter is from the split delivery version. Computational 

results are presented and compared to dynamic programming algorithms. 


� MF-16 


2.2.14 

Vehicle and crew rostering 



Chair: Ana Paias, DEIO/CIO, University of Lisbon, Portugal, 

ampaias@fc.ul.pt 

Chair: Marta Mesquita, ISA / CIO, Technical University of Lisbon, 

Tapada da Ajuda, 1349-017, Lisboa, Portugal, marta@math.isa.utl.pt 

1 - The integrated vehicle-crew-roster problem with daysoff 

pattern 

Marta Mesquita, ISA / CIO, Technical University of Lisbon, 

Tapada da Ajuda, 1349-017, Lisboa, Portugal, 

marta@math.isa.utl.pt, Margarida Moz, Ana Paias, Margarida 

Pato 

In this talk we present a new mathematical model for the integrated vehiclecrew-roster 

problem that follows the days-off pattern of a specific group of 

drivers from a bus company. We propose an heuristic approach with embedded 

column generation and branch-and-bound techniques within a Benders decomposition. 

Taking advantage of the mathematical model structure, the decomposition 

approach alternates between the solution of an integrated vehicle-crew 

scheduling problem and the solution of a rostering problem. We report on computational 

experience with data from a bus company operating in Lisbon 

2 - Robust Airline Schedule Planning: Minimising Propagated 

Delay in an Integrated Routing and Crewing 

Framework 

Gary Froyland, Mathematics and Statistics, University of New 

South Wales, School of Mathematics and Statistics, University of 

New South Wales, 2052, Sydney, NSW, Australia, 

g.froyland@unsw.edu.au 

The airline scheduling problem has traditionally been decomposed into stages 

with the decisions from one stage imposed upon decisions made in subsequent 

stages. This unfortunately fails to capture the many dependencies between the 

stages. As delays are commonly transferred between late running aircraft and 

crew, it is important that aircraft routing and crew pairing decisions are made 

together. We introduce a new approach to accurately calculate and minimise 

the total cost of propagated delay, in a framework that integrates aircraft routing 

and crew pairing. 

3 - Heuristic for Multiple Trips Vehicle Routing and 

Scheduling Problem with Time Window 

San Nah Sze, Econormetrics & Business Statistics, University of 

Sydney, Room 282, The Economics&Business Building (H69), 

2006, Sydney, New South Wales, Australia, 

susana_sze@hotmail.com, Suk Fung Ng 

In this study we investigate on how to apply heuristic effectively into Multiple 

Trips Vehicle Routing and Scheduling Problem with Time Window 

(MTVRSTW). The problem consists of establishing start-times for the customers 

and creating a roster for serving teams. Two-staged heuristic is proposed 

to solve MTVRSTW due to its fast responses of solution generation and 

good result for large size instances. Since solving this problem can be financially 

significant, the primary objective of this study is to minimize the number 

of servicing teams required to fulfill all the operational constraints. Computational 

results are given to demonstrate the robustness and efficiency of the 

insertion algorithm. 

4 - A maximum covering formulation for the integrated vehicle 

and crew scheduling problem 

Teresa Galvão Dias, DEIG, Faculdade de Engenharia da 

Universidade do Porto, Rua Dr. Roberto Frias, s/n, 4200-465 

Porto PORTUGAL, 4200-465, Porto, Portugal, 

tgalvao@fe.up.pt, Jorge Pinho de Sousa, Bruno Prata 

The Vehicle and Crew Scheduling Problem (VCSP) is a hard, widely studied 

Combinatorial Optimization problem. Traditionally set covering based models 

have been used to approach the problem. This research proposes a new 

mathematical formulation for the VCSP without changeovers (i.e. drivers cannot 

use more than one vehicle) based on the maximum covering problem. A 

reactive GRASP heuristic was also developed for the problem. Preliminary 

computational results with instances from the literature and real instances from 

Fortaleza (Brazil) show the potential of the developed approaches. 

95


� MF-17 


1.3.14 

Long-term Transportation Planning 



Chair: Rajeev Namboothiri, CIRRELT, Montreal, Canada, C.P. 6128, 

succursale Centre-ville, H3C 3J7, Montreal, Quebec, Canada, 

rajeev@crt.umontreal.ca 

1 - A comprehensive evaluation of the impact of aggregation/disaggregation 

of data on the strategic planning of 

freight transportation systems 

Rajeev Namboothiri, CIRRELT, Montreal, Canada, C.P. 6128, 

succursale Centre-ville, H3C 3J7, Montreal, Quebec, Canada, 

rajeev@crt.umontreal.ca, Teodor Gabriel Crainic, Michel 

Gendreau, Alexandre Savariradjou 

In this talk, we analyze the strategic national/regional planning of multicommodity 

multi-modal freight transportation systems using an integrated 

evaluation platform. A comprehensive and realistic representation of the current 

state of such a system was developed, incorporating the various components 

of these systems and their complex interactions. Computational results 

provide detailed analysis of the impact of aggregation/disaggregation of data 

on strategic planning decisions as well as day-to-day operational decisions. 

2 - The influence of subcontracting on location decisions 

of small package shippers 

Andreas Stenger, IT-based Logistics, Goethe University 


stenger@wiwi.uni-frankfurt.de, Michael Schneider, Michael 

Schwind 

Location routing problems (LRP) are used to determine the optimal number 

and location of depots considering vehicle routing. Published models still 

lack important characteristics of real-world delivery networks. In particular, 

outsourcing unprofitable areas to subcontractors is one major trend in small 

package shipping that strongly influences location decisions. We contribute by 

developing a solution method for an LRP that includes the choice between selfoperating 

and subcontracting a depot. In numerical studies, we show the effect 

of subcontracting on the total network design and costs. 

3 - Integrated Facility Location and Multi-Trip Vehicle Routing 

Problem: Solution and Value of Integration 

Zeliha Akca, Investment Management, Strategic Planning and 

Investments, Turkish Airlines Inc., Istanbul, Turkey, 

zelihaakca@gmail.com, Rosemary Berger, Ted Ralphs 

We investigate the problem of simultaneously determining the location of facilities 

and the design of multi-trip vehicle routes to serve customer demands. For 

a version of this problem with capacitated facilities, time- and capacity-limited 

vehicles, we describe two versions of a branch-and-price algorithm, a one-stage 

and a two-stage version which is based on the idea of restart. We demonstrate 

the performance of the algorithms using instances up to 40 customers and we 

assess the benefit of integrated optimization. We obtain results confirming our 

interest in integrated problem. 

� MF-18 


1.3.15 

Markov Chains 











96 

1 - absorbent chains of markov: study of costs 

Silvana Ligia Vincenzi Bortolotti, Matemática/Estatística, 

UTFPR/UFSC, Rua Jornalista Tito de Carvalho, 155, 88040480, 

Florianopolis, Santa Catarina, Brazil, sligie@globo.com, Rosely 

Antunes de Souza, Afonso Farias Sousa Junior, Antönio Coelho 

This work brings a study and analysis of products costs of a micro-company 

using the absorbent chains of Markov. The data was summarized through a 

project of productive system, considering aspects such as production, inspection, 

dispatching, rejection. The analysis of products cost in this productive 

system is done starting from the determination and estimative of all the costs 

that involve the phases of production, inspection and dispatching.Using the absorbent 

chains of Markov it was possible to verify its efficient contribution for 

taking decisions in processes of costs management. 

2 - Hamiltonicity-Trace Conjecture for Singularly Perturbed 

Markov Chains 

Vladimir Ejov, Mathematics and Statistics, University of South 

Australia, 46 Aver Avenue, 5041, Daw Park, SA, Australia, 

contactways@yahoo.com, Nelly Litvak, Giang Nguyen, Peter 

Taylor 

We prove that the trace of the fundamental matrix of a singularly perturbed 

Markov chain that corresponds to a stochastic policy, feasible for a given graph, 

is minimised at policies corresponding to Hamiltonian cycles. 

This is joint work with Nelly Litvak, Giang T. Nguyen and Peter G. Taylor. 

3 - Modelling operational decisions in start-up firms 

Thomas Archibald, Management School, University of 

Edinburgh, 50 George Square, EH8 9JY, Edinburgh, United 

Kingdom, tarchibald@ed.ac.uk, Kuangyi Liu 

A general Markov decision process model for operations management with a 

profit maximizing objective is presented and extended to a survival maximizing 

objective with a constraint on capital. It is argued that the survival maximizing 

objective may be more suitable for start-up firms. The model has been used to 

address inventory, capacity expansion and marketing decisions. Analysis of the 

models under different assumptions about the operating environment provides 

insight into the successful management of start-up firms. 

4 - Prospective customer equity measurement and monitoring 

Nadine Losch, Department of Management Accounting, 

University of Goettingen, Platz der Göttinger Sieben 3, 37073, 

Göttingen, Germany, nadine.losch@wiwi.uni-goettingen.de, 

Klaus Moeller 

Customer equity (CE) is gaining increasing importance in both academia and 

practice. Companies seek to measure and monitor CE, as a key driver of their 

financial performance. The prediction of CE is essential in a strategic management 

characterized by growing uncertainty through dynamics in the customer 

base. A markov chain model will be used to explicitly model the dynamics 

of CE over time. Furthermore the use of classification and regression trees 

(CART) enables to create a segment specific CE. The prospective CE enables 

the derivation of a monitoring system and broad strategic trade-offs. 

� MF-19 



Game Theory and Statistics 



Chair: Alberto Pinto, Mathematics, University of Minho, Campus de 

Gualtar, 4710-057, Braga, aapinto@math.uminho.pt 




1 - Reliability of test score 

Mikhail Lutsenko, Mathematics, St. Petersburg Transport 

University, Moskovskij pr.9, 195257, St.Petersburg, Russian 

Federation, ml4116@mail.ru

Reliability of test scores without assumption about normality of levels knowledge 

of test takers is found. A statistical game between Statistician and Testtaker 

for solution of the problem is constructed. The worse a priori distribution 

of parameter, the best randomizer decision function of Statistician and reliability 

of testing are evaluated. The problems by MS Excel are solved when test 

has 10 items. In many important cases the reliability of assessment turns out to 

be very low. 

2 - On k-th order arbitrage 

Fabio Bellini, Department of Quantitative Methods, University 

of Milano - Bicocca, P.zza Ateneo Nuovo 1, 20126, Milano, 

Italy, fabio.bellini@unimib.it 

We introduce a nested sequence of vector stochastic orderings in a space of 

payoffs and provide the corresponding notion of k-th order arbitrage. We characterize 

absence of k-th order arbitrage by means of the existence of suitable 

state price densities. We show the link with option pricing bounds based on 

stochastic dominance considerations, introduced by Perrakis and Ryan (1984), 

Levy (1985) and Ritchken (1985) among others. We show how these results 

can be generalized to markets with bid/ask spreads and provide numerical and 

empirical examples. 

3 - SMED: statistical analysis of setup time reduction in 

plastic injection machines 

Helena Alvelos, DEGEI, Universidade de Aveiro, Campus 

Universitário de Santiago, 3810-193, Aveiro, Portugal, 

helena.alvelos@ua.pt, Telmo Correia, Ana Raquel Xambre 

The work was developed in a department that includes 50 plastic injection machines. 

In order to improve their utilization rate and increase flexibility SMED 

(single minute exchange of die) principles were applied to the mould changing 

operation. Setup times were registered, before and after SMED application, 

and the impact was statistically analyzed using data from 21 machines. The 

study is a quantitative analysis of the benefits of SMED in a real world situation 

and contributes to a better understanding of the impact these methods have 

in the reduction of waste inside organizations. 

4 - Statistical analysis of Public and National accounts related 

with expenses of previous years 

Paula Santos, Faculdade de Ciênciais da Economia e da 

Empresa, Universidade Lusíada de Lisboa, R. da Junqueira 188, 

Lisboa, Portugal, paula.santos@oniduo.pt, Manuela Sarmento 

Portugal has to comply with the ESA 95 criteria, in which expenses of previous 

years assumes special significance since they are considered in the national accounts 

deficit. Thus, this paper intends to present how those expenses are being 

calculated. The results presented are based on surveys, carried out in General 

Government, in 2008. The methodology was based on the exact results of Pearson 

chi-square test and contingency tables. As a main conclusion, it is pointed 

out that the criteria adopted to recognize the expenses of previous years are not 

harmonized, limiting the convergence between Public and National Accounts 

and, especially, conditioning the calculation of the deficit reported to EU. 

� MF-20 


1.3.33A 




Chair: Pedro Castro, UMOSE, LNEG, 1649-038, Lisbon, Portugal, 

pedro.castro@ineti.pt 

1 - An exact approach for the 1D contiguous bin packing 

problem 

Marat Mesyagutov, Numerical Mathematics, Dresden University 

of Technology, Zellescher Weg 12 - 14 (C 319), 01069, Dresden, 

Germany, mmesyagutov@googlemail.com, Guntram 

Scheithauer, Gleb Belov 

The problem of the 1D contiguous bin packing problem is considered as the 

main problem of the talk. For its solution we propose an exact algorithm based 

on branch and bound method using linear programming. Solution of the posed 

problem is used to obtain improved lower bounds for the 2D strip packing problem. 

The computational results, which will be presented, are obtained on 2D 

strip packing problem instances. 


2 - Codifications in Evolutionary Algorithms for the Multi- 

Objective 2D Guillotine Strip Packing Problem 

Jesica de Armas, Estadística, I. O. y Computación, University of 

La Laguna, Avda. Astrofísico Francisco Sanchez, S/N, 38271, La 

Laguna, Spain, jdearmas@ull.es, Gara Miranda, Coromoto Leon 

The 2D Strip Packing Problem can be posed as a multi-objective optimisation 

problem. For the problem solution, we have applied some of the most-known 

multi-objective evolutionary algorithms. We proposed a solution codification 

which is based on a complete representation of pattern layouts. This approach 

was promising but wasn’t suitable to afford large instances. That is why we 

have focused on the design of a codification which can be much more competitive. 

So, we have proposed several hyperheuristic-based codifications covering 

different regions of the search space for large instances. 

3 - An improvement of the Best-Fit heuristic for 2D Strip 

Packing Problem 

Miroslav Rada, University of Economics in Prague, Czech 

Republic, miroslav.rada@vse.cz 

Paper deals with the Best-Fit heuristic for 2D-regular-ODP (2D strip packing 

problem), proposed by Burke et. al in 2004. The algoritm is based on finding 

the lowest possible gap and placing a dynamically selected rectangle inside 

it, and was shown to give high quality solutions. Strategy of selecting rectangle 

for the placement is modified in the paper. Efficient implementation of 

the strategy with time complexity O(n*log(n)) is suggested. Original algorithm 

can be implemented in O(n*log(n)), too. The proposed modification achieves 

better results for more than one half benchmark problems. 

4 - New MILP Model for the 2D Strip Packing Problem 

Pedro Castro, UMOSE, LNEG, 1649-038, Lisbon, Portugal, 

pedro.castro@ineti.pt, Jose Fernando Oliveira 

We propose a model relying on the concept of events to continuously locate 

the rectangles along the strip height. It is a mixed discrete/continuous-space 

model, continuous in the y-axis and discrete in the x-axis. The latter is divided 

into slots of unitary size and each partial strip can be associated to a different 

space grid. The location of event points will vary from one grid to the next and 

the challenge is to ensure that a given event point has the same y-coordinate in 

consecutive grids whenever there is a rectangle assigned to them. Results for a 

set of 29 instances are given. 

� MF-21 


6.2.47 

Optimization Modeling III 



Chair: David Canca, School of Engineers, University of Seville., Av. 

de los Descubrimientos s/n, Isla de la Cartuja, 41092, Seville, Spain, 

dco@us.es 

1 - Management of loading operations in the cement industry 

Tuomo Takkula, Logistics Systems Division, Inform GmbH, 

Pascalstraße 23, 52076, Aachen, Germany, 

tuomo.takkula@inform-ac.com 

In a typical cement plant a variety of products is loaded onto trucks, trains and 

ships. The products are moved from the silos to the loading points via a mechanical 

transportation layer. In general, the loading of a product at one loading 

point prevents the loading of certain products at other loading stations, turning 

load scheduling into a complex problem. We present a two-stage MIP-based 

multicommodity flow model which deals with these and other constraints and 

permits the scheduling of all load operations in an cost-effective way. 

2 - A Diet Planning Model for Malaysian Boarding School 

using Integer Programming 

Suliadi Firdaus Sufahani, Applied Mathematics, University Of 

Sheffield, Hicks Building, Hounsfield Road, S3 7RH, Sheffield, 

United Kingdom, app08sfs@sheffield.ac.uk 

97


Integer Programming is a mathematical method that can solve this type of problem 

and determines the most nutritious and palatable meals while considering 

the constraints of the Recommended Nutrient Intake for Malaysian children 

aged 13 to 18 years old, the cost of the menu items, the budget provided by 

the government, and the variety of menus. The problem is large and integer 

programming will be used. A system will be developed for the caterers of the 

Malaysian boarding schools by using LPSolve with Matlab. 

3 - Diet Scheduling-A Means To Aid Rural Health Care 

Sanchita Jha, CSE, PESIT, India, ruchi.superstar@gmail.com, 

Suraj Agarwal, Rishil Murukan Mypalli, Vinu Agrawal 

Diet scheduling is a relatively less explored problem in the field of operations 

research. So we have built a health care tool that takes into account the food 

habits,age,BMI, gender, physical activity, economical viability and nutritional 

requirement of any individual and outputs the minimum cost balanced diet; 

thereby combating ignorance and poverty, the two major problems of rural areas. 

This problem falls within the purview of linear programming model of 

operations research involving 108 decision variables and 16 constraints. We 

have used our own optimization algorithm to solve the same. 

4 - A mixed integer programming model applied to the optimization 

of cleaning procedures in a sunlight collector 

field. 

David Canca, School of Engineers, University of Seville., Av. de 

los Descubrimientos s/n, Isla de la Cartuja, 41092, Seville, Spain, 

dco@us.es, Pedro L. Gonzalez-R, Gabriel Villa 

Renewable energy, and more specifically those based on solar energy, are a 

clear alternative to the fossil fuel energy. There are many solar energy plants 

that are opening throughout the world. The solar plants efficiency is closely 

related with an appropriate policy in cleaning and maintaining processes. Unfortunately, 

these aspects are not usually taken into account in the actual design 

of plants. This work presents a mathematical model to address this issue. 

� MF-22 



Quantitative Health Care Policy Decision 

Making 

Stream: Health Care Management 


Chair: Marion Rauner, Dept. Innovation and Technology 

Management, University of Vienna, Bruennerstr. 72, A-1210, Vienna, 

Austria, marion.rauner@univie.ac.at 

1 - Disaster planning for ambulance services: a training 

tool using DES 

Marion Rauner, Dept. Innovation and Technology Management, 

University of Vienna, Bruennerstr. 72, A-1210, Vienna, Austria, 

marion.rauner@univie.ac.at, Helmut Niessner 

Due to severity, timing, location, and number of injured people related to a disaster, 

emergency staff faces different disaster situations and has to best cope 

with them. Besides fixed rules for triage of patients and treatment prioritization, 

emergency officers have to decide on organizational structures and on 

allocating ambulcance officers to triage tents or ambulance cars depending on 

the disaster situation. We have provided the Austrian Samaritan Organization 

which a graphical simulation tool for training their emergency staff to best handle 

decision making for numerous disaster sites. 

2 - Hospital productivity and technical progress across 

Austrian public hospitals 

Margit Sommersguter-Reichmann, Institute of Industrial 

Management, University of Graz, Universitaetsstrasse 15, 

Resowi, G2, 8010, Graz, Austria, 

margit.sommersguter@uni-graz.at, Adolf Stepan, Mex 

Glawischnig 

We examine productivity changes of Austrian hospitals (1999-2007) with a 

Malmquist productivity index. Former studies have revealed artificial rather 

than actual productivity progress as a result of using credit points in the output 

vector. The revision of the activity based system, however, has not resulted in 

the re-calculation of credits based on the obsolete system so that any structural 

changes are directly reflected in the productivity index. We use linear regression 

to analyse whether productivity changes can be explained by structural 

changes beyond the changes in the financing system. 

98 

3 - Estimating the Impact of Stochasticity in Operating 

Theatres 

Jean-Sébastien Tancrez, Operations Management, EPFL, Ecole 

Polytechnique Fédérale de Lausanne, EPFL - TOM, Odyssea 

2.19, Station 5, 1015, Lausanne, Switzerland, 

jean-sebastien.tancrez@epfl.ch, Benoît Roland, Jean-Philippe 

Cordier, Fouad Riane 

Even if it is clear, the stochastic nature of operating theatres is often bypassed in 

their management, or handled using simple rules. Our goal is to help managers 

rationalizing the consideration of stochasticity in OT. Based on the Markov 

theory, our approach allows estimating the impact of the randomness coming 

from the operating times, the unexpected emergencies and the blocking due to 

the recovery unit. For example, our tool evaluates the waiting time of the emergencies, 

the disruption of the planning, or the number of operations to plan so 

that the overtime is kept limited. 

� MF-23 


6.2.49 

MOO: Network Territorial Partition Problems 



Chair: José Rui Figueira, Instituto Superior Tecnico, Technical 

University of Lisbon, Av. Cavaco Silva, Tagus Park, 2780 - 990 Porto 

Salvo, 2780 - 990, Lisbon, Portugal, figueira@ist.utl.pt 

1 - Models for the integration of MCDA and GIS: Usage and 

weaknesses 

Karim Lidouh, Computer and Decision Engineering (CoDE), 

Université Libre de Bruxelles, Bvd du Triomphe CP210/01, 

1050, Brussels, Belgium, klidouh@ulb.ac.be, Esteban Zimányi, 

Yves De Smet 

Since a few years there has been a tendency to integrate methods from the 

MCDA field in geographical information systems (GIS). This integration allows 

taking multiple criteria into account when dealing with spatial decision 

problems. Several integration models have already been proposed to explain 

how some multicriteria methods should interact with GIS. However they often 

present some shortcomings and ultimately cannot be applied to any kind of 

spatial decision problem. In this contribution, we analyze some of the existing 

models and propose adaptations to broaden their domain of application. 

2 - Aggregating Census Units with a Multi-Objective Genetic 

Algorithm: Preliminary Results of a Case Study 

in Canada 

Dilip Datta, Mechanical Engineering, National Institute of 

Technology Silchar, India., Silchar, Cachar, Assam, 788010, 

Silchar, Assam, India, datta_dilip@rediffmail.com, José Rui 

Figueira, Jacek Malczewisk 

Small census units of London, Ontario are aggregated into census tracts (CTs) 

in conjunction with Statistics Canada. The aggregation aims at compactness 

of CTs, their population and area based homogeneities, as well as other subjective 

considerations. A degree of compactness is achieved, but with substantial 

discrepancies in the homogeneities. Therefore, a multi-objective genetic 

algorithm is investigated for improving the aggregation. Circular- and squareshaped 

CTs are studied by maximizing their compactness, as well as population 

and area based homogeneities. The results are considerably better than the 

actual CTs, in terms of all the three objectives for the two measures of compactness. 

3 - Graph Partitioning by a Real-Coded Multi-Objective Genetic 

Algorithm 

José Rui Figueira, Instituto Superior Tecnico, Technical 

University of Lisbon, Av. Cavaco Silva, Tagus Park, 2780 - 990 

Porto Salvo, 2780 - 990, Lisbon, Portugal, figueira@ist.utl.pt, 

Dilip Datta 

The graph partitioning problem is usually solved by problem-specific versions 

of an algorithm. Moreover, although various population-based metaheuristics 

are now in great consideration towards different problem-domains, these are 

yet to be studied widely to this problem. In this work, some mechanisms are 

proposed for handling the problem by a general real-coded multi-objective genetic 

algorithm. Applying to two large-size two-objective test cases, reasonably 

good solutions could be obtained. Moreover, the proposed real-coded 

algorithm is found outperforming an integer-coded genetic algorithm.

� MF-24 


6.2.50 

Natural Computation in BioInformatics 


Medicine 


Chair: Mario Pavone, Department of Mathematics and Computer 

Science, University of Catania, v.le A. Doria 6, 95125, Catania, Italy, 

mpavone@dmi.unict.it 

1 - Metabolic P Models for Biological Phenomena 

Giuditta Franco, Department of Computer Science, University of 

Verona, Italy, Ca’ Vignal 2 - Strada Le Grazie 15, 37134, Verona, 

Italy, giuditta.franco@univr.it, Vincenzo Manca 

Metabolic systems are traditionally modeled by differential equations, exhibiting 

intrinsic limitations in the actual evaluation of kinetic reaction rates. In this 

work we present MP systems, which are multiset grammars regulated by flux 

maps, that allow to deduce, from time series observed at macroscopic temporal 

scales, a discrete model of the bioprocess under investigation. This methodology 

opens the possibility to understand systemic logic of the process. The 

theory of MP systems is based on algebraic and algorithmic procedures. 

2 - A Formal Framework of Computational Modeling for 

Ecosystems Based on Membrane Computing 

M. Angels Colomer, University of Lleida, 25198, Lleida, 

colomer@matematica.udl.es, Antoni Margalida, Ignacio 

Pérez-Hurtado, Mario J. Pérez-Jiménez 

Our aims is to present a new computing bioinspired model in the formal framework 

of Membrane Computing, an emergent branch of Natural Computing. 

This new model of computation has been introduced with the aim of defining 

computing devices that abstract from the structure and the functioning of the 

living cells. Ecosystems are described by means of multienvironment P systems 

consisting of a finite number of environments, each of them having a specific 

P system with active membranes. Each rule has associated a probabilistic 

constant which depends on the left-hand side of the rule and the runtime. 

3 - Optimization Algorithms for the Protein Structure Prediction 

Problem 

Mario Pavone, Department of Mathematics and Computer 

Science, University of Catania, v.le A. Doria 6, 95125, Catania, 

Italy, mpavone@dmi.unict.it, Giuseppe Nicosia, Vincenzo 

Cutello 

Discrete models for protein structure prediction embed the protein amino acid 

sequence into a discrete spatial structure where an optimal tertiary structure 

is predicted on the basis of simple assumptions relating to the hydrophobichydrophilic 

character of amino acids in the sequence and to relevant interactions 

for free energy minimization. A quick, state-of-the-art survey of discrete 

models and evolutionary algorithms for protein structure prediction is 

presented, and the main design and performance features of an immunological 

algorithm are illustrated in a tutorial fashion. 

4 - Bio-Inspired Reverse Engineering Methodologies to Infer 

the Gene Regulatory Networks 

Mario Pavone, Department of Mathematics and Computer 

Science, University of Catania, v.le A. Doria 6, 95125, Catania, 

Italy, mpavone@dmi.unict.it, Natalio Krasnogor 

The purpose of Reverse Engineering is to elicit the internal structures of a given 

system from external observations. RE plays a central role in systems biology 

since is important also understand the dynamics of the genes and proteins. 

Identifying genes and proteins is not enough to understand the complexity of 

the system but need to know how these objects interact each other. Classical 

methods have been used to infer the genes interaction whose drawback is to 

map genes in a binary state. RE based on bio-inspired algorithms is able to 

effectively reconstruct the networks in dynamic models. 


� MF-25 


6.2.48 

Financial Risk Management 






Chair: Apostolos Fertis, IFOR, D-MATH, ETH Zurich, 101 

Rämistrasse, HG G 22.1, 8092, Zürich, Switzerland, 

afertis@ifor.math.ethz.ch 

1 - Robust Risk Management 

Apostolos Fertis, IFOR, D-MATH, ETH Zurich, 101 

Rämistrasse, HG G 22.1, 8092, Zürich, Switzerland, 

afertis@ifor.math.ethz.ch 

Often, the distribution of the returns of financial assets is given in two levels. 

The first level comprises of the non-accurate distribution over a set of scenarios, 

while the second level contains the precise distributions under the scenarios. 

We construct the class of robust risk measures,that tackles this uncertainty by 

applying robust optimization in the first level and a traditional risk in the second 

level. We define Robust CVaR as the robust risk corresponding to CVaR and 

design a way to compute it. We compare optimal-RCVaR and optimal-CVaR 

portfolios in real NYSE and NASDAQ data. 

2 - Australian Electricity Market and Price Volatility 

Ghazaleh Mohammadian, School of Mathematics and Statistics, 

University of South Australia, School of Mathematics and 

Statistics, Mawson Lakes Campus, 5095, Adelaide, South 

Australia, Australia, ghazalehmohammadian@yahoo.com, Jerzy 

Filar, Asef Nazari 

Australian Electricity Market has experienced high price volatility since the 

deregulation in early 1990’s. Considering variance as a measure of price risk, 

we demonstrate that this risk is increasing dramatically in high demand periods. 

We consider two-sided risk in terms of prices exceeding, or falling below, 

the mean. Closer examination shows that bidding behaviour of generators increases 

risk of volatility in general and the left-sided risk more than right-sided 

risk. Finally we suggest a new bidding strategy for generators which will results 

in alleviating this price volatility. 

3 - Forecasting Stock Index Realized Volatility with an 

Asymmetric HAR-FIGARCH Model: The Case of S&P 

500 and DJIA Stock Indices 

Dimitrios Louzis, Management Science and Technology, Athens 

University of Economics and Business, 3, Amerikis Str, 10564, 

Athens, Greece, dlouzis@aueb.gr, Spyros Xanthopoulos, 

Apostolos Refenes 

The Heterogeneous Autoregressive (HAR) model proposed by Corsi is extended 

in order to account for leverage effects in the realized volatility process 

and the long memory of the conditional variance of the HAR residuals. Estimation 

results reveal a heterogeneous component structure in asymmetric effects 

and a significant long memory property in the "volatility of realized volatility’. 

Compared with established HAR and ARFIMA realized volatility models, the 

proposed model exhibits superior out-of-sample volatility and Value at Risk 

forecasting performance. 

4 - Dual representation of choice and aspirational preferences 

David Brown, Duke University, United States, 

dbbrown@duke.edu, Melvyn Sim, Enrico De Giorgi 

We study a general class of preferences that favor diversification, except perhaps 

on a subset of sufficiently disliked acts, where concentration is preferred. 

This structure encompasses a number of known models. We show that such 

preferences can be expressed in dual form in terms of a family of risk measures 

and a target function. One special case that we explore in detail is that of a 

bounded target function. This case corresponds to a type of satisficing and has 

descriptive relevance. The model is amenable to large-scale optimization. 

99


� MF-26 


3.1.11 

Neural Network Applications 



Chair: Sureyya Ozogur-Akyuz, Department of Mathematics and 

Computer Science, Bahcesehir University, Bahcesehir University, 

Dept of Mathematics and Computer Science, Cıragan cad. Besiktas, 

34353, Istanbul, Turkey, sureyya.akyuz@bahcesehir.edu.tr 

Chair: Gurkan Ozturk, Industrial Engineering, Anadolu University, 


Eskisehir, Turkey, gurkan.o@anadolu.edu.tr 

1 - Performances of the ANN, SVM, and K-means clustering 

methods for recognizing different environmental 

sounds 

Inci Saricicek, Industrial Engineering Dept., Eskisehir 

Osmangazi University, ESOGU, Bati Meselik, 26480, Eskisehir, 

Turkey, incid@ogu.edu.tr, Buket Barkana, Burak Uzkent 

Automatic environmental sound classification has become a very active subject 

of research during last decades. This study investigates the categorization 

of environmental sounds using Support Vector Machine, K-means clustering 

and Neural Networks. With this study, we present the performances of three 

methods for classifying three commonly encountered sounds: Rain, engine and 

restaurant 

2 - Optimization of the automatic manipulator trajectory 

using stochastic algorithm and neural networks 

Alina Fedossova, Mathematics, Universidad Nacional, Cll 159 # 

56-75, t 4 ap 403, 12345, Bogota, Cundinamarca, Colombia, 

afedosova@unal.edu.co 

We present the stochastic outer approximations algorithm for robot trajectory 

planning. Initially, it is an automatic manipulator of three degrees of freedom. 

The proposed problem consists of finding total optimal time of displacements 

which must adjust to the trajectory using cubic splines constrained by the speed, 

acceleration and jerk. It’s a problem of semi-infinite optimization. 

After we find the minimum travel time of robot using neural networks that are 

trained on the results of the previous algorithm. For this purpose we use three 

types of neural networks in MATLAB. 

3 - Paraconsistent Artificial Neural Network — PANN and 

EEG Analysis 

Jair Minoro Abe, Paulista University, Rua Dr Bacelar 1212, 

04026-002, Sao Paulo, SP, Brazil, jairabe@uol.com.br, Fábio 

Romeu de Carvalho 

Abstract: Since the first works on ANN, several theories have been developed. 

Recently, we have introduced a new ANN, namely the Paraconsistent Artificial 

Neural Network — PANN, based on Paraconsistent Annotated Evidential 

Logic Et. Such logic (and so PANN) can manipulate concepts like impreciseness, 

inconsistency, and paracompleteness in its interior without trivialization. 

Some applications in pattern recognition, specifically in EEG analysis are discussed. 

4 - Estimating Correlated Constraint Boundaries from 

timeseries data: The multi-dimensional German Tank 

Problem 

Abhilasha Aswal, Infosys Technologies Limited, Bangalore, 

India, #44 Electronics City, Hosur Road, 560100, Bangalore, 

Karnataka, India, abhilasha.aswal@iiitb.ac.in, G. N. Srinivasa 

Prasanna, Sheela Siddappa 

We present a multidimensional generalization of the 1-D German Tank problem. 

N-D correlated data samples are drawn from a uniform distance, in a 

region bounded by linear constraints. These constraints have to be estimated 

minimizing variance without bias (UMVU estimator). Our new UMVU uses 

k-means clustering of the facets of the convex hull of data. Convergence time 

is polynomial in accuracy. We apply this to estimating constraints in robust 

optimization approaches to supply chain mgmt. As opposed to general MSE, 

our parameters are structurally limited to forming a non-empty polytope. 

100 

� MF-28 



Stochastic Programming - General 

Methodology 



Chair: Martin Branda, Department of probability and mathematical 

statistics, Charles University in Prague, Ke Karlovu 3, 121 16, 

Prague, Czech Republic, martin.branda@seznam.cz 

1 - Line-search with variable-number sample size 

Natasa Krklec, Department of Mathematics and Informatics, 


Sad, Serbia, natasa.krklec@dmi.uns.ac.rs, Natasa Krejic, 

Katarina Vla 

We are dealing with simulation-based optimization. The objective function is 

assumed to be in the form of mathematical expectation and it is approximated 

by a sample mean. The sample size is updated in every iteration. Since analytical 

form of the objective function is unknown, the gradient is also unavailable. 

Therefore we are using derivative free approach, such as regression, to obtain 

the model and the search direction in every iteration. We are using line search 

techniques to determine the candidate for a next iteration. Convergence is analyzed 

and numerical results are discussed. 

2 - Reformulation of general chance constrained problems 

using the penalty functions 

Martin Branda, Department of probability and mathematical 

statistics, Charles University, Sokolovska 83, 18675, Prague, 

Czech Republic, martin.branda@seznam.cz 

We explore reformulation of nonlinear stochastic programs with several joint 

chance constraints by stochastic programs with suitably chosen penalty-type 

objectives. We show that the two problems are asymptotically equivalent. We 

discuss solving both problems using Monte-Carlo simulation techniques for the 

case when the set of feasible solution is finite which appears in bounded integer 

programs. The approach is applied to the financial optimization problem with 

Value at Risk constraint, transaction costs and integer allocations. 

3 - Some Recent Results in Stochastic Gradient Estimation 

Michael Fu, Smith School of Business, University of Maryland, 

Van Munching Hall, 20742, College Park, MD, United States, 

mfu@umd.edu 

We review some recent results in stochastic gradient estimation, where direct 

unbiased estimators are obtained for simulation models. The techniques considered 

are perturbation analysis, the likelihood ratio (score function) method, 

and weak derivatives. Applications to stochastic activity networks and in financial 

engineering are considered. In particular, performance measures involving 

indicator functions and quantiles are addressed. 

� MF-29 


8.2.11 

Boolean methods in system analysis, 

learning and circuit synthesis 



Chair: Tiziano Villa, Dipartimento d’Informatica, Universita’ di 

Verona, Ca’ Vignal, 2, Strada Le Grazie, 15, 37134, Verona, Italy, 

tiziano.villa@univr.it 

1 - Symbolic Model Checking of Boolean Models in Biology 

Arlindo Oliveira, INESC-ID/IST, R. Alves Redol 9, 1000, 

Lisbon, Portugal, aml@inesc-id.pt

Boolean models of gene regulatory networks can be used to analyze and characterize 

the behavior of these complex systems. The application of symbolic 

model checking techniques, developed for the analysis of state spaces of digital 

systems, enables researchers to answer important questions on the nature 

of the state spaces of gene regulatory networks, such as the robustness of the 

system to perturbations and the characteristics of the basins of attraction. I will 

describe the application of these techniques to the gene regulatory models of 

two biological networks in Yeast and the fruit fly. 

2 - Boolean Machine Learning as a Tool for Model Inference 

from Data 

Diego Liberati, Elettronica e Informazione, Politecnico, CNR 

IEIIT, INFN MiB, Politecnico di Milano, Italy, Piazza Leonardo 

da Vinci 32, 20133, Milano, Italy, liberati@elet.polimi.it 

Synthesis of Boolean functions can be applied to machine learning, to infer directly 

from data both the very salient variables, and their even nonlinear static 

interplay. More complex adaptive Bayesian networks just add ranking among 

such salient variables, useful in pathways discovery in systems biology from 

microarrays analysis. The analysis of the dynamic interplay among the salient 

variables thus identified is made easy by piecewise affine identification, providing 

a linearized model of the switching behaviour of the investigated plant 

within the hybrid systems paradigm. 

3 - An Efficient Heuristic Approach to Solve the Unate Covering 

Problem 

Fabrizio Ferrandi, Dipartimento d’Elettronica e Informazione, 

Politecnico di Milano, Milano, Italy, 20133, Milano, Italy, 

ferrandi@elet.polimi.it 

We describe a constructive heuristic algorithm for two-level logic minimization 

that combines advances in data structures (the use of Binary Decision Diagrams) 

with lagrangian relaxation. This technique allows an effective choice of 

the elements in the solution, as well as cost-related reductions of the problem 

and a good lower bound on the optimum. The results show that on a wide set 

of benchmark problems, the algorithm nearly always hits the optimum, and in 

most cases proves it to be such. On the problems whose optimum is actually 

unknown, the best known result is strongly improved. 

4 - Boolean functions in cryptography 

Enes Pasalic, FAMNIT, University of Primorska, Glagoljaska 8, 

6000, Koper, Slovenia, enespasalic@yahoo.se 

In this talk a broad range of applications of Boolean functions in cryptography 

is discussed. Boolean functions are basic cryptographic primitives in the 

design of symmetric cryptographic systems, such as stream and block ciphers, 

and many other cryptographic algorithms. Its design is closely related to certain 

combinatorial problems, among others to integer optimization methods and 

some problems related to graph theory. We briefly consider this connection, and 

in addition different state-of-the-art approaches in designing cryptographically 

"secure" Boolean functions are addressed. 

� MF-30 


8.2.13 

New issues in aggregation-disaggregation 

philosophies 


other Issues 


Chair: Evangelos Grigoroudis, Decision Support Systems Laboratory, 

Technical University of Crete, University Campus, Kounoupidiana, 

73100, Chania, Greece, vangelis@ergasya.tuc.gr 

Chair: Yannis Siskos, Department of Informatics, University of 

Piraeus, Karaoli Dimitriou 80, 18534, Piraeus, Greece, 

ysiskos@unipi.gr 

1 - A DSS for Robustness Analysis in Multicriteria Satisfaction 

Evaluation 

Nikos Tsotsolas, Department of Statistics and Insurance Science, 

University of Piraeus, 2, Feidiou Str, 15236, Penteli, Greece, 

ntsotsol@unipi.gr, Yannis Siskos 


MUSA methodology, as an MCDA disaggregation approach to customer satisfaction 

evaluation problems, encompasses robustness concern, a necessary step 

towards the inference of robust satisfaction functions which shall be optimally 

consistent with expressed customers’ judgements. This paper presents a DSS 

which aims in providing substantial help to MUSA users, in three levels: (1) 

in benchmarking different robustness procedures (analytic and heuristic ones) 

for a given set of data, (2) in selecting the most appropriate parameters of the 

method and (3) in producing the full range of MUSA results. 

2 - Extensions of the MUSA method: Modeling additional 

preferences 

Evangelos Grigoroudis, Decision Support Systems Laboratory, 

Technical University of Crete, University Campus, 

Kounoupidiana, 73100, Chania, Greece, 

vangelis@ergasya.tuc.gr, Yannis Siskos 

The MUSA method is a preference disaggregation approach following the main 

principles of ordinal regression analysis. This study presents several extensions 

of the method, which include additional DMs’ preferences or desired properties 

of the inferred preference system. For example, additional preferences about 

the importance of the criteria are presented and additional constraints regarding 

special properties of the assessed model variables are discussed. Finally, this 

study presents alternative objective functions during the post-optimality analysis 

process of the MUSA method. 

3 - Multi-attribute Utility Theory without Preference Independence 

— Conception of a real multiplicative Preference 

Elicitation and Aggregation Method. 

Johannes Siebert, Business Administration, University of 

Bayreuth, Universitätsstr. 30, 95440, Bayreuth, Bayern, 

Germany, Johannes.Siebert@uni-bayreuth.de 

This paper introduces a new model called aggregate utility coefficient model 

(AUCM) that extends the MAUT that dependent preferences can be accounted 

for. Decision makers articulate their preferences relatively to an average of an 

alternative by means of interaction coefficients. The dimensional-specific preferences 

are multiplied to yield utility level on which the final decision is based. 

The terms obtained by a second or higher order Taylor expansion of the product 

at the average preference level are economically interpreted and used to model 

preferences and their interdependencies. 

� MF-32 


8.2.17 

OR in Animal Production 



Chair: Anders Ringgaard Kristensen, Department of Large Animal 

Sciences, University of Copenhagen, Faculty of Life Sciences, 

Groennegaardsvej 2, Frederiksberg C, DK-1870, Copenhagen, 

Denmark, ark@life.ku.dk 

1 - Optimization of heifer rearing strategy in Polish beef 

herds 

Anna Stygar, Department of Cattle Breeding, University of 

Agriculture in Krakow, al. Mickiewicza 24/28, 30-059, Crakow, 

Poland, astygar@ar.krakow.pl, Anders Ringgaard Kristensen, 

Joanna M. Makulska 

Beef cow lifetime productivity, measured by the number and weight of the 

calves weaned in the whole reproductive period, is considerably affected by 

heifer rearing strategy. The key elements of this strategy are the decisions on 

heifer growth rate and the moment of her first breeding. The objective of the 

study was to apply the multi-level hierarchic Markov process to determine the 

economically optimal rearing strategy for Polish beef heifers. The optimized 

decisions concerned the intensity of heifer feeding, the age, body weight and 

season at her breeding and the time for replacement. 

2 - A multistage stochastic programming model for planning 

piglet production 

LluisM Pla, Mathematics, University of Lleida, JaumeII,73, 

25001, Lleida, Spain, lmpla@matematica.udl.es, Sara Verónica 

Rodríguez-Sánchez, Victor Albornoz 

101


The planning of piglet production is central into modern pig production in connection 

with supply chain coordination. The complexity of piglet production 

is mainly affected by the stochastic nature of biological processes involved, 

highly variability on markets and rational scheduling of tasks on farm. In this 

way, the farmer has to make decisions today at different levels of uncertainty 

that are affecting future performances of the farm and chain. Hence a preliminary 

multistage stochastic programming model with recourse for planning 

piglet production under finite time horizon is presented. 

3 - Optimization of reproduction and production cycles for 

high yielding dairy cows 

Joanna M. Makulska, Department of Cattle Breeding, University 

of Agriculture in Krakow, al. Mickiewicza 24/28, 30-059, 

Kraków, Poland, rzmakuls@cyf-kr.edu.pl, Andrzej W˛eglarz, 

Anna Stygar, Barbara Tombarkiewicz 

The aim of the study was to use a multi-level hierarchical Markov decision 

process for the optimal management of reproduction and production cycles of 

dairy cows. The parameters of the developed model were estimated using the 

data records from the specific herd of high yielding cows in Poland. The optimized 

decisions concerned the length of days open and days dry periods in 

the consecutive calving intervals of a cow and the time of her replacement with 

a heifer. The criterion of optimality was assumed as the maximization of the 

total present value of the expected net returns per cow place. 

4 - Replacement policies for dairy cows 

Lars Relund Nielsen, Centre for Operations Research 

Applications in Logistics (CORAL), Aarhus School of Business, 

University of Aarhus, Fuglesangs Allé 4, 8210, Aarhus V, 

Denmark, lars@relund.dk, Anders Ringgaard Kristensen, Erik 

Jørgensen 

In a recent paper a hierarchical Markov decision processes (MDP) with finite 

state and action space was formulated for the dairy cow replacement problem 

with stage lengths of 1 d. Bayesian updating was used to predict the performance 

of each cow in the herd and economic decisions were based on the prediction. 

The model can be used to assist the farmer in replacement decisions on 

a daily basis and is based on daily milk yield measurements that are available 

in modern milking systems. This talk will present the results of the paper and 

discuss directions for further research. 

� MF-33 


8.2.19 

Environmental Management II 



Chair: Christos Ioakimidis, MECHANICAL ENGINEERING, 

INSTITUTO SUPERIOR TECNICO (IST) - UNIVERSIDAD DE 

LISBOA (UTL), Av. Rovisco Pais, 1049-001, LISBON, Portugal, 

christos.ioakimidis@dem.ist.utl.pt 

1 - The effect of environmental uncertainty on operations 

priorities and business performance 

Ahmed Attia, business management, alexandria university, 

elgomohoria st., 000, damanhour, elbehera, Egypt, 

ahmed-attia2005@hotmail.com 

Operations strategy is a one of the Functional strategies which help the company 

to deal with the environmental uncertainty,the operations strategy could 

be translated into operations priorities (cost,Flexibility,quality,delivery). The 

current research investigated the effect of environmental uncertainty on operations 

priorities,then the research investigated the effect of operations priorities 

on buisness performance. The current research data was collected from egyptian 

Pharmaceutical industry 

2 - Integrating Environmental Considerations into the Supplier 

Selection Criteria 

102 

Ceyda ¸Sen, Yildiz Technical University, Turkey, 

cgungor@yildiz.edu.tr, Mercan Ozturk, Özge Nalan Alp 

In the traditional supplier selection process, standart criteria like cost, quality 

and delivery performance are considered. However, growing environmental 

concerns have led the purchasing professionals to develop green purchasing 

strategies, and to rethink the supplier selection process which has neglected environmental 

impacts. To facilitate the process of the green supplier selection, 

this paper presents an extensive literature review on environmental considerations, 

and a hierarchical criteria structure that integrates these factors into the 

traditional supplier selection criteria. 

3 - Airline Service Network Design for Economic and Ecologic 

Sustainability 

Amir José Daou Pulido, Karlsruhe Institute of Technology, 

Klosterweg 28, Zimmer b16, 76131, Karlsruhe, 

Baden-Württemberg, Germany, amir.pulido@student.kit.edu, 

Cornelia Schoen 

In an effort to help reduce CO2 emissions and mitigate the climate impacts, the 

EU recently announced that as from 2012 the aviation sector will be included in 

the EU Emissions Trading Scheme. In particular, all flights starting and landing 

in the EU will be subject to a cap on their emissions. In preparation of this regulation, 

airlines are spurred to explore how they can improve their operations 

for economic and ecologic sustainability. We present an optimization model for 

airline service network design that integrates profit and environmental goals to 

comply with the new regulation 

4 - Enhancing electric energy management decisions with 

an Energy box 

Luis Oliveira, Mechanical Eng., Instituto Superior Técnico, Rua 

Sarmento Beires n o 15 5 o direito, 1900-410, Lisboa, Lisboa, 

Portugal, luis.jose.oliveira@gmail.com, Christos Ioakimidis, 

Paulo Ferrao 

This study presents a future where the consumer has an Energy box, a 24/7 

processor operating on a PC, to manage the electrical energy at his residence 

or small business, integrated in a microgrid. The considered model is a house 

with microrenewables production, air conditioner, and an electric vehicle that 

provides storage capacity. We propose an algorithm for the energy box with 

forecasted parameters (weather, electricity prices) and deterministic parameters 

(house’s loads, temperatures) as input, to obtain increases in the renewable 

energies penetration, power quality, and cost reduction 

� MF-34 


8.2.23 


Applications of Lot Sizing and Scheduling II 


Quantity 


Chair: Reinaldo Morabito, Dept. of Production Engineering, Federal 

University of São Carlos, CP 676, 13565-905, São Carlos, Sao Paulo, 

Brazil, morabito@ufscar.br 

1 - A production scheduling problem on an aerospace assembly 

fixture 

Horacio Yanasse, LAC, INPE, Av. dos Astronautas 1758, CP 515 

- INPE/CTE, 12227-010, São José dos Campos, SP, Brazil, 

horacio@lac.inpe.br, Bruno Silva, Reinaldo Morabito 

In this work we consider a production scheduling problem in an aerospace industry 

where workstations are located side by side in an assembly fixture. Depending 

on the jobs, adjacent workstations cannot process them at the same 

time due to space limitations. Moreover, jobs have precedence constraints, due 

dates, arrival times and technological constraints. The objective is to minimize 

the makespan. The problem is formulated as an MIP and solved by CPLEX. 

The solution obtained can improve significantly the production schedule when 

compared with the one used in practice. 

2 - A hybrid general lot-sizing and scheduling formulation 

for a production process with a two-stage product 

structure 

Sandra Transchel, Department of Supply Chain and Information 

Systems, Penn State University, 461 Business Building, 16802, 

University Park, PA, United States, sxt37@psu.edu

The paper is the result of a project with a chemical company. We present a 

hybrid mixed-binary optimization model based on the General Lot-sizing and 

Scheduling Problem (GLSP) for a production process characterized by a twostage 

product structure. Several company-specific requirements were considered 

which are not found in many classic lot-sizing models. We present two 

alternative reformulations based on the simple-plant-location problem. An extensive 

computational experiment based on real industry data is presented. 

3 - Apply local branching to solve lot sizing problems 

Renato Paiva, ICMC - USP, 13560970, São Carlos, São Paulo, 

Brazil, renatodepaiva@gmail.com, Franklina Toledo 

This work focuses on the capacitated lot-sizing problem (CLSP), which is one 

of the central tasks involved in production planning. The CLSP studied the 

following characteristics: multiple items; set-up times; capacity constraints; 

backlogging; carry over and overlapping. This work has two main goals: a) to 

study the influence of the solution of the CLSP when you consider the possibilities 

described above; b) to apply local branching to solve these problems. 

To solve the proposed instances, we used CPLEX 11 and a local branching 

approach proposed here. Good results were obtained. 

� MF-35 


6.2.46 

Facilitated Decision Analysis I 

Stream: Facilitated Modelling in OR 


Chair: Gilberto Montibeller, Dept. of Management (OR Group), 

London School of Economics, Houghton Street, WC2A 2AE, 

London, United Kingdom, g.montibeller@lse.ac.uk 

1 - Structuring Multicriteria Approaches to Developing 

Country Problems 

Theodor Stewart, Statistical Sciences, University of Cape Town, 

Upper Campus, 7701, Rondebosch, South Africa, 

theodor.stewart@uct.ac.za 

We review the application of causal mapping approaches to eliciting values 

from various groups in diverse developmental problem settings such as fisheries 

rights allocation, management of informal shops in informal settlement 

areas, and the establishment of food banking. In each case, the resulting structure 

was expressed in formal multicriteria terms for purposes of evaluataing 

future courses of action. Stakeholders in the workshops came from a variety of 

backgrounds, some with very little formal educational background. We shall 

discuss potential adaptations of the approach to link in with systems dynamics 

models for multicriteria policy evaluation. 

2 - A multi-stage, multicriteria framework to support facilitation, 

problem structuring and choice in outsourcing 

and supplier selection 

Miloslawa Fink, Département d’Informatique, Université de 

Fribourg, Bd de Pérolles 90, C310, 1700, Fribourg, Switzerland, 

miloslawa.fink@unifr.ch, Valerie Belton, Marino Widmer 

Decisions about outsourcing and supplier selection can significantly impact on 

organisational performance. Reported applications of MCDA in this context 

frame the decision as a single stage choice; however, in reality the situation can 

be much more complex, comprising several stages and multiple considerations. 

This paper presents a framework, initially derived from the literature and subsequently 

tested with practitioners, which makes explicit and helps to structure 

the different stages of the decision process. A case study to explore the use of 

the framework in practice is also presented. 

3 - Building Decision Support Systems with Facilitated 

Modelling 

Gilberto Montibeller, Dept. of Management (OR Group), 


London, United Kingdom, g.montibeller@lse.ac.uk, L. Alberto 

Franco, Victor Del Rio Vilas 

Most applications of facilitated modelling have been developed in the context 

of decision conferencing, to support a one-off decision. This type of approach 

is unsuitable for supporting repetitive decisions or those which involve a large 

set of options. In this paper we suggest an alternative use of facilitated modelling: 

to build up decision support systems (DSSs) for recurring decisions. 

This approach is illustrated by a real-world use, in which a DSS is being developed 

to aid a multi-criteria assessment of threats and vulnerabilities associated 

with animal health issues for Defra in the UK. 


4 - Facilitated Logistics Analysis: A decision support tool 

for locating logistic facilities with optimisation and 

multi-criteria value analysis 

Hugo Yoshizaki, Engenharia de Produção, Universidade de São 

Paulo, Av. Prof. Almeida Prado Trav 2 n.128, Cidade 

Universitária, 05508-900, São Paulo, SP, Brazil, hugo@usp.br, 

Gilberto Montibeller 

Locating logistic facilities, such as plants and distribution centres, in an optimal 

way, is a crucial decision for manufacturers. We suggest a multi-criteria framework 

to analyse such problems, which combines the value from the topology of 

a network (such as total cost or resilience) with the value of its discrete nodes 

(such as specific benefits of a particular location). In this framework the focus 

is on optimising the overall logistic value of the network. A decision support 

tool was developed to facilitate the decision process. An illustrative case will 

be presented. 

� MF-36 


3.1.05 

Fuzzy Optimization and Decision Analysis 1 


Intelligence 


Chair: Jaroslav Ramik, Dept. of Math. Methods in Economics, 

Silesian University, School of Business, University Sq. 1934/3, 

73340, Karvina, Czech Republic, ramik@opf.slu.cz 

Chair: Josef Vícha, Mathematical Institute in Opava, Silesian 

University in Opava, Na rybničku 1, 74601, Opava, Czech Republic, 

Josef.Vicha@math.slu.cz 

1 - Multiple Criteria Academic Staff Evaluation Models 

Jana Talasova, Dept. of Mathematical Analysis and Applications 

of Mathematics, Faculty of Science, Palacky University 

Olomouc, tr. 17. listopadu 1192/12, 771 46 , Olomouc, Czech 

Republic, talasova@inf.upol.cz, Jan Stoklasa 

Various academic staff evaluation models were subjected to detailed analysis. 

The routine use of weighted mean as a sole aggregation operator proved inappropriate 

for aggregating evaluations from different academic areas (lecturing, 

R&D, management). Even more general aggregation operators (OWA, 

WOWA) still leave some room for improvement. To objectively assess benefit 

of an individual staff member, the use of a fuzzy rule base in aggregating partial 

evaluations proves optimal. Our proposed linguistic fuzzy evaluation model is 

currently being implemented at Palacký University Olomouc. 

2 - Fuzzy Models of Decision Making as a Research Tool in 

Architecture 

Zuzana Talasova, Cabinet of Architectural Modelling, Faculty of 

Architecture, Czech Technical University in Prague, 

Jugoslavskych partyzanu 3, Praha 6, 16000, Prague, Czech 

Republic, zuza.talasova@post.cz 

This paper explores possibilities of using linguistic fuzzy models as a new research 

tool in the history of architecture. Namely, Adolf Loos’s villas are studied 

applying the models. The aim is to create a linguistic fuzzy model that 

would characterize the designing process of a "Loosian’ villa. Rules applied 

by A. Loos in villa design are expressed in the form of fuzzy rule bases. The 

formulation of rules results from the analysis of individual villa designs and is 

based on theoretical concepts pertaining to Loos’s work - in particular, on his 

concept of a Raumplan (space plan). 

3 - Fuzzified Choquet Integral in Multiple Criteria Evaluation 

Models 

Iveta Bebcakova, Department of mathematical analysis and 

mathematical application, Palacky University Olomouc, 

Olomouc, Czech Republic, bebcakova.iveta@post.cz, Jana 

Talasova 



Within evaluation models, the integrand of the discrete Choquet integral (CI) 

represents partial evaluations of an alternative with respect to given criteria, 

while the fuzzy measure, i.e. the generalized monotonic measure, stands for 

the weights of the sets of criteria. We focus on fuzzified CI. We study the firstlevel 

fuzzified CI that handles partial fuzzy evaluations. Then we employ the 

second-level fuzzified CI where also the weights of the sets of criteria are in 

the form of fuzzy numbers. The user-friendly SW pertaining to this framework 

will be presented. 

4 - Dual Optimization Problems with Generalized Relational 

Inequality Constraints 

Karel Zimmermann, Faculty of Mathematics and Physics, 

Applied Mathematics, Charles University, 11800, Prague, Czech 

Republic, zimmermann@seznam.cz, Martin Gavalec 

Relational inequalities and equations on fuzzy algebras, which were introduced 

and applied to some medical problems [E. Sanchez] are generalized. Dual pairs 

of optimization problems with generalized relational inequality constraints are 

studied. It is proved that dual pairs posses both weak and strong duality property 

from the duality theory. Application of the results to finding saddle points 

of matrices and their relation to some games on fuzzy algebras is discussed. 

Perspectives of further research are outlined. 

� MF-37 


3.1.09 

OR for Development and Developing 

Countries II 









1 - Recent Trend of World Productivity Growth: A Stochastic 

Frontier Application 

Sangho Kim, Departmernt of International Trade, Honam 

University, 506-714, Gwangju, Korea, Republic Of, 

shkim@honam.ac.kr 

Based on Cuesta (2000), we develop a stochastic frontier production model 

that allows for different groups of countries to have patterns of technical efficiency. 

We apply our model to Penn World Tables’s 53 countries (2000-2003) 

to decompose total factor productivity growth into technical efficiency change 

and technical progress. Our results indicate that technical inefficiency is a serious 

impediment to higher productivity. They also identify varying paths for 

countries to derive their growth from a combination of components. 

2 - Export functions analysis. Czech Republic and EU. 

Šárka Krkosková, Department of econometrics, University of 

economics, Prague, Czech Republic, xeroxy@email.cz, Adéla 

Ráčková 

In this paper export functions were formulated and a case of the Czech Republic 

was studied. Domestic and foreign income represented by GDP, domestic 

and foreign prices, exchange rates, monetary aggregates, foreign direct investment, 

the Czech Republic‘s EU entry, neighbouring countries‘ EMU accession 

and start of global financial crises were used to define the functions. Models 

with combination of mentioned variables, different lags and time period were 

discussed. GDPs and exchange rates were the significant factors of the export 

function in most of the models. Prepared under IGA, F4/13/2010. 

3 - Development of Logistic Centers in Poland 

Iwona Otola, Management Faculty, Czestochowa University of 

Technology, ul. Armii Krajowej 19 B, 42- 200, Czestochowa, 

Poland, iotola@zim.pcz.pl 

The logistic centre is the most advanced form of the central unit of logistic 

networks. The logistic net rich in the linear and scoring infrastructure offering 

diverse alternatives of elastic configuring chains of supplies is an essential 

element of functioning of economic networks. The modern logistic centre 

is forced by the more and more competitive economy to apply modern technologies. 

Modern centers use computer systems such as expert systems and 

databases, which managing the object of this type isn’t possible without. 


4 - Supplier selection in supply ch ain management of accommodation 

enterprises ( a research in the hotels of 

turke y) 

Mehmet Sarioglan, Tourism, Social Science, Turkey, University 

Of Balikesir, Balikesir, University, Turkey, 

mehmets@balikesir.edu.tr, Cevdet Avcikurt, Murat Do ˘Gdubay 

In this study, the necessity of a research on supply chain mana gement in accommodation 

enterprises was detected in order to contribute to the literature 

and to determine the tendencies of supplier selection criteria in ac commodation 

enterprises. For this purpose, in this research the tendencies of s upplier 

selection criteria in supply chain management were tried to determine i n 73 

accommodation enterprises which operate in 26 cities of Turkey. Significan t 

differences and similarities between the criteria, which are used by accommod 

ation enterprises in the selection of supplier, and the size of accommodation e 

nterprises were determined in this study. 

� MF-39 


6.2.45 

Limit Behaviour and Approximations I 



Chair: Vladimir Veliov, Institute of Mathematical Methods in 

Economics, Vienna University of technology, ORCOS, 

Argentinierstr. 8/119, 1040, Vienna, Austria, veliov@tuwien.ac.at 

1 - Optimization Methods for Optimal Control Problems 

with Control and Pointwise State Constraints 

Ion Chryssoverghi, Mathematics, National Technical University 

of Athens, Agiou Meletiou 93, 11251, Athens, Attica, Greece, 

ichris@central.ntua.gr, John Coletsos, Basil Kokkinis 

We consider an optimal control problem for systems defined by nonlinear ordinary 

differential equations, with control and pointwise state constraints. The 

problem is formulated in the classical and in the relaxed form. Various necessary 

conditions for optimality are first given for both formulations. For the 

numerical solution of these problems, we then propose a penalized gradient 

projection method generating classical controls, and a penalized conditional 

descent method generating relaxed controls. Using relaxation theory, we study 

the behavior in the limit of sequences constructed by these methods. Finally, 

numerical examples are given. 

2 - High-Order Approximations in Optimal Control 

Vladimir Veliov, Institute of Mathematical Methods in 


Argentinierstr. 8/119, 1040, Vienna, Austria, 

veliov@tuwien.ac.at 

The direct numerical methods for solving optimal control problems require: (i) 

approximation of the admissible controls by a finitely parameterized controls; 

(ii) discretization of the underlying differential equation. The second issue is 

well developed, while the error analysis in the first issue is still in a challenging 

issue. The talk is devoted to new ideas and results related to the last issue. In 

particular, we investigate the relation between the information pattern of the 

approximations in (i) and the accuracy of approximation, focusing on higher 

than first order approximations. 

3 - Euler Approximation of Linear-Quadratic Control Problems 

with Bang-Bang Solutions 

Walter Alt, Applied Mathematics, Friedrich-Schiller-Universitaet 

Jena, Ernst-Abbe-Platz 2, 07743, Jena, Germany, 

walter.alt@uni-jena.de, Robert Baier, Matthias Gerdts 

We investigate Euler discretizations for a class of optimal control problems 

with control appearing linearly. First we show first order convergence for the 

optimal values. Under the additional assumption that the optimal control has 

bang-bang structure we also derive error estimates for the controls.

� MF-40 


6.2.52 

Telecommunications 

Stream: Network Optimization [c] 


Chair: Andras Farago, Computer Science, University of Texas at 

Dallas, 800 W. Campbell Rd., 75080, Richardson, Texas, United 

States, farago@utdallas.edu 

1 - Suboptimal solutions to team optimization problems 

with statistical information structure 

Marcello Sanguineti, DIST, University of Genoa, Via Opera Pia, 

13, 16145, Genova, Italy, marcello@dist.unige.it, Mauro 

Gaggero, Giorgio Gnecco 

Network team optimization problems with statistical information structure are 

investigated. A team of n Decision Makers (DMs), each having at disposal 

some information (obtained, e.g., by measurement devices or by exit polls) and 

various possibilities of decisions, coordinate their efforts to achieve a common 

goal, expressed via a team utility function. Decisions are generated by the DMs 

via strategies, on the basis of the available information y1, . . . , yn that each of 

them has and in the presence of uncertainties in the "state of the external world’ 

x (which the DMs do not control). Such uncertainties are modeled via a joint 

probability density p(x, y1, . . . , yn). For these problems, optimal solutions in 

closed form can be derived only in special cases, so a methodology of approximate 

solution is proposed. Suboptimal solutions are searched for, taking the 

form of linear combinations of elements from sets of basis functions, possibly 

with adjustable "inner’ parameters. Upper bounds on the accuracy of such 

suboptimal solutions are obtained. The estimates are expressed in dependence 

of the number of trigonometric and Gaussian basis functions. The trade-off 

between the level of decentralization and the smoothness assumptions on the 

utility function and the probability density, required to derive the upper bounds, 

is investigated. Numerical results are presented for an instance of the network 

team optimization problem under study, which models optimal production in a 

multidivisional firm. 

2 - On call admission control with nonlinearly constrained 

feasibility regions 

Giorgio Gnecco, DIST, University of Genoa, Via Opera Pia, 13, 

16145, Genova, Italy, giorgio.gnecco@dist.unige.it, Marco 

Cello, Mario Marchese, Marcello Sanguineti 

A simple criterion is proposed to improve suboptimal coordinate-convex policies 

in Call Admission Control problems with nonlinearly constrained feasibility 

regions. To test the criterion, numerical simulation results are given. 

Finally, some structural properties of the optimal coordinate-convex policies 

are proven, which do not depend on a complete knowledge of the nonlinear 

boundary of the feasibility region. 

3 - Evaluating downside risks in reliable networks 

Megha Sharma, Production & Quantitative Methods, Indian 

Institute of Management Ahmedabad, D-230 IIM Ahmedabad, 

Vastrapur, 380015, Ahmedabad, Gujarat, India, 

meghas@iimahd.ernet.in, Diptesh Ghosh 

Many real world operations require networks spread over large geographical 

areas. Elements of such networks are often prone to failure. Such networks 

are modeled as reliable networks, i.e., networks in which arcs and/or nodes can 

fail from time to time. If an operation is critical, it is useful to measure network 

performance when it is not performing well. In this paper, we propose 

two metrics to measure such downside performance, and develop a state-space 

enumeration based algorithm to evaluate the metrics. We report computational 

performance of our algorithm on randomly generated instances. 

4 - The Multigraph Advantage and Its Application in Network 

Analysis 

Andras Farago, Computer Science, University of Texas at Dallas, 

800 W. Campbell Rd., 75080, Richardson, Texas, United States, 

farago@utdallas.edu, Dung Trung Tran 


In a number of network models one has to distinguish different types of edges 

between the same nodes. This occurs quite often in communication networks. 

For example, in an optical network, the different wavelengths can define distinct 

logical links between the same nodes. Similar phenomenon occurs in radio 

networks, when more than one frequency bands are being used simultaneously. 

In such situations a multigraph, possibly with colored edges, is a more natural 

model of the network than a simple graph. 

Our goal is to investigate the gain that can be obtained in some important graph 

parameters via allowing parallel edges in the network model. We introduce the 

concept of the multigraph sum, in which several simple graphs on the same 

node set are merged to form a multigraph. We prove that when the component 

graphs are chosen randomly, then the edge-connectivity of the resulting 

multigraph has an unexpectedly high surplus over the sum of the original connectivities. 

Moreover, the surplus remains significant and asymptotically nonvanishing, 

even if it is measured not by its absolute size, but in relative terms, 

as a percentage gain. 

We call this phenomenon the "multigraph advantage", and show that it occurs 

with quite a few different graph parameters. Thus, we find several cases when 

"the whole is strictly more than the sum of its parts". Translating it back into the 

motivating communication network models, we conclude that it often pays off 

to implement multiple types of links in a network, and, moreover, the obtained 

gain can be quantified using our results. 

� MF-41 


3.1.06 

Pricing in networks 






Chair: Houyuan Jiang, Judge Business School, University of 

Cambridge, Trumpington Street, CB2 1AG, Cambridge, United 

Kingdom, h.jiang@jbs.cam.ac.uk 

1 - A column generation approach for a bilevel pricing 

problem 

Aurélie Casier, Départment d’Informatique, ULB - Université 

Libre de Bruxelles, 1050, Brussels, Belgium, acasier@ulb.ac.be, 

Martine Labbé, Bernard Fortz 

Consider the product pricing problem (PPP) in which a company sets prices 

for products in order to maximize its revenue and reacting to these prices the 

customers buy, among all products on the market, the one providing them the 

biggest utility. Initially modeled as a bilevel program, PPP can be reformulated 

as a single level nonlinear model. From this nonlinear formulation, we derive a 

new IP formulation containing an exponential number of variables and propose 

a column generation solution approach. 

2 - Optimisation of forecourt fuel pricing 

David McCaffrey, Research, KSS Ltd, St. James’s Buildings, 79 

Oxford Street, M1 6SS, Manchester, United Kingdom, 

mccaffreyd@kssg.com, Barbara Jenkins, Tom Liptrot 

We present an application of price optimisation to the sale of automotive fuel on 

the forecourt. We present a detailed analysis of the results of a trial of price optimisation 

software at a US fuel retailer. We demonstrate an increase in gross 

profitability of around 0.6 US cents per gallon, obtained without significant 

loss of volume against target. We then set out an extension to the methodology 

which allows for price-volume trade off between retail sites. We show 

how this sits as a supervisory level above the site level optimisation and present 

simulation results. 

� MF-42 


3.1.07 

Optimization and Data Mining I 

Stream: Optimization and Data Mining 


Chair: Emilio Carrizosa, Estadistica e Investigacion Operativa, 

Universidad de Sevilla, Matematicas, Reina Mercedes s/n, 41012, 

Sevilla, Spain, Spain, ecarrizosa@us.es 



1 - Support Vector Machine for Time Series Regression 

Renato De Leone, Dipartimento di Matematica e Informatica, 

Università di Camerino, via Madonna delle Carceri 9, 62032, 

Camerino, MC, Italy, renato.deleone@unicam.it 

Support Vector Machines (SVMs) have been extensively used in classification 

and regression. In this talk we will show how SVMs can be used to predict 

spcific aggregated values from a time series. Applications in finance will also 

be discussed. 

2 - Application of svm in the classification of ship of 

acoustic signals using frequency and rpm data 

Mikey da Silva Neto, Departamento de Engenharia Aeronáutica 

e Mecânica, Instituto Tecnológico de Aeronáutica, Rua H9 B 

Apto 102, 12228-611, São José dos Campos, São Paulo, Brazil, 

mikey@ita.br, Moacyr Machado Cardoso Junior, Rodrigo 

Scarpel 

Ship class recognition by submarine sonar is very important, for navigation 

safety as well as in belligerence situations. The aim of this paper is to analyze 

sound frequency and propeller-axle revolutions per minute (RPM) gained 

through acoustic signal, through Lofargram-Demongram for classification purpose. 

Support vector machine (SVM) techniques were used in conjunction of 

grid search algorithm in order to extract kernel function optimal parameters. 

The results showed a good performance of ship classification using the features 

selected, as well as, a relative low computational cost. 

3 - An Online Monitoring Procedure Using Support Vector 

Regression 

Seong-Jun Kim, Industrial Eng., Gangneung-Wonju National 

University, 120 Univeristy Road Gangneung, 210702, 

Gangneung, KW, Korea, Republic Of, sjkim@gwnu.ac.kr, 

Inyong Seo, Sungho Choi 

Online monitoring (OLM) is becoming widespread in power plants. It is essential 

to secure both efficiency and safety in the plant operation. Many OLM 

systems make use of an auto-associative model in order to estimate operation 

parameters from transferred signals. This paper deals with model building for 

the auto-association. A support vector regression is employed for the modeling 

in this paper. The monitoring performance of the proposed OLM technique is 

compared with feedforward neural network. An illustration will be given by 

real-world example. 

� MF-43 


8.2.02 

Revenue Management I 

Stream: Demand, Pricing and Revenue Management 


Chair: Alf Kimms, Mercator School of Management, University of 

Duisburg-Essen, Lotharstr. 65, LB 125, 47057, Duisburg, Germany, 

alf.kimms@uni-due.de 

1 - A Dynamic Programming Decomposition for Revenue 

Management with Substitution 

Jochen Goensch, Department of Analytics & Optimization, 

University of Augsburg, Universitätsstraße 16, D-86159, 

Augsburg, Germany, jochen.goensch@wiwi.uni-augsburg.de, 

Claudius Steinhardt 

In markets with highly uncertain demand, supply-side substitution between resources 

(e.g. through opaque products) can help firms to improve capacity 

utilization and revenue. This talk presents an extension of a standard capacity 

control approach making use of the additional flexibility. We show that the underlying 

decomposition yields a tighter upper bound than the widely used DLP. 

In numerical experiments the approach is applied to airline network revenue 

management with opaque products. It shows that the approach is tractable and 

outperforms widely used bid price approaches. 

2 - Capacity Control with Planned Upgrades in the Car 

Rental Industry 


Claudius Steinhardt, Department of Analytics & Optimization, 

University of Augsburg, Universitätsstraße 16, 86159, Augsburg, 

Germany, claudius.steinhardt@wiwi.uni-augsburg.de, Jochen 

Goensch 

Even though upgrades are common in car rental, practical revenue management 

implementations usually resort to rather simple heuristics, like successive 

planning, and use virtual capacity as given for standard capacity control. In 

this talk, we present approaches that extend the traditional decomposition for 

capacity control by simultaneously considering upgrades as well as capacity 

control decisions. Based on data from a major car rental company, we perform 

computational experiments which show that the proposed approaches are 

tractable and outperform widely-used successive planning approaches. 

3 - Examples illustrating the importance of integrating revenue 

management and supply chain decision making. 

Peter Bell, Richard Ivey School of Business, University of 

Western Ontario, N6A 3K7, London, Ontario, Canada, 

pbell@ivey.ca 

We examine two firms with traditional supply chain issues (excess capacity, 

high transportation costs, and highly seasonal demand) and show that these issues 

can be profitably resolved using revenue management tools. We also show 

that management of these firms will likely make inappropriate strategic and tactical 

decisions unless their decision model integrates both revenue management 

and supply chain decision variables. 

4 - Nucleolus Based Revenue Sharing in Airline Alliances 

Alf Kimms, Mercator School of Management, University of 

Duisburg-Essen, Lotharstr. 65, LB 125, 47057, Duisburg, 

Germany, alf.kimms@uni-due.de, Demet Cetiner 

A major problem in airline alliance revenue management operations is to construct 

allocation rules which define how the alliance revenue should be shared 

among the airline partners. In this paper, we provide fair revenue proportions 

for airlines based on the nucleolus solution concept. Through a computational 

study on randomly generated alliance networks, we show that the nucleolus 

revenue proportions provide reasonable allocation rules for airline alliances and 

may serve as a benchmark for decentralized approaches. 

� MF-44 


8.2.03 

SD modelling of Scarcity and Sustainable 

Development 



Chair: Erik Pruyt, Policy Analysis, Delft University of Technology, 

Visitors Address: Jaffalaan 5, Delft, The Netherlands, 

Correspondence Address: P.O. Box 5015, 2600 GA , Delft, 

Zuid-Holland, Netherlands, E.Pruyt@tudelft.nl 

1 - Scarcity of Minerals and Metals: Making Sense of 

Generic and Specific Exploratory System Dynamics 

Models 

Erik Pruyt, Policy Analysis, Delft University of Technology, 

Visitors Address: Jaffalaan 5, Delft, The Netherlands, 

Correspondence Address: P.O. Box 5015, 2600 GA , Delft, 

Zuid-Holland, Netherlands, E.Pruyt@tudelft.nl 

Possible short, medium and/or long term scarcity of minerals/metals may actually 

pose a threat to modern societies. Its potentially disruptive societal consequences 

qualify this issue for exploration from a world/regional security point 

of view. An generic exploratory System Dynamics model that could be used to 

explore the dynamic complexity of potential mineral/metal scarcity under deep 

uncertainty is presented. The generic model is turned into specific models for 

answering particular questions. 

2 - Scarcity of Minerals and Metals: An Application of Exploratory 

Modelling and Analysis 

Caner Hamarat, Policy Analysis, Delft University of Technology, 

Jaffalaan 5, 2628 BX, Delft, Netherlands, c.hamarat@tudelft.nl, 

Erik Pruyt

Possible short, medium and/or long term scarcity of minerals/metals may actually 

pose a threat to modern societies. Its potentially disruptive societal consequences 

qualify this issue for exploration from a world/regional security point 

of view. Different exploratory SD models are used here as alternative scenario 

generators for Exploratory Modelling and Analysis to explore the dynamic 

complexity of potential scarcity under deep uncertainty. Tens of thousands of 

scenarios are analysed and used to determine the effectiveness and robustness 

of promising policies. 

3 - System Dynamics Modeling for the Urmia Saline Lake 

Management 

Mahdi Zarghami, Faculty of Civil Engineering, University of 

Tabriz, 29 Bahman Blvd., 51664, Tabriz, zarghaami@gmail.com, 

Elmira Hassanzadeh, Yousef Hassanzadeh 

About five millions of people are living in Urmia lake basin, Iran. Because of 

the intensive irrigation, the water resources are overused and then the regional 

water companies require the simulation of the lake future. This study is defined 

to partition the impact of climate change and hydrostructures by using the 

System Dynamics approach. The model includes water supply sources, evaporation, 

demand sources and management tools, modeled by Vensim software. 

Several scenarios have been studied, which provide sensitive results needed for 

the effective evaluation of the Urmia Lake problem. 

� MF-45 


8.2.12 

Recent developments from Nonconvex 

Programming 


Approaches 


Chair: Tao Pham Dinh, INSA Rouen, 76131, Rouen, France, 

pham@insa-rouen.fr 

1 - New DC Programming Approaches for BMI and QMI 

Feasibility Problems 

Yi-Shuai Niu, LMI, INSA de Rouen, BP08 - Avenue de 

l’Université, 76801, Rouen, France, niuyishuai@hotmail.com, 

Tao Pham Dinh 

We propose new DC programming approaches for solving BMI and QMI Feasibility 

Problems which are important in robust control. The inherent difficulty 

lies in the nonconvexity of the feasible set. We reformulate QMI as a DC 

program, and then using an efficient DC Algorithm (DCA) for the numerical 

solution. A hybrid method combining DCA with an adaptive B&B is proposed 

for guaranteeing the feasibility of QMI. A partial solution concept is proposed 

for reducing the computational time and for solving more large-scale problems. 

Some numerical simulations and comparison with PENBMI are also reported. 

2 - Siting and Sizing of Facilities under Probabilistic Demands 

Luís Fernandes, Matemática, IPT/IT, Estrada da Serra - Quinta 

do Contador, 2300 - 313, Tomar, Portugal, luism@ipt.pt, 

Joaquim Judice, Hanif Sherali, Antonio Antunes 

We describe a discrete location model for non-essential service facilities planning 

which seeks the number, location and size of facilities that maximizes the 

total expected demand attracted by the facilities. Demand for service is assumed 

sensitive to distance and size. Facilities are assumed to satisfy a threshold 

level of demand. An MINLP model is proposed and a branch-and-bound 

algorithm designed for solving the MINLP; convergence is established. Numerical 

results are reported to illustrate its efficacy and efficiency in practice. 

3 - A new approach for solving Value-at-Risk constrained 

Optimization using the DC programming and DC Algorithm 

(DCA) 

Manh Nguyen Duc, LMI, INSA de Rouen, France, Poussin 107, 

Cite du Bois, 76130, Rouen, Rouen, France, 

duc.nguyen@insa-rouen.fr, Hoai An Le Thi, Tao Pham Dinh 


In this paper, we will consider a well-known problem in the Markowitz style 

portfolio selection: maximize the rate return with Value-at-Risk constrain when 

the distributions of returns of the considered assets are given in the form of finite 

scenarios. The problem is initially reformulated as a DC (Difference of 

Convex functions) program. In the next step, the problem is then written in the 

form minimizing a polyhedral concave function on a convex polyhedron via a 

new penalty technique. Finally, DC programming and DCA (DC algorithm) 

have been investigated to solve the resulting DC program. 

� MF-46 


8.2.14 

Semi-Infinite Optimization II 

Stream: Semi-Infinite Optimization 


Chair: Shunsuke Hayashi, Graduate School of Informatics, Kyoto 

University, Yoshida-Honmachi, Sakyo-Ku, 606-8501, Kyoto, Japan, 

shunhaya@amp.i.kyoto-u.ac.jp 

1 - Polyhedral cells of a Voronoi diagram 

Ina Voigt, Fakultät für Mathematik, Technische Universität 

Dortmund, Vogelpothsweg 87, 44227, Dortmund, 

Nordrheinwestfalen, Germany, ina.voigt@tu-dortmund.de, 

Stephan Weis 

We identify a cell of a Voronoi diagram, i.e. a nearest neighbor region, with 

the feasible set of a semi-infinite system. Utilizing a theorem from the theory 

of semi-infinite programming, we investigate the geometry of a Voronoi cell. 

We prove that a Voronoi cell of an infinite discrete point set is polyhedral if and 

only if its corresponding characteristic cone is a polyhedron. This connects 

computational geometry with semi-infinite optimization. 

2 - Optimality conditions and regularized explicit exchange 

method for convex semi-infinite programs with 

infinitely many conic constraints 

Shunsuke Hayashi, Graduate School of Informatics, Kyoto 

University, Yoshida-Honmachi, Sakyo-Ku, 606-8501, Kyoto, 

Japan, shunhaya@amp.i.kyoto-u.ac.jp, Takayuki Okuno, Masao 

Fukushima 

Although the semi-infinite program is normally represented with infinitely 

many “inequality” constraints, we focus on a problem with infinitely many 

“conic” constraints, say SICP for short. We first show that, under a generalized 

Slater constraint qualification (GSCQ), an optimum of the SICP satisfies the 

KKT conditions that can be represented only with a finite subset of the conic 

constraints. We next propose a hybrid algorithm combining the regularization 

technique with the explicit exchange method, and show that the algorithm is 

globally convergent under the GSCQ. 

3 - Semi-infinite Optimization of Probe Placement in Radiofrequency 

Ablation 

Sabrina Haase, CeVis, University of Bremen, Universitaetsallee 

29, 28359, Bremen, Bremen, Germany, 

sabrina.haase@mevis.fraunhofer.de, Anton Winterfeld, Tobias 

Preusser, Karl-Heinz Küfer 

Radiofrequency ablation is a minimally invasive therapy for the treatment of 

tumors. A needle-shaped probe is placed in the tumor and connected to an 

electric generator. Thus the induced electric current causes tumor cell death 

due to the evolving heat. Since the whole tumor is supposed to be destroyed, 

an optimal probe placement is needed where the presence of structures that 

must not be harmed (blood vessels, colon) has to be taken into account. This 

task results in a semi-infinite optimization problem which is presented in the 

talk together with some first numerical results. 



� MF-47 


8.2.16 

Advances in Discrete-Continuous Optimal 

Control 1 

Stream: Discrete Optimal Control 





Chair: Dmitrii Lozovanu, Institute of Mathematics and Computer 

Science, Academy of Sciences of Moldova, Academiei 5, IMI, 

MD-2020, Chisinau, Moldova, Moldova, Republic Of, 

lozovanu@math.md 

1 - Regularity conditions in discrete optimal control problems 

Boban Marinkovic, Applied Mathematics, University of 

Belgrade, Faculty of Mining and Geology, Djusina 7, 11000, 

Belgrade, Serbia, mboban@verat.net 

We consider general optimal control problems for discrete-time systems with 

equality type of constraints on endpoints and control. We discuss different 

types of regularity conditions for such problems. We present first-order necessary 

optimality conditions which are meaningful under the new nontriviality 

condition. Note that by nontriviality we mean that the Lagrange multiplier that 

corresponds to the cost functional is nonzero. 

2 - Solving discretized degenerate optimal control problems 

with state constraints 

Catarina Avelino, Mathematics, UTAD, Departamento de 

Matemática, Edifício das Ciências Florestais, Quinta de Prados, 

5001-801, Vila Real, Vila Real, Portugal, cavelino@utad.pt 

We consider a class of nonlinear optimization problems that arise from the discretization 

of optimal control problems with bounds on both state and control 

variables. We are particularly interested in degenerate cases, i.e., when the 

linear independence constraint qualification is not satisfied. We analyse the 

basic global convergence properties and the numerical behaviour of a multiplier 

method that updates multipliers corresponding to inequality constraints. 

Numerical results are included and indicate that this method is robust on degenerate 

cases. 

� MF-48 


8.2.04 

OR/MS: Beyond Mathematics I 

Stream: OR/MS: Beyond Mathematics 


Chair: Heiner Müller-Merbach, Wirtschaftswissenschaften, 

Universität Kaiserslautern, Postfach 3049, 67653, Kaiserslautern, 

Germany, hmm@bior.de 

1 - A.K. Erlang — my distinguished relative 

Jakob Krarup, Dept. of Computer Science, University of 

Copenhagen, Ydervang 4, DK-3460, Birkeroed, Denmark, 

krarup@diku.dk 

Agner Krarup Erlang (1878-1929), Head of the laboratory (Copenhagen Telephone 

Company, 1908-29) and pioneer in the application of probability theory 

to problems of telephone traffic, is among the inductees in IFORS’s Operational 

Research Hall of Fame (Intl. Trans. in Op. Res. 11, 2004), and recognized as 

the first operational researcher in Denmark. A brief account of his works is 

provided ... together with fragments of his genealogical tree. 

2 - OR/MS — Beyond Mathematics 


Heiner Müller-Merbach, Wirtschaftswissenschaften, Universität 

Kaiserslautern, Postfach 3049, 67653, Kaiserslautern, Germany, 

hmm@bior.de 

Traditional OR/MS used to have its focus on interdisciplinary problem solving 

— far beyond mathematics. Promoters of this kind of OR/MS were pioneers 

such as C. West Churchman, Hugh Miser, Russell L. Ackoff, Samuel Eilon 

etc. "OR/MS — Beyond Mathematics’ has its emphasis on the understanding 

of man-machine systems with their technical (mechanical, thermodynamic, 

chemical, other physical) aspects and their social and economic aspects well as 

their ethical aspects. Most important in the interdisciplinary OR/MS approach 

are the steps to analyse the system under study. 

3 - Estimating Business and Management Journal Quality 

from the 2008 Research Assessment Exercise in the UK 

John Mingers, Kent Business School, Kent University, CT2 7PE, 

Canterbury, Kent, United Kingdom, j.mingers@kent.ac.uk, 

Maria Paola Scaparra 

The 2008 UK RAE peer reviewed over 10,000 journal articles. Each output 

was graded on a 5-point scale. These grades were accumulated for each department 

to provide an overall quality profile. This data provides the possibility 

of reconstructing the judgements made by the Panel. We have used linear programming 

to produce the best estimate of the grades awarded to papers from 

each journal that had more than three entries. This provides both a grade profile 

for each journal and a single quality estimate.

Tuesday, 9:00-10:20 

� TA-01 


Aula Magna 




Chair: Jose Fernando Oliveira, Faculty of Engineering / INESC 

Porto, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465, 

Porto, Portugal, jfo@fe.up.pt 

1 - You want them to Remember? Then Make it Memorable! 

James Cochran, Department of Marketing and Analysis, 

Louisiana Tech University, PO Box 10318, 71272, Ruston, LA, 

United States, jcochran@cab.latech.edu 

Each of us has key concepts we want our students to understand and remember, 

but lecturing to students on these concepts often fails to engender their 

deep comprehension or long term retention. So how can instructors of operations 

research/management science effectively accomplish these pedagogical 

goals? In this session Professor Cochran will discuss the use of several interesting 

and novel active learning exercises, classroom cases, and live projects that 

can dramatically improve student comprehension and retention of key concepts. 

Throughout the session Professor Cochran will emphasize his points with live 

demonstrations of active learning exercises. Card tricks, classroom versions 

of television game shows, and a teaching case with integrated active learning 

will be featured. Because many of these exercises are easily transferable across 

topics, instructor/classroom styles, cultures, national borders, institutions, faculties, 

programs, levels of technology, and class sizes, it is very likely you will 

walk away from this session with ideas on how to improve your own teaching 

(indeed, Professor Cochran will be very disappointed if you don’t!). 

� TA-02 


3.2.14 




Chair: Aristide Mingozzi, Department of Mathematics, University of 

Bologna, C.d.L. Scienze dell’Informazione, Via Sacchi, 3, 47023, 

Cesena, FC, Italy, mingozzi@csr.unibo.it 

1 - Theoretical investigations on maximal dual feasible 

functions 

J. M. Valério de Carvalho, Departamento de Produção e 

Sistemas, Universidade do Minho, 4710 053, Braga, Portugal, 

vc@dps.uminho.pt, Jürgen Rietz, Cláudio Alves 

Dual feasible functions are used to get valid inequalities and lower bounds for 

integer linear problems. We illustrate their use in some examples, we provide a 

simpler proof for maximality, and we describe new results concerning the extremality 

of functions of the literature. Extremal functions are a dominant class 

of dual feasible functions. 

2 - Monoidal Cut Strengthening Revisited 

Andrea Qualizza, Tepper School of Business, Carnegie Mellon 

University, 5000 Forbes Avenue, 15213, Pittsburgh, PA, United 

States, qualizza@cmu.edu, Egon Balas 

We discuss an enhancement of the Balas-Jeroslow procedure for strengthening 

disjunctive cuts for mixed 0-1 programs. When applied to a split cut derived 

from a source row of the simplex tableau, the enhanced procedure yields, besides 

the mixed integer Gomory cut, also cuts that are sometimes stronger. 

3 - Models and Algorithms for Multi-Echelon Distribution 

Networks 

Roberto Wolfler-Calvo, LIPN, Université Paris Nord, 93430, 

Villetaneuse, France, roberto.wolfler@lipn.univ-paris13.fr, 

Aristide Mingozzi, Roberto Baldacci 

EURO 24 - Lisbon 2010 TA-04 

In Multi-Echelon Distribution Networks the deliveries from the production 

plants to the customers are managed through intermediate depots by means 

of two or more levels of delivery. Generally, two different types of decisions 

must be addressed in designing this type of distribution networks. The first 

one is of strategic type and consist in opening one or more depots, on a given 

set of a priori defined depot locations, and to design, for each opened depot, a 

number of routes in order to supply the demands of a given set of customers. 

This problem is called Location Routing Problem (LRP) and the objective is to 

minimize the sum of the route costs and the fixed costs of the opened depots. 

The second one is of tactical type and consist in designing the set of routes that 

must be operated from the different level of the networks. If the deliveries from 

the production plant to the customers are managed by means of two levels of 

delivery, the problem is known as Two Echelon Vehicle Routing Problem (2E- 

VRP) and the objective is to minimize the sum of the routing costs of the first 

and second level routes. In this paper we describe an exact method for solving 

both the LRP and the 2E-VRP. Computational results on benchmark instances 

from the literature are also presented. 

4 - Reference Point based Solution Approach for the Resources 

Constrained Shortest Path Problem 

Luigi Di Puglia Pugliese, D.E.I.S.: Department of Electronics, 

Computer Science and Systems, University of Calabria, Via 

ponte P. Bucci, 87036, Rende, Italy, Italy, 

ldipuglia@deis.unical.it, Francesca Guerriero 

The Resources Constrained Shortest Path Problem (RCSPP) is one of the most 

studied problem in combinatorial optimization. The aim is to find the shortest 

path under additional constraints, representing upper bounds on the consumption 

of resources along the path. In the scientific literature, different approaches 

have been defined to solve the RCSPP. In this work we propose an innovative 

interactive method to address the RCSPP, based on a novel search strategy in 

the criteria space. The performance of the proposed approach is evaluated on 

the basis of an extensive computational study. 

� TA-04 


3.2.13 

Project scheduling 






Chair: José Carlos Reston Filho, P&D, Fundação André Nunes 

Coelho, Rua 4 N o 65 Casa 3, Cj Pq Tropical Parque 10, 69055-743, 

Manaus, Amazonas, Brazil, jcreston@gmail.com 

1 - Towards the minimization of project duration and cost 

in stochastic project management environments 

José Carlos Reston Filho, P&D, Fundação André Nunes Coelho, 

Rua 4 N o 65 Casa 3, Cj Pq Tropical Parque 10, 69055-743, 

Manaus, Amazonas, Brazil, jcreston@gmail.com, Anabela 

Pereira Tereso, Lino Costa 

There is a general consensus that the rational use of available resources is, 

nowadays, an important task, since project managers are interested in the maximization 

of profit and the minimization of risk. In this work, it is proposed 

a multi-objective model to optimize the use of resources in a stochastic environment. 

In this multi-objective formulation the aim is to minimize the project 

duration and the total project cost at the same time. In the future, we intend to 

solve this problem using multi-objective evolutionary algorithms. 

2 - An electromagnetism-like algorithm for a project 

scheduling problem with discounted cash flows 

Marisa Toste, Ciências e Tecnologia, Escola Superior de 

Tecnologia e Gestão de Oliveira do Hospital do Instituto 

Politécnico de Coimbra, Rua General Santos Costa, 3400-124, 

Oliveira do Hospital, Portugal, marisa.toste@estgoh.ipc.pt, 

Dalila Martins Fontes 

An electromagnetism algorithm is developed to address a resource constrained 

project scheduling problem. The chosen problem involves discounted cash 

flows and therefore the objective is to find a schedule that maximizes the project 

net present value. The electromagnetism method (EM) is a population based 

meta-heuristic algorithm utilizing an attraction-repulsion mechanism to move 

sample points (i.e., solutions) towards the optimality. Computational results, 

on problem instances found in previous literature, are reported. 


TA-05 EURO 24 - Lisbon 2010 

3 - Modeling Project Resource Scheduling Problem with 

Activity Compressibility 

M. Karimi-Nasab, Industrial Engineering, Iran University of 

Science and Technology, Narmak, 1684613114, Tehran, Tehran, 

Iran, Islamic Republic Of, mehdikariminasab@yahoo.com, B. 

Shah-Hoseini 

Resource scheduling is a major part of every real project management. The 

problem is here analyzed via a new mixed integer model. As the problem is 

known to be NP-Hard, a new population-based heuristic solution method is developed 

to obtain a near-optimal resource schedule in a deterministic project 

planning problem. The algorithm is run on a set of test data and computational 

experiences report about the superior performance of the algorithm than the 

existing methods. 

� TA-05 


3.2.16 

EURO Excellent in Practice Award 2010 

Stream: EURO Excellence in Practice Award 2010 


Chair: M. Grazia Speranza, Dept. of Quantitative Methods, 

University of Brescia, C.da Santa Chiara, 50, 25122, Brescia, Italy, 

speranza@eco.unibs.it 

1 - Billerud optimizes its bleaching process using online 

optimization 




mikael.ronnqvist@nhh.no, Patrik Flisberg, Stefan Nilsson 

The company Billerud is a leading packaging company and has four integrated 

pulp and paper mills. Since 2004, the paper mill Skärblacka is using an online 

process control system, OptCab, for its bleaching operations. The core of 

OptCab is an online optimization system that dynamically updates a process 

description and establishes an optimized bleaching control. Between 2004 and 

2008, the use of chemicals has decreased by about 10% which corresponds to a 

saving of 2 million Euros. Additional savings are less negative environmental 

impact and that the final brightness quality is more stable. 

2 - Game Theory for Security: Lessons Learned from Deployed 

Applications 

Milind Tambe, Computer Science and Industrial & Systems 

Engineering Departments, University of Southern California, 

3737 Watt Way, PHE 410, 90089, Los Angeles, CA, United 

States, tambe@usc.edu, Fernando Ordonez, Manish Jain, 

Christopher Kiekintveld, Jason Tsai, Shyamsunder Rathi, James 

PIta 

Intelligently allocating limited security resources to protect important terrorist 

targets is a key concern around the world. Game theory is ideally suited for 

such security allocation, as it accounts for differences in target priorities and adversary 

responses to security strategies. We present our research on fast gametheoretic 

algorithms and its deployed applications: (1) ARMOR, deployed at 

Los Angeles Airport (LAX) since 2007 to randomize checkpoints and canine 

patrols. (2) IRIS, in use since 2009 by the US Federal Air Marshals Service to 

generate random flight schedules for air marshals. 

3 - Catch-Up Scheduling for Childhood Vaccination 

Pinar Keskinocak, Georgia Tech, United States, 

pinar@isye.gatech.edu, Faramroze Engineer, Larry Pickering 

We developed a decision support tool for constructing catch-up schedules for 

childhood immunization, to ensure that a child receives timely coverage against 

vaccine-preventable diseases. We show that the catch-up scheduling problem 

is NP-hard, and develop a dynamic programming algorithm that exploits the 

typical size and structure of the problem. Our approach is unique in methodology, 

information, strategy, and advice it offers to the user. The tool is being 

advocated by the Centers for Disease Control and Prevention and the American 

Academy of Pediatrics. 


� TA-06 


8.2.30 

DEA Methodology VI 



Chair: Giannis Karagiannis, economics, university of macedonia, 156 

egnatia str.„ 54006, thessaloniki, Greece, karagian@uom.gr 

1 - The Malmquist Productivity Index and its Decomposition 

for Radial DEA Models with Single Constant Input 

Giannis Karagiannis, economics, university of macedonia, 156 

egnatia str.„ 54006, thessaloniki, Greece, karagian@uom.gr 

In this paper we explore the Malmquist productivity index and its decomposition 

for radial DEA models with single constant input to evaluate the research 

achievements of each faculty member in the Department of Economics at the 

University of Macedonia, Greece. In this case, scale efficiency change effect 

and the input bias index have no contribution to productivity changes. In the 

proposed setting the single constant input corresponds to each faculty member 

and we consider two outputs (e.g., journal articles and other publications) and 

one attribute (quality of journal articles). 

2 - A Slack-based measure for measuring the efficiency of 

decision-making unit with negative data, using DEA 

Fuh-Hwa Liu, Industrial Engineering & Management, National 

Chiao Tung University, 1001 Ta Shueh Road, Dept. IE&M, 300, 

Hsin Chu City, Taiwan, Taiwan, fliu@mail.nctu.edu.tw, 

Ling-Chuan Tsai 

In this research, we introduce a slack-based measure for Data Envelopment 

Analysis (DEA) problem with negative data. DEA introduced to OR/MS literatures 

by Charnes, Cooper, and Rhodes (1978), which measuring the efficiency 

of a set of decision making units such as firms or public sector agencies 

with technologies characterized positive data only. DEA problems with 

positive and negative data have been one of the main issues for performance 

evaluation and there have been various approaches for them. Emrouznejad, 

Anouze, and Thanassoulis (2010) have reviewed the literature and proposed 

a semi-radial measure. We employed their data set to have the comparisons 

between the results of those approaches and ours. 

3 - Technical Economies of Scope using Data Envelopment 

Analysis 

Ozren Despic, Aston Business School, Aston University, Aston 

Triangle, B4 7ET, Birmingham, West Midlands, United 

Kingdom, o.despic@aston.ac.uk, Konstantinos Bakoulas, 

Emmanuel Thanassoulis 

This paper presents a DEA approach on identifying the existence of technical 

economies of scope. We introduce the notion of technical economies of scope, 

which exists when lower inputs are needed to produce an output bundle by a 

single diversified firm rather than to produce each individual output by separate 

specialized firms. The paper reviews previous DEA-based approaches and 

develops a new approach to identifying and measuring technical economies or 

diseconomies of scope. 

� TA-07 


8.2.47 

DEA Application X 



Chair: Luis C. Dias, Faculdade de Economia / INESC Coimbra, 

University of Coimbra, Av Dias da Silva 165, 3004-512, Coimbra, 

Portugal, lmcdias@fe.uc.pt 

1 - Efficiency changes in Turkish airports for the period 

2006-2009 

Erhan Berk, Hava Duragi Lojmanlari 4.Blok Daire:1, 06790, 

ANKARA, Turkey, eberk@kho.edu.tr, Diyar Akay

The global economic crisis has been an impact on aviation sector like the other 

sectors in recent years. Turkey is one of the countries not adversely affected 

from this global crisis in terms of aviation sector. There are several factors 

that make this success possible, and of the important one among them are airports. 

In this study, efficiency of the airports in Turkey is measured for the 

period 2006-2009, and their changes with respect to the size and the regions 

are observed and discussed. 

2 - Benchmarking university libraries: The influence of different 

environments 

Gerhard Reichmann, Department of Information Science and 

Information Systems, University of Graz, Universitätsstr. 15/F3, 

8010, Graz, Austria, gerhard.reichmann@uni-graz.at, Margit 

Sommersguter-Reichmann 

In this contribution we analyse performance differences across university libraries 

from different countries from a cross-section and a longitudinal perspective, 

thus comprising libraries which act in a strongly regulated environment 

(Germany, Austria) and others which operate in a more competitive, marketoriented 

environment (US, Canada). We use the Malmquist index approach to 

disentangle environmental efficiency from managerial efficiency and to decompose 

productivity changes between 1999 and 2008 into changes in technical 

efficiency and changes in technology. 

3 - Chance-constrained FDH model by genetic algorithm in 

uncertain environment 

Hiroshi Morita, Department of Information and Physical 

Sciences, Osaka University, 2-1 Yadama-oka, 5650871, Suita, 

Japan, morita@ist.osaka-u.ac.jp 

We consider the chance-constrained DEA under uncertainty, where the input 

and output may include some stochastic variations. To overcome the difficulty 

to solve the joint chance-constrained model, we adopt the genetic algorithm in 

uncertain environments. The FDH model is considered under uncertainty, since 

this algorithm is useful to solve the combinatorial optimization problem. The 

possibility of extension to the conventional DEA model is also discussed. 

4 - Discriminating efficient units using L1 norm FDH 

Shinn Sun, Department of Management, Fo Guang University, 

No. 160, Linwei Road, 26242, Jiaosi, Yilan County, Taiwan, 

ssun@mail.fgu.edu.tw 

The purpose of this paper is two-fold: to introduce L1 norm into A&T 

FDH proposed by Agrell and Tind for ranking efficient decision making units 

(DMUs); to compare this new model with FDH by Deprin et al., A&P FDH by 

Puyenbroeck, 0-1 LP FDH by Jahanshahloo et al. The proposed method is able 

to remove the existing difficulties in the FDH. In this study, we examine two 

questions: which FDH model has the highest discrimination power of ranking 

efficient units and what results of ranking are different by these three FDH 

model. In developing this method, we are influenced by Jahanshahloo et al. 

and we exploit the leave-one-out idea and L1 norm. Three numerical examples 

with few DMUs and many DMUs are used for the mathematical comparison of 

these three FDH models. 

� TA-08 


6.1.36 

Project Scheduling: New Results and 

Applications 



Chair: Jenny Nossack, Managment Information Sciences, University 

of Siegen, Hölderlinstraße 3, 57068, Siegen, North Rhine-Westphalia, 

Germany, jenny.nossack@uni-siegen.de 

1 - Multi-Project Scheduling with 2-Stage Decomposition 

Gündüz Ulusoy, Industrial Engineering, Sabanci University, 


gunduz@sabanciuniv.edu, Anıl Can 

A non-preemptive, zero time lag multi-project scheduling problem with multiple 

modes and renewable and nonrenewable resources is considered. A 2-stage 

decomposition approach is adopted to formulate the problem as a hierarchy of 

0-1 mathematical programming model. At stage 1, each project is reduced to 

an activity and the resulting project network is solved so as to maximize NPV. 

Using the starting times and resource profiles obtained in stage 1 each project 

is solved in stage 2 for minimum makespan. Lagrangian relaxation and Branch 

& Cut are applied. 


2 - A Constraint Integer Programming Approach for 

Resource-Constrained Project Scheduling 

Stefan Heinz, Zuse Institute Berlin, Germany, heinz@zib.de, 

Timo Berthold, Marco Lübbecke, Rolf Möhring, Jens Schulz 

We propose a hybrid approach for solving the resource-constrained project 

scheduling problem which is an extremely hard to solve combinatorial optimization 

problem of practical relevance. Jobs have to be scheduled on resources 

subject to precedence constraints such that the resource capacities are 

never exceeded and the latest completion time of all jobs is minimized. We 

show that lower bounds from the linear relaxation of the integer programming 

formulation and conflict analysis are key ingredients for pruning the search 

tree. For five instances of the PSPLIB we report a new lower bound. 

3 - Improving a Make-to-Order Production Process using 

Resource-Constrained Project Scheduling 

Wolfgang Summerauer, Department of Business Administration, 

University of Vienna, Bruenner Strasse 72, 1210, Vienna, 

Vienna, Austria, wolfgang.summerauer@univie.ac.at, Christian 

Almeder, Richard Hartl 

We present a case study of a small Austrian plastics and elastics producer which 

tries to optimize its make-to-order production process considering aspects of 

sustainability like e.g. work conditions, energy consumption and amount of 

waste. The large product portfolio, the need for skilled workers and the scarce 

resources (e.g. furnaces and work stations) make the problem interesting, but 

also challenging to solve. Our approach to overcome those difficulties is to 

tackle it as a resource-constrained project scheduling problem (RCPS). 

4 - Decision support system for continuous production 

Krystsina Bakhrankova, Faculty of Economics, Informatics and 

Social Sciences, Molde University College, Specialized 

University in Logistics, Britvegen 2, 6411 , Molde, Norway, 

krystsina.bakhrankova@himolde.no 

This paper develops a model-based decision support system (DSS) for an existing 

European chemical plant with a multi-stage continuous production process. 

The system comprises two modules: energy cost minimization and output 

maximization. Following the production description, a gist of the underlying 

formulations is presented. Then, the DSS is tested on real data instances with 

a focus on its configuration, practical implications, and implementation challenges. 

The system reflects the essence of the researched process, provides 

substantial cost savings, and improves capacity utilization. 

� TA-09 


6.2.53 

Mathematical Modeling on Timetabling 

Problems: Models and Analytic Network 

Process Approach 



Chair: Mujgan Sagir, IE, Eskisehir Osmangazi University, ESOGU 

IE Dept., Meselik, Eskisehir, Turkey, mujgan.sagir@gmail.com 

1 - EXAM SCHEDULING: PARAMETER ESTIMATION WITH 

the ANP and MATHEMATICAL MODELING ON THE SO- 

LUTION 

Mujgan Sagir, IE, Eskisehir Osmangazi University, ESOGU IE 

Dept., Meselik, Eskisehir, Turkey, mujgan.sagir@gmail.com, 

Zehra Kamisli Ozturk 

An invigilator is a person who supervises students during examinations. Assigning 

invigilators to exams is considered as a phase during the exam scheduling 

process. The problem has its own constraints and a multi objective structure. 

To prioritize the objectives of this problem, an ANP model is constructed 

with ten clusters. All the participants and the criteria with their interrelations 

are defined. Different objectives of the problem are treated as alternatives. 

Relative weights are obtained by using the Super Decisions software. Then 

we solve a previously developed mathematical model by using this new set 

of parameters. It is seen that with the new parameters, model gives different 

solution. 

111

TA-10 EURO 24 - Lisbon 2010 

2 - A Genetic Algorithm for Course Scheduling Problem 

Zehra Kamisli Ozturk, Open Education Faculty, Anadolu 

University, Anadolu University, Faculty of Open Education 

Yunusemre Campuse No: 318, 26470, Eskisehir, Turkey, 

zkamisli@anadolu.edu.tr, Mujgan Sagir 

In this study, a genetic algorithm(GA) is developed to solve multi-objective university 

course scheduling problem. The proposed GA’s based on random keys 

that mean the solutions are encoded by using random keys. Random key based 

genetic algorithm encodes the chromosomes with a length of just the number of 

courses and contains problem’ all information. Another advantage of the algorithm 

is the fact that, we do not need to use problem specific genetic operators 

and repair mechanisms. The algorithm’s tested on some problem instances and 

it’s concluded that competitive results’re obtained. 

3 - Vehicle Routing Problem: An Application Study 

Ediz Atmaca, Industrial Engineering, Gazi University Faculty of 

Engineering and Architecture, Gazi University Faculty of 

Engineering and Architecture, Industrial Engineering 

Department, 06570, Ankara, Turkey, hediz@gazi.edu.tr 

The Vehicle Routing Problem(VRP)can be defined as a problem of finding the 

optimal routes of delivery or collection from one or several depots to a number 

of customers.In this study,VRP with flexible time windows are discussed 

for a supermarket that provides material distribution from main warehouse to 

its stores. The mathematical programming model is developed to improve existing 

routes and solved.The problem that is treated as cost reduction,can be 

solved by using multi-purpose decision making techniques for decision makers 

who want to achieve their multiple objectives in real time. 

4 - A solution approach for staff scheduling in service systems 

Burcin Ozsoydan, Industrial Engineering, Osmangazi University, 

Turkey, fbozsoydan@ogu.edu.tr, Aydin Sipahioglu 

In a service system, staff should be used effectively without causing any breakdowns. 

However, to find an appropriate staff schedule that they desire is not 

easy. This kind of problem generally appears in service systems where they 

serve whole day or large part of the day as in post office, hospital etc. Balancing 

the working hours among the staff or minimizing the total cost may be 

chosen as objective function. In this study, a mathematical model is proposed 

to determine staff schedule considering their individual preferences. And, an 

effective solution approach is offered to solve the model. 

� TA-10 


6.2.56 

Graphs and Networks VI 



Chair: Marc Demange, ESSEC Business School, Bucharest, 

Romania, demange@essec.edu 

1 - A family of Torus Grids that are G-Graphs 

Cerasela Tanasescu, ESSEC Business School, Bucharest, 

Romania, tanasescu@essec.edu, Alain Bretto 

The most famous link between abstract group theory and the theory of graphs 

is certainly provided by Cayley Graphs. A much more recent studied family 

of graphs constructed from groups are G-graphs that have also highly-regular 

properties. One first step in the study of G-graphs is to better understand their 

connection with Cayley Graphs. In this context, many well known graph families 

were recently identified as G-graphs. Here we show that 2n x n Torus 

Grids, for even n (a well known family of Cayley Graphs), are G-graphs. 

2 - On Inverse Matching Problems 

Marc Demange, ESSEC Business School, Bucharest, Romania, 

demange@essec.edu, Laurent Alfandari, Jerome Monnot 

Inverse combinatorial optimization aims to modify as least as possible an instance 

of a combinatorial problem so that a fixed solution becomes optimal. 

Considering the Maximum Matching Problem we mainly consider the boolean 

case where one has to delete the least possible number of edges (changing a 

weight from 1 to 0) to make the fixed matching a maximum one. We also consider 

the extended framework where the aim is to make the fixed matching only 

k-optimal for a fixed k. We propose hardness results and a polynomial case in 

bipartite graphs. 

112 

3 - Design of optical wdm networks 

Amal Benhamiche, CORE-TPN, Orange Labs R&D, 38-40 rue 

du Général Leclerc, 92794, Issy-Les-Moulineaux, France, 

amal.benhamiche@orange-ftgroup.com, A. Ridha Mahjoub, 

Nancy Perrot 

We consider the Grooming, Routing and Wavelength Assignment problem in 

optical WDM mesh networks. This is a network design problem which consists 

in grooming demands in lightpaths, assigning a wavelength to each lightpath 

and routing the traffic on these with minimum cost. We first give an Integer 

Linear Programming formulation for the problem, then we discuss some 

pre-processing procedure and propose a fast heuristic which shows to be very 

efficient for solving large and real instances. We finally provide an illustrative 

application of the proposed heuristic for a real network instance. 

� TA-12 


8.2.39 

ANP 01 


Processes 


Chair: Birsen Karpak, Management, Youngstown State University, 

One University Plaza, WCBA 635, 44555, Youngstown, OH, United 

States, bkarpak@ysu.edu 

1 - Prioritizing the Criteria in Renewable Energy Pricing 

Mehmet Yalçın Altunta¸s, Industrial Engineering, Istanbul 

Technical University Management Faculty, Tanzimat sokak. 

58/20 Göztepe, 34730, ˙Istanbul, Turkey, 

yalcinaltuntas@gmail.com, Gulgun Kayakutlu 

Renewable energy sources are still far from being cost advantageous, despite 

the reduction in costs of technologies. Success in reducing the operational 

pricing will certainly support public acceptance. This study aims to construct a 

pricing model specific to renewable energies maintaining variables unique for 

renewable energies. Almost sixty factors are depicted in literature to be clustered 

as economic, technologic, social, political and environmental attributes. 

Analytical Network Process (ANP) is applied to determine the most important 

factors to be considered in the pricing structure. 

2 - Access of Green Supply Chain Management for SMEs 

Irem Duzdar, Industrial Engineering, Istanbul Arel University, 

Tepekent Campus, Kucukcekmece, Istanbul, Turkey, 

iremd82@gmail.com, Gulgun Kayakutlu 

During recent years, supply chain managers tried more to consider environmental 

issues in their decision making process. Green supply chain management 

(GSCM) is not just about considering environment in supply chain decision 

making processes, but also about productivity and making more profit. Green 

supply chain management can be defined as associating environmental thinking 

into supply-chain management, beginning from the product design; material 

sourcing and selection, manufacturing processes, delivery of the final product 

to the consumers as well as end-of-life management of the product after its 

useful life. Objective of this study is to identify the main criterions of GSCM 

for small and medium size companies (SMEs). The green criterions for those 

companies are defined with Causual Map. Assessment of those criterions is a 

multi criteria decision making problem. Analytical Network Processing (ANP) 

technique is used for this study. After becoming definite the most important 

criterions GSCM culture start to be formed. This paper will be designed behavioural 

strategies for utilizing the GSCM system for SMEs. 

3 - Success factors of for a manufacturing SME in Ohio: 

Women perspective 

Birsen Karpak, Management, Youngstown State University, One 

University Plaza, WCBA 635, 44555, Youngstown, OH, United 

States, bkarpak@ysu.edu, Anne M. McMahon 

This study implements an Analytical Network Process (ANP) based framework 

developed for small /medium manufacturing enterprises to prioritize measures 

of success and related antecedents (collectively called success factors) to 

a woman owned manufacturing company in Ohio. A female expert has judged 

the impact of each factor on other success factors for her business. Key priorities 

are identified. The study results will guide the authors as they pursue 

distinctive issues related to women owned small businesses in Ohio.

� TA-13 


2.2.21 

Continuous Location I 






1 - Line location with restrictions on the slope of the line 

Thorsten Krempasky, Numerische und Angewandte Mathematik, 

Georg August Universität Göttingen, Lotzestraße 16-18, 37083, 

Göttingen, Germany, krempask@math.uni-goettingen.de, Anita 

Schoebel 

Coming from the field of robust estimators, this talk draws a connection to 

location theory. We start with the so-called "Regularized Least Absolute Deviation’ 

(RLAD) regression and look at it as a line location problem. We present 

a solution path for the RLAD regression by solving line location problems with 

vertical distance and restrictions on the slope of the line. Looking at these 

problems with different distance functions leads to interesting results in line 

location as well as in regression theory. 

2 - Hybrid Methods for Solving Location-Allocation Problems 

with Dense Demand 

Alper Murat, Industrial and Manufacturing Engineering, Wayne 

State University, 4815 Fourth Street, q, 48202, Detroit, MI, 

Afghanistan, amurat@wayne.edu, Vedat Verter, Gilbert Laporte 

We develop hybrid methods for solving continuous location-allocation problems 

using gradient-based search methods and the alternate location allocation 

(ALA) heuristic of Cooper (1964). We present the hybridization of ALA 

with two improvement approaches which prioritize location and allocation decisions. 

We illustrate their applications to problems with dense demand. Results 

show that the proposed hybridization improves the efficiency of ALA. 

3 - DCM optimization in continuous location 

Rafael Blanquero, Estadística e Investigación Operativa, 

Universidad de Sevilla, Faculty of Mathematics, Tarfia S.N. 

41012-Seville (Spain), 41012, Seville, Spain, rblanquero@us.es, 

Lenys Bello, Emilio Carrizosa, Eligius M.T. Hendrix 

A function is said to be DCM if it can be expressed as a difference of two 

convex monotonic functions. For DCM functions, it is possible to combine 

convexity and monotonicity properties to obtain sharp bounds, to be used in 

global optimization procedures such as Branch and Bound. In this talk we deal 

with two models for locating a competitive facility in the plane that can be 

successfully solved by using DCM optimization. 

4 - Robust Solutions for Location Problems with Uncertainties 

Kathrin Klamroth, Department of Mathematics and Informatics, 

University of Wuppertal, Gaussstr. 20, 42097, Wuppertal, 

Germany, klamroth@math.uni-wuppertal.de, Markus Kaiser 

We consider location problems where uncertainty not only occurs in the demand 

and position of the existing facilities, i.e., in the location objective, but 

also in the constraints of the problem like, for example, the size and location of 

the feasible region and/or the occurrence and position of barriers to travel. Different 

models to cope with uncertainty like minmax regret models (to handle 

uncertain demands) and recent robustness concepts like recoverable robustness 

(to handle uncertainty in the constraints) are discussed. 

� TA-14 


2.2.15 

Closed-Loop Supply Chains 



Chair: Aysegul Demir, Department of Industrial Engineering, Middle 

East Technical University, Endustri Muhendisligi Bolumu Oda No: 

325, ODTU/Cankaya, 06531, Ankara, Turkey, 

demir.aysegul@gmail.com 


1 - Inventory Management in Closed-loop Supply Chains 

under Non-stationary Demand 

Ratna Babu Chinnam, Industrial & Manufacturing Engineering, 

Wayne State University, 4815 Fourth Street, 48202, Detroit, MI, 

United States, r_chinnam@wayne.edu, Ibrahim Dogan 

This study aims to analyze remanufacturer’s inventory control policy under 

non-stationary demand. The objective is to decide on virgin product replenishment 

quantities under used product returns. The exact solution to this inventory 

control problem in our setting is complex and time demanding. We offer and 

analyze a number of different sub-optimal policies. 

2 - Coordination and uncertainty in closed loop supply 

chains 

Juan Pedro Sepúlveda, School of Engineering, Universidad del 

Desarrollo, Avenida La Plaza 0680, Las Condes, Santiago de 

Chile, Chile, jpsepulveda@udd.cl, Felipe Baesler 

There is a lack of research about coordination in closed loop supply chains, 

Ketzenberg (2008).Principally the works in this area use approaches like game 

theory or contracting (Guide and Wassenhove, 2009), so we think that there is a 

gap in the studies about the value of coordination on operational decisions like 

inventory management. In this paper we will analyze quantitatively the gains of 

coordination, through the analysis of ad-hoc inventory models, in the reverse 

logistics setting and we evaluate the gains for different levels of uncertainty 

about customer demand and returns. 

3 - Pricing and Production Decisions in Reusable Container 

Systems 

Busra Atamer, Industrial Engineering, Middle East Technical 

University, ODTU Endustri Muhendisligi Bolumu, Oda No:326, 

Orta Dogu Teknik Universitesi Kampusu, 06531, Ankara, 

Turkey, busra@ie.metu.edu.tr, Ismail Serdar Bakal, Z. Pelin 

Bayindir 

In this study, we consider decision problems on reusable container systems. 

The return rate of containers depends on customer demand and deposit price; 

and the producer determines the return rate via modifying the latter.Production 

planning and pricing decisions are to be made simultaneously because the producer 

also has to decide on the quantity of brand new reusable containers to 

be purchased.In this environment, we maximize the manufacturer’s profit by 

utilizing constrained non-linear optimization techniques.We gain significant insights 

with our analytical and computational observations. 

4 - On the value of Radio Frequency Identification (RFID) 

technology for managing pools of Returnable Transport 

Items (RTIs) 

Aysegul Demir, Department of Industrial Engineering, Middle 

East Technical University, Endustri Muhendisligi Bolumu Oda 

No: 325, ODTU/Cankaya, 06531, Ankara, Turkey, 

demir.aysegul@gmail.com, Simme Douwe Flapper, Sedef Meral 

Limited asset visibility is a key problem in the management of RTIs. One way 

of increasing asset visibility is RFID technology. However, RFID requires high 

investment cost and intense efforts for implementation. In this study, we investigate 

the value of increase in asset visibility and improvement opportunities 

provided by RFID technology for the management of RTI pools in a closedloop 

supply chain setting both considering its costs and benefits with the help 

of mathematical models. We present the results obtained with the models for a 

case study, as well as topics for further research. 

� TA-15 


2.2.12 

Vehicle Routing Problems with Pickups and 

Deliveries 



Chair: Gunes Erdogan, Industrial Engineering, Ozyegin University, 

Özye˘gin Üniversitesi Lojmanları, Re¸sat Bey Sok. No: 21 Apt. B6, 

34000, ˙Istanbul, Turkey, gunes.erdogan@ozyegin.edu.tr 

113


1 - An optimization model for the Vehicle Routing Problem 

with Multiple Container Load 

Massimo Di Francesco, Department of Land Engineering, 

UNIVERSITY OF CAGLIARI, Piazza d’armi, 09123, 

CAGLIARI, Italy, mdifrance@unica.it, Teodor Gabriel Crainic, 

Paola Zuddas 

We study a vehicle routing problem with pickups and deliveries of full container 

loads from/to an intermodal terminal. Trucks can carry more than one 

container and are requested to wait containers emptied after deliveries to satisfy 

pick-up demands. To address this problem, we propose an optimization model. 

We will provide randomly generated experiments as well as comparisons to the 

decisions of real-world carriers. 

2 - Optimization of Batteries Distribution and Route Planning 

with Pickups and Deliveries 

Sónia Cardoso, Departamento de Engenharia e Gestão, Instituto 

Superior Técnico, Portugal, soniaraquel20@hotmail.com, Ana 

Paula Barbósa-Póvoa 

Nowadays logistics plays a key role in improving companies’ efficiency and 

ensuring competitive advantages. One of the main focuses of logistics is the 

distribution and in order to optimize the resources used, vehicle routing problems 

have been widely studied. This paper presents the solution of a distribution 

problem of a Portuguese company producer and retailer of batteries. A 

MILP model was developed to define the optimal set of routes that minimize 

total costs in terms of the number of vehicles and the total travelling distance, 

taking into account the vehicles’ capacity. 

3 - The Traveling Salesman Problem with Pickups, Deliveries 

and Handling Costs 

Daniele Vigo, DEIS, University of Bologna, Via Venezia 52, 

47023, Cesena, Italy, daniele.vigo@unibo.it, Maria Battarra, 

Gunes Erdogan, Gilbert Laporte 

This paper introduces a new variant of the One-to-Many-to-One Single Vehicle 

Pickup and Delivery problem that incorporates the handling cost incurred 

to rearrange the load at customer locations. The simultaneous optimization 

of routing and handling costs is difficult and the resulting loading patterns are 

hard to implement in practice. Branch-and-cut algorithms based on integer 

linear programming formulations of simplified loading policies are proposed. 

Extensive testing on instances involving up to 25 customers indicate that the 

simplified policies favorably compare with the optimal one. 

4 - A Branch-and-Cut Algorithm for the Non-Preemptive 

Capacitated Swapping Problem 

Gunes Erdogan, Industrial Engineering, Ozyegin University, 

Özye˘gin Üniversitesi Lojmanları, Re¸sat Bey Sok. No: 21 Apt. 

B6, 34000, ˙Istanbul, Turkey, gunes.erdogan@ozyegin.edu.tr, 

Jean-Francois Cordeau, Gilbert Laporte 

This paper models and solves a capacitated version of the Non-Preemptive 

Swapping Problem. This problem is defined on a complete digraph, at every 

vertex of which there may be one unit of supply of an item, one unit of 

demand, or both. The objective is to determine a minimum cost capacitated 

vehicle route to transport the items in such a way that all demands are satisfied. 

The vehicle can carry more than one item at a time. Three mathematical programming 

formulations, several classes of valid inequalities are provided, as 

well as the results of extensive computational experiments. 

� TA-16 


2.2.14 

OR models and algorithms 



Chair: Federico Perea, Matemática Aplicada 2, Universidad de 

Sevilla, 41092, Seville, Spain, perea@us.es 

1 - Wardrop Equilibria with Risk Averse Users 

114 

Fernando Ordonez, Industrial and Systems Engineering, 

University of Southern California, 3715 McClintock Ave, GER 

240, 90089, Los Angeles, CA, United States, fordon@usc.edu, 

Nicolas Stier-Moses 

Traffic in a transportation network can be modeled as the outcome of users optimizing 

their costs in a network game. There are, however, random events that 

impact user’s traffic costs. In this work we extend network games by adding 

random deviations to congestion costs. Given a model of user risk behavior, we 

show that an equilibrium for this game always exists. We introduce different 

specific equilibrium models that tradeoff accuracy with tractability and show 

computationally that these models are closer to an equilibrium where users optimize 

their desired percentile cost than the classic Wardrop equilibrium. 

2 - A Customer Oriented Model for Public Transportation 

Timetables 

Miguel Angel Pozo, Universidad de Sevilla, Alcalde Isacio 

Contreras, N o 8, 1 o A, 41003, Seville, Spain, miguelpozo@us.es, 

Francisco A. Ortega 

The Transit Network Timetabling Problem (TNTP) aims to determine optimal 

timetables for each line in a transit network by establishing departure and arrival 

times at each station. The main objective in public transportation consists 

of providing an optimal service for passengers under fleet size constraints. This 

paper deals with the optimal timetable for a given number of vehicles when 

users differ between their desired travel times. According to different transport 

policies, the problem is formulated through two location-allocation models and 

is solved for real size instances. 

3 - Demand based railway timetables, capacity and quality 

of service. 

David Canca, School of Engineers, University of Seville., Av. de 

los Descubrimientos s/n, Isla de la Cartuja, 41092, Seville, Spain, 

dco@us.es, Alejandro Zarzo, Gabriel Villa, Encarnación Algaba 

Railway scheduling and timetabling have been widely studied in the literature. 

Currently, most efforts focus on the process of solving integer programming 

models both using meta-heuristics and exact approaches. However, in the case 

of medium-frequency, there are only a few approaches based on passenger demand. 

This paper focuses in an integrated approach to modeling the scheduling 

problem as well as considers new issues such as capacity analysis or quality of 

service provided to customers, which are inseparable from the scheduling problem. 

4 - GRASP Algorithms for the Railway Network Design 

Problem 

Federico Perea, Matemática Aplicada 2, Universidad de Sevilla, 

41092, Seville, Spain, perea@us.es, Antonio J. Lozano, Juan A. 

Mesa 

We propose a GRASP algorithm to solve the railway network design problem 

and the robust version of it. These problems are modeled as integer linear programming 

problems, and design a topological railway network maximizing the 

trip coverage in the presence of a competing mode, assuming that links can 

fail. Since this problem is computationally intractable for realistic sizes, grasp 

heuristics are designed for solving it. Some experiments have been carried 

out, and their results make us think that this kind of heuristics are suitable for 

railway network design problems. 

� TA-17 


1.3.14 

Contracts and Sub-contracting for 







1 - A web platform for effective planning of green combined 

transportation in EU 

Vasileios Zeimpekis, Financial & Management Engineering, 

University of the Aegean, Kountouriotou 41, Chios, 82100, 

Chios Island, Greece, vzeimp@fme.aegean.gr, Dimitrios 

Saliaris, Konstaninos Mamasis, Ioannis Minis

This paper presents the design and implementation of an innovative web platform 

that supports shippers and forwarders to seek near optimal intermodal 

transport solutions that take into account cost, transportation time, risk and CO2 

emissions. The paper analyses initially the related requirements of 35 freight 

companies. Subsequently, the design of the platform is presented, which takes 

into account these requirements in order to a) develop an attractive transportation 

plan using available data and related OR techniques, b) request offers from 

registered providers, and c) monitor delivery. 

2 - Solving a milk collection and delivery problem 

Nadia Lahrichi, Management et technologie, CIRRELT, 

ESG/UQAM, CP8888, succ centre-ville, H3C3P8, Montreal, Qc, 

Canada, nadia.lahrichi@cirrelt.ca, Teodor Gabriel Crainic, 

Michel Gendreau, Walter Rei, Louis-Martin Rousseau 

The Fédération des producteurs de lait du Québec is a coalition of dairy farmers 

in province of Quebec. The producers have delegated to the Fédération the 

responsibility for negotiating all conditions of transporting milk, and especially 

costs, from the farm to the plant on their behalf with the transporters. The objective 

is to study the process of contracting with carriers for determining the 

fare of transportation. This process typically starts with assigning dairy farms 

to vehicles and delivering the volume collected to plants. 

3 - Optimisation models and algorithms for waste management 

in South African rural areas 

Elias Jakobus Willemse, Built Environment, CSIR, PO Box 395, 

Pretoria, 0001, Pretoria, Gauteng, South Africa, 

ewillemse@csir.co.za, Johan Joubert 

In rural areas of South Africa, local entrepreneurs are subcontracted by municipalities 

to collect the household waste and transport it to bulk waste containers. 

In this paper optimisation models and algorithms are presented that will help 

municipalities to better plan and implement this strategy. The optimisation 

models are able to determine the location of transfer stations within the service 

area; to sector an area into balanced collection zones—each zone is assigned to 

a subcontractor; and to generate a collection route for each subcontractor. 

� TA-18 


1.3.15 

Supply Chain Management 











1 - An application of the revised weighting method in vendor 

selection with price breaks 

Tunjo Perić, Bakeries Sunce, Komedini 1, 10090, Zagreb, 

Croatia, tunjo.peric1@zg.t-com.hr, Zoran Babic 

This paper investigates the possibility of application of the revised weighting 

method in the problem of vendor selection and quantities supplied with 

price breaks. Goals and criteria are expressed in a hierarchical structure and 

the change of prices depends upon the size of order quantities. The proposed 

methodology is tested on the example flour purchase by the firm dealing with 

production of bread and pastry. 

2 - Interaction analysis of participants in supply chains 

Ivana Kovacevic, Human resource management, Faculty of 

organizational sciences, Jove Ilica 154, 11 000, Belgrade, 

Republic of Serbia, Serbia, kivana@fon.rs, Biljana Panic 

Supply chain efficacy is dependent on fine coordination of participants, precisely 

on quality of their interactions. So, communication could facilitate or 

jeopardize the whole process. Simulation of the beergame example of supply 

chain process revealed some behavioral patterns that are described by transactional 

analysis terminology. Transactional analysis is used as a theoretical 

frame for diagnosing and understanding social interactions. Some suggestions 

for overcoming potential communication problems in supply chains are given 

in the conclusion. 


3 - Simulation of supply chain dynamic behaviour by 

means of discrete event tools 

Christos Papanagnou, Automation, Alexander TEI Thessaloniki, 

Sindos, 54100, Thessaloniki, Greece, chrobo@yahoo.com, 

Panagiotis Tzionas 

The scope of this research is to study the dynamic behaviour of supply chains 

by examining initially the impact of different standard forecasting methods 

such as moving average and exponential smoothing on the bullwhip effect. 

Further, we investigate efficient ordering policies under different scenarios followed 

in cases where information feedback patterns are taken into consideration 

by participants. Finally, we derive the total operational cost of each participant 

and we examine moving average vs exponential smoothing competition 

schemes by minimizing the corresponding operational costs 

4 - Strategic Cost Management and Supply Chain Management: 

lessons from a case study 

Manuel Nunes, Production and Systems, University of Minho, 

Campus de Azurém, 4800-058, Guimarães, Portugal, 

lnunes@dps.uminho.pt, Paulo Afonso 

Strategic management of costs requires analysis that goes beyond the frontiers 

of the firm in order to know the value supply chain as a whole, i.e., right from 

the origins of the technology, materials, human and financial resources that it 

uses, up to the final consumer. In this context, product and/or service costing 

systems must be complemented with methods that are capable of reflecting the 

full impact of costs on a firm’s strategy. In this paper, a methodology based 

on strategic management of costs (applied to the supply chain) that supports 

decision-making is presented. 

� TA-19 



Economical Models 



Chair: Alberto Pinto, Mathematics, University of Minho, Campus de 

Gualtar, 4710-057, Braga, aapinto@math.uminho.pt 

Chair: José Martins, University of Minho, 4700, Braga, 

jmmartins@estg.ipleiria.pt 

1 - Adaptive forecasting methods for time series with complex 

trend patterns 

Vitor Mendonca, CIO - FCUL, University of Lisbon, Edificio 

C6, Campo Grande, 1749-016, Lisboa, Portugal, 

vitormendonca@gmail.com, Antonio Rodrigues 

Growth or development processes may exhibit trend patterns more complex 

than usually assumed by traditional models, including pure sigmoidal growth 

in innovation diffusion. We propose three new modelling approaches, using recursive 

estimation methods, designed to better account for modulated but aperiodic 

changes in curvature and convexity in trends. This is confirmed from comparative 

experiments of their application to an heterogeneous set of time series 

from very different origins – marketing, energy, climate, technology, health, 

etc. 

2 - An economic analysis of biofuels production, processing, 

and distribution in the US 

Hayri Önal, Agricultural and Consumer Economics, University 

of Illinois, 305 Mumford Hall, 1301 W. Gregory Dr., 61801, 

Urbana, Illinois, United States, h-onal@illinois.edu 

Recent energy policies in the US aim to increase the share of renewables in 

transportation fuel. Specifically, 36 billion gallons of ethanol derived from corn 

and cellulosic biomass is required to be blended with gasoline by the year 2022. 

The impact of this on land use and equilibrium in food markets is addressed in 

this paper. We use a large scale spatially explicit mathematical model to determine 

the dynamic equilibria in commodity markets and an optimum network 

for production, processing, and distribution of ethanol. We present the methods 

used and some empirical results of the model. 

115

TA-20 EURO 24 - Lisbon 2010 

3 - Mathematical modeling for sugarcane residual biomass 

exploitation 

Helenice Florentino, Bioestatística, IB UNESP, Depto de 

Bioestatística IB UNESP, Bairro Rubião Júnior S/N, 18618-000, 

Botucatu, SP, Brazil, helenice@ibb.unesp.br 

Brazil is the world’s largest sugarcane producer. There is great concern about 

the crop system used, because the common practice is manual harvesting with 

prior straw burning. The Brazilian authorities have approved a law prohibiting 

the burning. However, the major difficulty in using this residue is how to economically 

transport this biomass from farm to processing center. The aim of 

this work was to develop a model to minimize the cost of the residual biomass 

transfer process and to maximize the residual biomass energy balance. For this 

0-1 multiobjective programming techniques were used. 

4 - A duopoly collusion, dumping and Cournot comparison 

José Martins, University of Minho, 4700, Braga, 

jmmartins@estg.ipleiria.pt, Alberto A. Pinto 

In this study, we consider an economical model where two firms of different 

countries can compete in collusion, Cournot and dumping followed punishment. 

We compare the profits of both firms in these three games in order to 

discover which game is more profitable for each firm. 

� TA-20 


1.3.33A 




Chair: José Fernando Gonçalves, LIAAD, Faculdade de Economia, 

Universidade do Porto, Rua Dr. Roberto Frias, s/n, 4200-464, Porto, 

Portugal, jfgoncal@fep.up.pt 

1 - A Parallel Multi-population Genetic Algorithm for a Constrained 

Two-dimensional Orthogonal Packing Problem 

José Fernando Gonçalves, LIAAD, Faculdade de Economia, 

Universidade do Porto, Rua Dr. Roberto Frias, s/n, 4200-464, 

Porto, Portugal, jfgoncal@fep.up.pt 

This paper addresses a constrained two-dimensional, non-guillotine restricted, 

packing problem, where a set of small rectangles has to be placed into a larger 

stock rectangle to maximize the value of the rectangles packed. The algorithm 

proposed hybridizes a novel placement procedure with a genetic algorithm 

based on random keys. Computational tests are presented using available 

instances taken from the literature. The results validate quality of the solutions 

and the approach. Supported by Fundação para a Ciência e Tecnologia (FCT) 

project PTDC/GES/72244/200. 

2 - An efficient algorithm for enumerating bottom-left stable 

positions and its applications 

Shinji Imahori, Department of Computational Science and 

Engineering, Nagoya University, Furo-cho, Chikusa-ku, 

464-8603, Nagoya, Japan, imahori@na.cse.nagoya-u.ac.jp, 

Yuyao Chien, Yuma Tanaka, Mutsunori Yagiura 

We propose an efficient algorithm for enumerating bottom-left stable positions 

when we are given a set of packed rectangles and one new rectangle. Our algorithm 

uses no-fit polygons of rectangles and it can enumerate all the bottomleft 

stable positions within a short computation time even when the number of 

packed rectangles is more than one million. We also propose a compaction 

algorithm that translates the already packed rectangles closer to one another. 

By combining these two algorithms, we present a heuristic algorithm for the 

two-dimensional rectangle packing problem. 

3 - A combined cutting stock and lot-sizing problem 

Elsa Silva, Centro de Investigação Algoritmi, Universidade do 

Minho, Braga, Portugal, elsa@dps.uminho.pt, Filipe Alvelos, J. 

M. Valério de Carvalho 

The two-dimensional cutting stock problem (2DCSP) consists in the minimization 

of the wastage when cutting a set of rectangular items from a (virtually 

infinite) set of plates. In this communication, we extend the 2DCSP by considering 

that the demand of the items is known for multiple time periods and there 

are storage costs. We propose an integer programming model which integrates 

the 2DCSP and lot-sizing problem. 

116 

4 - Two-Dimensional Cutting Stock Management in Fabric 

Industries and Optimizing the Large Object’s Length 

Shaghayegh Rezaei, Textile Engineering, , Islamic Azad 

University, Science and Research Branch, Ashrafi esfehani 

Highway, Hesarak str., Tehran, Tehran, Iran, Islamic Republic 

Of, shaghayegh.rezaei@gmail.com, Mahdi Shadalooee 

Selection of cutting patterns to minimize production waste is an important issue 

in operations research. In this study, the two-dimensional cutting stock problem 

has been investigated to reduce the cutting waste, with focus on men’s clothing. 

Regular and irregular shapes are enclosed by rectangles; and the metaheuristic 

algorithm of simulated annealing (SA) was used. This research showed that 

the amount of required stock can be reduced in fabric cutting by using the SA 

algorithm. Moreover, if the length of pieces is not fixed, incontrollable stock 

can be changed into controllable ones. 

� TA-21 


6.2.47 

Optimization Modeling IV 



Chair: Manuel Soler Arnedo, Statistics and Operations Research, 

Universidad Rey Juan Carlos, C/ Tulipán s/n, 28933, Móstoles, 

Madrid, Spain, Manuel.soler@urjc.es 

1 - Nonlinear optimization in a thermo-economic analysis 

of a small cogeneration system 

Ana Ferreira, DPS, University of Minho, Azurém, Guimarães, 

4800-058, Guimarães, Portugal, cristina_ferreira0@hotmail.com, 

Celina Pinto Leão, Manuel Nunes, Senhorinha Teixeira, Luís B. 

Martins 

Combined heat and power means that both electrical and thermal energy are 

produced simultaneously. The main objective of this work is the use nonlinear 

optimization techniques in a thermo-economic analysis in the design of a 

small cogeneration system. The maximization of annual worth of the cogeneration 

plant was performed. According to the results obtained from the tested 

scenarios, it can be concluded that the optimal value of cogeneration systems 

capacity is sensitive to capital cost and mostly to the energy prices (gas buying 

and electricity grid-selling prices). 

2 - A MINLP Model for the Hydro Scheduling of an Electricity 

Producer 

Javier Diaz, Sistemas e Informatica, Universidad Nacional de 

Colombia, Cra 80 65-223, Facultad de Minas Bloque m8A 

oficina 212, 1, Medellin, Antioquia, Colombia, 

javidiaz@unal.edu.co 

An MINLP model is developed to maximize the profit of a hydroelectric generation 

company considering the technical efficiency and shut up costs of the 

units. The decision variables are the water discharges. It optimizes the integrated 

operation of a system of cascaded hydroelectric plants. Statistical regressions 

are used to estimate both the technical efficiency of a Francis turbine 

from its ’hill diagram’ and the head-volume relationship of the reservoirs. We 

analyze several case studies of realistic dimensions. 

3 - Billiards games: an optimization approach 

Jean-François Landry, Computer Science, Université Blaise 

Pascal/Université de Sherbrooke, France, 

jean-francois.landry2@usherbrooke.ca, Jean-Pierre Dussault, 

Philippe Mahey 

This presentation will address the difficulties related to solving the discrete/continuous 

aspects present in various billiards games. A short description 

of a few known billiards games will first be provided, along with the distinct 

aspects that characterize them. With the problematic clearly identified, 

an optimization model will then be introduced, which takes advantage of the 

knowledge that can be extracted from this type of problem. Possible optimization 

approaches to solve this model will finally be discussed, along with a short 

demonstration.

4 - Hybrid Optimal Control for Commercial Aircraft Trajectory 

Shaping 

Ernesto Staffetti, Department of Statistics and Operations 

Research, University Rey Juan Carlos, C/ Tulipan s/n, 28933, 

Mostoles, Madrid, ernesto.staffetti@urjc.es, Manuel Soler 

Arnedo, Alberto Olivares 

Given the sequence of phases and flight modes conforming the flight profile of 

a commercial aircraft, the initial and final states and a set of path constraints, 

we solve the problem of finding control inputs and the corresponding trajectory 

of the aircraft that minimizes fuel consumption. Such optimal control problem 

is solved using a direct approach, a collocation method, in which the highly 

nonlinear dynamics of an aircraft is solved using an Interior Point Optimizer, 

IPOPT, an open source software package for large-scale nonlinear optimization. 

� TA-22 



Health Care Scheduling I 



Chair: Tugba Cayirli, Management, Ozyegin University, 

KUSBAKISI CAD. NO: 2, ALTUNIZADE - USKUDAR, 34662, 

ISTANBUL, Turkey, tugba.cayirli@ozyegin.edu.tr 

1 - Optimization of Daily Scheduling for Extramural Health 

Care Services 

Andrea Trautsamwieser, Institute of Production and Logistics, 


Feistmantelstraße 4, 1180, Vienna, Austria, 

andrea.trautsamwieser@boku.ac.at, Patrick Hirsch 

The demand for extramural health care services is rising tremendously. Therefore, 

powerful algorithms are required to assist decision making. A model 

formulation and a solution approach for the daily planning are presented. The 

objective is to minimize the traveling- and waiting times of the nurses and to 

minimize the dissatisfaction level of clients and nurses. A feasible solution has 

to observe suitable assignments of clients to nurses considering for example 

qualification levels. Additionally, working time restrictions, hard time windows, 

and mandatory breaks have to be kept. 

2 - A Universal Appointment Rule for a Single-Server System 

with Walk-ins & No-Shows 

Tugba Cayirli, Management, Ozyegin University, KUSBAKISI 

CAD. NO: 2, ALTUNIZADE - USKUDAR, 34662, ISTANBUL, 

Turkey, tugba.cayirli@ozyegin.edu.tr, Kum-Khiong Yang, Ser 

Aik Quek 

This study introduces a universal appointment rule presented as a mathematical 

function of environmental factors that include probabilities of no-shows, walkins, 

number of patients per session, coefficient of variation of service times and 

relative cost of doctor’s time to patient’s time. Nonlinear regression is conducted 

on simulated data to derive the association between clinic parameters 

and rule formulation. The performance of the rule is evaluated in terms of the 

total cost of system as a weighted sum of patients’ wait time, doctor’s idle time 

and overtime. 

3 - Metaheuristic solutions for the or planning and 


Paolo Landa, Department of Economics and Quantitative 

Methods (DIEM), University of Genova, Via Vivaldi 5, 16126, 

Genova, Italy, paolo.landa@yahoo.it, Elena Tanfani, Angela 

Testi, Roberto Aringhieri, Patrick Soriano 

This paper presents a metaheuristic algorithm to determine Operating Rooms 

(ORs) plans using a block-scheduling system. Given a surgery department 

made up of different sub-specialties sharing a fixed number of ORs, the problem 

is that of determining the allocation of OR time blocks to sub-specialties 

together with the subsets of patients to be operated on over a given planning 

horizon. The problem is formulated as a 0-1 LP model with operative, capacity 

and demand constraints. Since the model is strongly NP-hard, we propose a 

tabu search approach to find good (suboptimal) solutions in reasonable CPU 

times. Computational results are reported and compared with those provided 

by a 0-1 LP based heuristic 


� TA-23 


6.2.49 

MOO: Scheduling Problems 



Chair: Jacques Teghem, MATHRO, Faculté Polytechnique de Mons, 

9, rue de Houdain, 7000, Mons, Belgium, 

jacques.teghem@fpms.ac.be 

Chair: Taicir Loukil, Faculté des Sciences Economiques et de 

Gestion, Route l’aérodrome km 4, BP 3018 Sfax, Tunisia, 3018, Sfax, 

Tunisia, Taicir.Loukil@fsegs.rnu.tn 

1 - A bicriteria flowshop problem with time lags 

Emna Dhouib, LOGIQ, Institu Supérieur de Gestion Industrielle 

de Sfax, Route El Mharza km 1.5, 3018, SFAX, 

emna-dhouib@yahoo.fr, Jacques Teghem, Daniel Tuyttens, 

Taicir Loukil 

We consider the permutation flowshop problem with minimal and maximal 

time lags constraints. The objective is to hierarchically minimize first the number 

of tardy jobs and then the makespan.A mixed integer linear programming 

is proposed and the problem is also solved by local search metaheuristics. Both 

approaches are compared numerically. 

2 - A bi-objective approach for rescheduling a flowshop 

Jacques Teghem, MATHRO, Faculté Polytechnique de Mons, 9, 

rue de Houdain, 7000, Mons, Belgium, 


Following the seminal paper of Hall and Potts, we consider a scheduling problem 

for which a set of new jobs arrives when original jobs are already scheduled 

-but nor yet processed - to minimize a classical objective. We analyze how to 

schedule the total set of jobs, original and new, to minimize simumltaneously 

the classical objective and a disruption objective of the initial schedule.The 

aim is to determine or to approximate the Pareto Front In a first step we apply 

a multiobjective metaheuristic to approximate this set of solutions. In a second 

step, we analyze various particular cases, with one or two machines,for which 

an optimal schedule minimizing the classical objective is obtained by a priority 

rule. For these cases, we propose exact algorithms to obtain the Pareto Front. 

3 - Benchmark generators for reactive multi-mode project 

scheduling problem after a mode modification 

Sonda Elloumi, Quantitative Methods, Faculty of Economic and 

Management Sciences, 3 Rue Ibn Khaldoun, Route de Gremda 

Km4, 3062, Sfax, Tunisia, elloumi_sonda@yahoo.fr, Philippe 

Fortemps, Taicir Loukil 

During its execution, a project can undergo unexpected events. Hence, reactive 

algorithms are crucial. Since no benchmark instances are available for reactive 

algorithms, we propose three complementary benchmark generators. All 

these generators lean on a baseline schedule. Then, they suppose that at a certain 

date of the project progress, the planned mode of a future activity should 

change. This fact is tackled with different manners according to the chosen 

generator, which in its turn depends on the user convenience. 

4 - Bi-criteria thermal-energy production scheduling, using 

genetic algorithms 

Dalila Fontes, Faculdade de Economia e LIAAD-INESC Porto 

L.A., Universidade do Porto, Rua Dr Roberto Frias, 4200-464, 

Portugal, fontes@fep.up.pt, Luís Roque, Fernando A. C. C. 

Fontes 

A genetic algorithm is developed to address a unit commitment problem, involving 

the minimization of the operating costs and of the pollution emission 

costs. Since minimizing pollution may lead to an increase in operating costs 

and vice versa, a bi-objective problem is faced. In order to schedule the energy 

production, one has to determine which plants will used, as well as, the quantities 

to be produced by each of these plants. The GA proposed approximates 

the Pareto front by finding a set of interesting solutions. Computational results 

are reported for a couple of standard test systems. 

117


� TA-24 


6.2.50 

Applications of OR in Life Science 

Informatics 


Medicine 


Chair: Erfan Younesi, Department of Bioinformatics, Fraunhofer 

Institute for Algorithms and Scientific Computing SCAI, Schloss 

Birlinghoven, C3, 53754, Sankt Augustin, NRW, Germany, 

erfan.younesi@scai.fraunhofer.de 

1 - A Temporal Logic Constraint Solving Approach to 

Model Coupling: Case Study on the Effects of Irinotecan 

Injections on the Mammalian Cell Cycle 

Elisabetta De Maria, INRIA, France, 78153, Paris, France, 

elisabetta.demaria@dimi.uniud.it 

In systems biology, the number of available models of cellular processes increases 

rapidly, but re-using models in different contexts or for different questions 

remains a challenging issue. We show how the formalization of experimental 

observations in temporal logic with numerical constraints can be used 

to automatically validate a coupled model and optimize unknown parameter 

values with respect to experimental data. In particular, we study the coupling 

of the mammalian cell cycle, the circadian clock, the p53/Mdm2 DNA-damage 

repair system, and the irinotecan metabolism. 

2 - Gene Network Analysis: Identifying Causal Relationships 

Jörg Zimmermann, Institute of Computer Science III, University 

of Bonn, Roemerstr. 164, 53117, Bonn, NRW, Germany, 

jz@iai.uni-bonn.de 

Microarrays enable the fast acquisition of gene expression profiles. In order to 

get functional models from these profiles, gene networks have emerged as an 

important concept, leading to two main problems: 

1. How to extract a specific network from biological data? 

2. How to analyze the network structure? 

Causal relationships can be modeled by directed networks. The resulting 

edge-orienting problem can be solved by using SNPs as "causal anchors" and a 

scoring algorithm, which measures the strength of evidence for a causal direction. 

3 - Detecting early breast cancer with Artificial Immune 

Recognition Systems 

Christos Katsis, Dept. of Applications of Information 

Technology in Administration & Economy, Technological 

Educational Institute of Ionian Islands, Kapodistriou 11, 

Neapolis, GR31100, Lefkas, Greece, ckatsis@cc.uoi.gr, I. 

Gogkou, Christos Papadopoulos, Yorgos Goletsis, Giorgos 

Stylios 

We present a two stage method for a decision support system in early breast 

cancer detection. In hard to diagnose cases, different examinations (mammography, 

ultrasonography, magnetic resonance imaging) provide contradictory results 

and patient is guided to biopsy for definite results. Our method is based 

on Artificial Immune Recognition Systems and is evaluated in real data from 

53 subjects with contradictory diagnoses. The obtained results are promising. 

Comparative results with other approaches are given. The application of such 

an approach can reduce the number of unnecessary biopsies. 

118 

� TA-25 


6.2.48 

Risk Management and Portfolio 

Optimization I 



Chair: Azar Karimov, Financial Mathematics, Institute of Applied 

Mathematics, Middle East Technical University, Eskishehir road, 

06530, Ankara, Afghanistan, azer.kerimov@gmail.com 




1 - Optimal portfolio with annuity purchase 

Tadashi Uratani, Industrial and System Engineerig, hosei 

university, nakacho 18, 162-0835, shinjuku, Japan, 

uratani@hosei.ac.jp, Takashi Kobayashi 

We study the self-annuitization and the dynamic optimal portfolio selection to 

minimize the probability of lifetime ruin. In order to avoid the risk of living after 

spending out his wealth, there are three financial instruments, a risky asset, 

risk free asset, and annuity which guarantee fixed income until death. As a retiree 

is getting older, the annuity price is becoming cheaper to purchase it. The 

problem is to find the optimal portfolio of three financial assets and the timing 

to buy annuity. The optimal solution is holding only the risky asset afterwards 

his wealth equals to the risky investment. We define a annuity price function 

and solve the purchase time as a first hitting time. 

2 - Portfolio rebalancing model using the criterion of entropy 

Wen Yi Lee, Information Management, National Chi Nan 

University, 470 University Road, Puli Nantou 545, Taiwan, 545, 

Puli, Taiwan, s96213502@ncnu.edu.tw, Jing-Rung Yu 

A diversified portfolio is required to form an efficient model with lower risk for 

investors. The maximization of entropy can generate a diversified portfolio. In 

addition to the criteria of entropy, risk and return, the short selling and the transaction 

cost are taken into account to form a more realistic rebalancing model. 

In order to transfer a non-linear entropy model into a linear entropy model, the 

ordered weighted averaging operator (OWA) is adopted. Our model is compared 

with the model with return and risk. A numerical example is illustrated 

the proposed model in more details. 

3 - Do High-Tech Stock Recommendations Have Investment 

Value? 

Chin-Tsai Lin, Graduate Institute of Business and Management, 

Yuanpei University, No. 306, Yuanpei St., 30015, Hsin Chu, 

Taiwan, ctlin@mail.ypu.edu.tw, Yi-Hsien Wang, Fu-Ju Yang, 

Tung-Cheng Hu, Yi-Shan Chen 

This study investigates the investment value of analysts’ stock recommendations 

on the performance of listed high-tech companies and examines whether 

these explanatory variables are associated with observed cumulative abnormal 

returns. These analytical results demonstrate that average abnormal returns are 

significantly positive pre- announcement date, and negative post- announcement 

date. Hence, the observed phenomena can be attributed to the analysts 

generally adopt conservative position following disclosure of information. 

4 - Extreme Value Theory: a new tendency in market risk 

management? 

Selena Totic, Operational Management and Statistics, Faculty of 

Organizational Sciences, Belgrade University, Patrijarha Varnave 

25, 11158, Belgrade, Serbia, selena@fon.rs, Milica Bulajic 

The recent financial crisis brought many questions concerning the adequacy 

of different risk measures and especially Value-at-Risk (VaR) methodologies. 

Among them highly ranked was the deficiency of common VaR methodologies 

to capture fat-tail risks. In this paper, the Extreme Value Theory as an 

alternative approach of measuring VaR is explored. The most important theoretical 

results of univariate EVT are introduced. Special emphasis is paid to 

two methodologies related to the Extreme Value Theory (EVT): the Peaks over 

Threshold (POT) and the Blocks Maxima (BM).

� TA-26 


3.1.11 

Operation research games 

Stream: Cooperative Game Theory 


Chair: Gustavo Bergantinos, Statisitics and O. R., University of Vigo, 

Facultade de Económicas, Universidade de Vigo, 36310, Vigo, 

Pontevedra, Spain, gbergant@uvigo.es 

1 - On Monotonic rules in minimum cost spanning tree 

problems 

Juan Vidal-Puga, Estadística e IO, Universidade de Vigo, 

Facultade de Ciencias Sociais e da Comunicación, Campus A 

Xunqueira sn, 36005, Pontevedra, Pontevedra, Spain, 

vidalpuga@uvigo.es, Gustavo Bergantinos, Anirban Kar 

We characterize, in minimum cost spanning tree problems, the family of rules 

satisfying monotonicity over cost and population. 

2 - A generalization of obligation rules for minimum cost 

spanning tree problems 

Gustavo Bergantinos, Statisitics and O. R., University of Vigo, 

Facultade de Económicas, Universidade de Vigo, 36310, Vigo, 

Pontevedra, Spain, gbergant@uvigo.es, Leticia Lorenzo, Silvia 

Lorenzo-Freire 

Tijs et al (2006) introduce the family of obligation rules for minimum cost 

spanning tree problems. We give a generalization of such family. We prove 

that our family coincides with the set of tules satisfying an additivity property 

and a cost monotonicity property. We also provide two new characterizations 

for the family of obligation rules using the previous properties. In the first one 

we add a property of separability; and in the second one we add core selection. 

3 - An axiomatic approach in minimum cost spanning tree 

problems with groups 

Gómez-Rúa María, Estatística e Investigación Operativa, 

Universidade de Vigo, Facultade de CC. Económicas e 

Empresariais. Lagoas-Marcosende, s/n, 36310, Vigo, 

mariarua@uvigo.es, Gustavo Bergantinos 

We study minimum cost spanning tree problems with groups, where agents are 

located in different villages or cities. The groups are the agents of the same 

village. In Bergantiños and Gómez-Rúa (2009, ET) we define the rule F as the 

Owen value of the irreducible game with groups and we prove that F generalizes 

the folk rule of minimum cost spanning tree problems. Bergantiños and 

Vidal-Puga (2007, JET) give two characterizations of the folk rule. In this paper 

we extend such characterizations to our setting. Some of the properties are 

the same and the other need to be adapted. 

4 - Linear Production Games with Externalities 

Manuel Alfredo Mosquera Rodríguez, Statistics and Operations 

Research, University of Vigo, Edifício Jurídico-Empresarial, 

Campus Ourense, 32004, Ourense, Ourense, Spain, 

mamrguez@uvigo.es, Natividad Llorca, Joaquin 

Sánchez-Soriano 

We deal with Linear Production situations in which there is a limited commonpool 

resource. In this case, if we are interested in studying these problems from 

a cooperative view-point, we should consider that the value of a coalition also 

depends on what the outsiders could do. In this sense, the use of a characteristic 

function in partition function form seems more suitable than the coalitional 

function form. We approach the definition of these functions from a non cooperative 

perspective and propose three different models. Finally, we analyze the 

core-related concepts for the three models. 


� TA-27 


8.2.06 

FREIGHT TRANSPORT AND LOGISTICS 

Stream: Transportation and Logistics [c] 


Chair: Katharina Grobleben, Logistics Management, WHU, 

Burgplatz 2, 56179, Vallendar, Germany, 

katharina.grobleben@whu.edu 

1 - Customer Agreement Characteristics and their Impact 

on Rail Freight Productivity 

Lars Backåker, Department of Science and Technology, ITN, 

Linköping University, Bredgatan 33, SE-60174, Norrköping, 

Östergötland, Sweden, lars.backaker@itn.liu.se, Johanna 

Törnquist Krasemann 

Modern rail freight operators are today continuously striving for increased productivity 

within their service networks. Since the rail freight industry is characterized 

by large-scale, strictly limited service networks, further improvements 

in the area have been proved difficult. Until now, several operation research 

contributions have focused on simplifying operational planning. Moving on, 

we find it necessary to consider how customer agreement characteristics affect 

planning and productivity. A literature review has been established. 

2 - The impact of the equal exchanges recovery strategy in 

closed-loop supply chains of returnable transport item 

(RTI) 

Ruth Carrasco-Gallego, Ingeniería de Organización, 

Administración de Empresas y Estadística, Escuela Técnica 

Superior de Ingenieros Industriales.Universidad Politécnica de 

Madrid., C/Jose Gutierrez Abascal, 2, 28006, Madrid, Spain, 

ruth.carrasco@upm.es, Simme Douwe Flapper, Eva 

Ponce-Cueto 

Many companies using RTIs, like crates, pallets, containers, for the distribution 

of their products have problems in timely getting back these items. One strategy 

used in practice to deal with the above is the equal exchange policy: at the 

moment of delivery the customer has to return as many RTIs as the number of 

RTIs used for delivering the order. An MILP model is used to study the behavior 

of, and the economic cost for the customer and the supplier under the above 

strategy for a number of different situations. The results obtained are discussed 

and directions for further research indicated. 

3 - Hybrid Optimization/Simulation Approach to Design the 

Reverse Logistics Network of a 3PL 

A. Cetin Suyabatmaz, Faculty of Engineering and Natural 

Sciences, Sabanci University, MBDF 1021, Sabanci Universitesi, 


csuyabatmaz@sabanciuniv.edu, F. Tevhide Altekin, Guvenc 

Sahin 

As the product take-back legislations enforced by the governments are put into 

effect, the percentage of companies using reverse logistics services of third 

party logistics service providers (3PLs) have reached to 33% in US and 15% in 

Europe. In this study, we present a hybrid optimization/simulation approach to 

design the reverse logistics network of a 3PL. A mixed integer programming 

model and simulation are iteratively used to design the network while incorporating 

uncertainties associated with the returns. The proposed approach is 

validated by a computational analysis. 

4 - Towards a modal shift in European freight transportation 

Katharina Grobleben, Logistics Management, WHU, Burgplatz 

2, 56179, Vallendar, Germany, katharina.grobleben@whu.edu 

In light of current developments in the transport sector, a modal shift is needed. 

The paper investigates which service attributes have the greatest impact on 

freight mode choice decisions in regard to road, rail-road and barge-road freight 

transport services. In more detail, the paper tests the validity of the rational approach 

for freight mode choice models. Based on a comprehensive literature 

review on intermodal transport and mode choice modeling, a refined empirical 

research methodology is used to investigate the choice behavior of decisionmakers 

for freight transport services. 

119


� TA-28 



Stochastic Programming Models 1 



Chair: Nilay Noyan, Manufacturing Systems/Industrial Engineering, 


nnoyan@sabanciuniv.edu 

1 - A scenario tree approach to Optimal Pharmaceutical 

Research and Development Portfolios using real option 

valuation method 

Majid Rafiee, Industrail Engineering, sharif University, 

Deprtment on Industrail Engineering,sharif University of 

Technology, Tehran, Iran, Tehran, Iran, Islamic Republic Of, 

ma_rafie@yahoo.com 

R&D managers in a pharmaceutical company’s pipeline must consider the entire 

portfolio in the face of market and technological uncertainty and resource 

constraints.This paper presents a stochastic optimization model of pharmaceutical 

multi-period project selection problem using a real option valuation 

method.Then it generates related scenario fan and, via existing algorithms, reduces 

scenarios to generate a new (smaller) scenario tree.Model with reduced 

scenarios has been solved in this paper.Numerical results indicate improved 

portfolios and show the effectiveness of the proposed algorithm. 

2 - A Stochastic Two-Period Model in Airline Seats Inventory 

Control 

Patricia Xufre, CIO-FCUL and ISEGI-UNL, Edificio C6, Campo 

Grande, 1749-016, Lisboa, Portugal, pxufre@isegi.unl.pt 

We formulate and analyse a stochastic two-period inventory control model for 

airline seats management. This is an extension of the classical newsboy model: 

it incorporates the possibility of a price increase between the two periods; and, 

it assumes the existence of consumer diversion, which will cause dependent 

sales. Futhermore, the formulation allows realistic assumptions such as cancellations, 

no-shows and overbooking. We derive the fare allocation limit, as well 

as the initial capacity for this model. We discuss the results obtained and some 

managerial implications. 

3 - A Stochastic Programming Approach for Stochastic 

Assembly Line Balancing with Line Stoppages 

F. Tevhide Altekin, Faculty of Management, Sabanci University, 


altekin@sabanciuniv.edu, Nilay Noyan, Selin Soner Kara 

The stochastic assembly line balancing problem (ALBP) seeks a feasible assignment 

of tasks with stochastic task times to a series of stations such that 

precedence relations are satisfied and some performance measure is optimized. 

This paper studies the stochastic ALBP under the assumption that whenever 

the work content required by at least one workstation exceeds the cycle time, 

the assembly line is stopped. A stochastic programming formulation that minimizes 

the total operating cost is proposed. To demonstrate the cost savings, the 

results of a computational study are also presented. 

4 - Managing Production in Supply Chain under uncertainty 

Yi-Shu Yang, Industrial Engineering and Technology 

Management, Da-Yeh University, 168, University Rd., 

Da-Tusen„ Chang-Hua, Taiwan, joanna@mail.dyu.edu.tw 

According to globalization, the competition of business is drastic. It is necessary 

for enterprises integrated into a supply chain to share information. The 

challenge for supply chain management is demand uncertainty. The purpose of 

this research is, satisfying the uncertain demands based on scenarios, to propose 

a two-stage stochastic programming model for supply chain in order to 

allocate production plans, inventory levels and logistics in a supply chain. Finally, 

a numerical example verifies the feasibility of stochastic programming 

for supply chain. 


� TA-29 


8.2.11 

Satisfiability: Structures and Complexities 



Chair: Stefan Porschen, Institut für Informatik, Universität zu Köln, 

Universität zu Köln, Pohligstrasse 1, 50969, Cologne, NRW, 

Germany, porschen@informatik.uni-koeln.de 

1 - The Strength of Parameterized Resolution 

Olaf Beyersdorff, Institut for Theoretical Computer Science, 

Leibniz University Hanover, Appelstr. 4, 30167 , Hannover, 

Germany, beyersdorff@thi.uni-hannover.de 

Parameterized complexity offers a fine view on classically intractable problems 

yielding efficient solutions for many hard optimization problems. We 

examine the proof-theoretic strength of parameterized resolution introduced 

by Dantchev, Martin and Szeider (FOCS’07). Using a game we obtain lower 

bounds to the proof size in parameterized tree-like resolution for the pigeonhole 

and the order principle. Conversely, we demonstrate the strength of the 

proof system by constructing short refutations of a number of principles hard 

for general resolution. 

Joint work with N. Galesi and M. Lauria. 

2 - On Random Ordering Constraints 

Andreas Goerdt, Fakultät für Informatik, Technische Universität 

Chemnitz, 09107, Chemnitz, Germany, 

goerdt@informatik.tu-chemnitz.de 

Ordering constraints are formally analogous to instances of the satisfiability 

problem in CNF, but instead of a boolean assignment we consider a linear ordering 

of the variables. For many types of constraints the satisfiability problems 

are NP-complete. We look at random ordering constraints. There is a sharp unsatisfiability 

threshold for certain types of constraints, of undetermined value. 

We pursue the problem of approximating the value of the threshold. We show 

that random instances of the betweenness constraint under certain conditions 

are satisfiable with high probability. 

3 - A Constructive Proof of the Lovász Local Lemma 

Robin Moser, Computer Science, ETH Zurich, 

Universitaetsstrasse 6, 8092, Zurich, ZH, Switzerland, 

moserro@gmail.com 

The Lovász Local Lemma is a powerful tool to non-constructively prove the 

existence of combinatorial objects meeting certain criteria. Beck demonstrated 

that a constructive variant can be given under certain more restrictive conditions. 

Simplifications of his procedure and relaxations of its restrictions were 

subsequently exhibited. We devised an poly-time algorithm to find the object 

guaranteed to exist by the Local Lemma. Here, I apply the algorithm to the 

sample application of bounded dependency k-SAT and use a simple information 

theoretic argument to bound the its running time. 

4 - Towards genetic theory of non-Hamiltonicity: A study 

of mutants and their descendents. 

Michael Haythorpe, School of Mathematics and Statistics, 

University of South Australia, Mawson Lakes Boulevard, 

Mawson Lakes, 5095, Adelaide, SA, Australia, 

michael.haythorpe@unisa.edu.au, Pouya Baniasadi, Vladimir 

Ejov, Jerzy Filar 

The Hamiltonian cycle problem is a famously difficult graph theory problem 

that can be stated simply - given a graph with N nodes, can a simple cycle of 

length N be found in the graph, or not? We look at the set of non-Hamiltonian 

cubic graphs and investigate why they contain no Hamiltonian cycles. We show 

that these graphs fall into three categories: bridge graphs, mutants, and descendents. 

Mutants are a new characterisation that is a superset to the set of all 

nontrivial snarks. Descendents are graphs which have a smaller graph as an 

ancestor, that can be identified in polynomial time.

� TA-30 


8.2.13 

MCDA II: Theoretical contributions 


other Issues 


Chair: Marc Pirlot, Mathematics and Operational Research, 

Université de Mons UMONS, Faculté Polytechnique, Rue de 

Houdain 9, B-7000, Mons, Belgium, marc.pirlot@fpms.ac.be 

1 - An empirical study of rank reversals in the 

PROMETHEE methods 

Céline Verly, CoDE-SMG, Université Libre de Bruxelles, Bd du 

Triomphe CP 210-01, 1050, Bruxelles, Belgium, 

celine.verly@ulb.ac.be, Yves De Smet 

The methods based on pair wise comparisons suffer from possible rank reversals 

when the set of alternatives is modified. Different authors have reported 

such "problems’ in AHP, ELECTRE, PROMETHEE, . . . In this talk, we empirically 

study this phenomenon in the scope of the PROMETHEE methods. At 

first, we present different types of rank reversal. Then, we quantify the likelihood 

of rank reversal occurrences in artificial data sets. Finally, we compare 

these results with an ELECTRE-like procedure. 

2 - On the bipolar-valued logical foundation of the classic 

outranking relation 

Raymond Bisdorff, CSC/ILIAS, University of Luxembourg, 6, 

rue Richard Coudenhove-Kalergi, L-1359, Luxembourg, 

Luxembourg, raymond.bisdorff@uni.lu 

Recently, Pirlot and Bouyssou have reported that a strict (asymmetric) outranking 

relation defined similarly to the classic outranking relation is in general not 

identical to its codual relation. This hiatus is problematic as the asymmetric 

part of an outranking relation is commonly identified with the codual relation. 

In this presentation we explore this hiatus in the context of our bipolar-valued 

epistemic logic. By considering an extended bipolar veto principle we are indeed 

able to preserve the identity between the asymmetric part and the codual 

of the outranking relation. 

3 - Subjective expected utility on the basis of ordered categories 

Denis Bouyssou, Université Paris Dauphine, CNRS-LAMSADE, 

Place du maréchal de lattre de tassigny, 75775, Paris Cedex 16, 

France, bouyssou@lamsade.dauphine.fr, Thierry Marchant 

This paper shows that subjective expected utility can be obtained using primitives 

that are much poorer than a preference relation on the set of acts. Our 

primitives only involve the fact that an act can be judged either “attractive”, 

“neutral” or “unattractive”. These categories may be interpreted as denoting 

the position of an act vis-à-vis a status quo. We give conditions implying that 

there are a utility function on the set of consequences and a probability distribution 

on the set of states such that attractive (resp. unattractive) acts have 

a subjective expected utility that is above (resp. below) some threshold. The 

numerical representation that is obtained has strong uniqueness properties. 

4 - Evaluations of Infinite Utility Streams: Pareto-Efficient 

and Egalitarian Axiomatics 

María D. García-Sanz, Economía e Historia Económica, 

University of Salamanca, Edificio FES, Campus Miguel de 

Unamuno, Salamanca, Spain, dgarcia@usal.es 

Two factors influence the resolution of the conflict among infinite generations: 

the consistency/ethical postulates requested; and the utilities that each generation 

can possess. We follow the Basu-Mitra approach to this problem. Firstly 

we examine efficiency and strengthened forms of Hammond Equity for the 

Future both when the utilities are in the unit interval and natural numbers. 

Secondly, we analyze the possibility of combining Pareto-efficiency and the 

spirit of the Hammond Equity principle for both specfications of utilities. We 

conclude that the Anonymity, Hammond Equity for the Future, and Hammond 

Equity ethics can be combined with weak specifications of the Pareto postulate. 


� TA-31 


8.2.15 

OR and Ethics I 

Stream: OR and Ethics 


Chair: Fred Wenstøp, Strategy and Logistics, BI Norwegian School 

of Management, Nydalsveien 37, 0483, Oslo, Norway, 

fred.wenstop@bi.no 

1 - From Policy Narratives to Policy Models 

Giorgio Gallo, Informatica, University of Pisa, Largo B. 

Pontecovo, 2, 56127, Pisa, Italy, gallo@di.unipi.it, Roberto 

Burlando 

A policy narrative is a "story’, having a beginning, middle and end, outlining 

a specific course of events which has gained the status of conventional wisdom 

within a certain community. The "Tragedy of the Commons’ is an example. A 

policy model, instead, describes a situation which, depending on the values of 

its parameters, can give origins to more than one story. Thus it is much more 

powerful in describing the knowledge we have about a particular situation. We 

will discuss positive and negative aspects of the use of policy stories/models, 

together with some of their ethical implications. 

2 - A Decision Framework for Ethics 

Cathal Brugha, Management Information Systems, University 



We use meta-theory to introduce eight meta-adjusting questions: what, where, 

who, which way, whether, whither, when and why, and four meta-adjusting 

dimensions: our commitments and convictions, how we adjust the world and 

how we adjust ourselves. Ethical decision-making uses reverse processes. It 

uses meta-adapting within the questions, first asking why, then when, and so 

on. It also uses meta-adapting within the four dimensions. For each dimension, 

it prioritises adapting ourselves, within adapting the world, within evincing 

what makes most sense, within adducing what is best for all. 

3 - Janus faced moral mathematics, self-interest and emotions 

Haavard Koppang, Gimle Terrasse 3„ 0264, Oslo, Norway, 

haavard.koppang@bi.no, Fred Wenstøp 

We address the question ’Why do we behave the way we do?’ with ’moral 

mathematics’ which we define as reasoning about moral and rationality with 

the help of pure logic and mathematics. The expression has two contrasting 

aspects to it: "moral’ and "mathematics’, a duality we will explore. By additionally 

looking at self-interest and emotions we will elucidate duality anon. 

We visit game theory, Parfit’s concept of self-interest, and the famous trolley 

problems where simple mathematics does not seem to count, as we approach 

an understanding of how moral works and the role of emotions. 

4 - Application of Chinese Doctrine of the Mean into Enterprise 

Leadership 

Tzu-Yin Liang, Department of Chinese, National Kaohsiung 

Normal University, No.116, Heping 1st Rd., Lingya District, 

802, Kaohsiung City, Taiwan, g9343706@yuntech.edu.tw, 

Shan-Yu Su, Shih-Chou Kao 

The aim of the study was to investigate the application of Chinese Doctrine 

of the Mean on the leadership in the enterprise. The content analysis (documentary 

analysis or informational analysis) was used to analyze the leadership 

concept in Chinese Doctrine of the Mean. Two points were addressed in this 

study from depicting the central thinking of Chinese Doctrine of the Mean, 

and having an insight into Western thinking of the leadership. Furthermore, 

to achieve the harmonization and to raise the competitive ability,how to implement 

Chinese Doctrine of the Mean into the Western leadership was described 

into this study. Based on the content analysis in Chinese Doctrine of the Mean, 

the object of the enterprise strategy is to reach the Neutralization Realm of Pre- 

Qin Confucianism. This strategy of the leadership involves the perfect ethics 

and most sincere. The tactic of the leadership is that a leader has to own the 

ability of the self-cultivation and respect the person with virtue. Moreover, the 

leader also has to sympathize with painstaking employees and to tolerate their 

unwitting mistakes. 

121


� TA-32 


8.2.17 

OR in Animal Science 



Chair: Lluis Miquel Pla, Mathematics, University of lleida, Fac. Dret 

i Economia, Jaume II, 73, 25001, Lleida, lmpla@matematica.udl.cat 

1 - Dynamic model of red deer population size 

Dominika Cywicka, Department of Cattle Breeding, University 


Kraków, Poland, cywicka@tlen.pl, Magdalena H˛edrzak 

The level of the game harvest in Poland is based on: population number estimated 

by an inventory, expected birth rate and mortality rate. In binding 

documentation additional information about harvest realised in last season is 

included. It seems quite interesting, that estimated number of red deer remains 

similar in the following years, even though the implemented harvest is lower 

than the planned one. The model of reed deer population dynamics, based on 

the official data, allows to verify the estimated number of red deer, as well as 

the assumptions regarded in the harvest planning. 

2 - Dynamic model of processes of hatching of chicken 

Magdalena H˛edrzak, Department of Cattle Breeding, University 


Kraków, Poland, magdziuch@tlen.pl, Marcin W. Lis, Barbara 

Tombarkiewicz, Krzysztof Pawlak, Jerzy W. Niedziółka 

The brood of wild precocial birds demonstrate the trend to leave of eggs (hatch) 

in the same time. The chicks in the commercial incubators also synchronize the 

stages of hatching (internal and external pipping, hatching). Dynamic of these 

processes depend on many factors e.g. race, age of hens, parameters of incubation, 

period of storage eggs. The aim of the study was to create the dynamic 

model of hatching processes. The decision variables are: layer and broiler type, 

age of the flock, hour of external pipped and hatched of following chicks. The 

Euler’s algorithm was used to computation. 

3 - Dynamic modelling of economic costs of foot and 

mouth disease (FMD) epidemics 

Jarkko Niemi, Economic Research, MTT Agrifood Research 

Finland, Luutnantintie 13, FI-00410, Helsinki, Finland, 

jarkko.niemi@mtt.fi, Heikki Lehtonen, Tapani Lyytikäinen, 

Leena Sahlström, Terhi Virtanen 

FMD is a highly contagious animal disease, which has the potential to cause 

catastrophic economic losses. This study estimates how costs to producers, 

consumers and taxpayers are accumulated. Disease spread among farms is simulated 

with an epidemiological Monte Carlo simulation model that uses spatial 

and temporal data. Market effects are estimated with a stochastic dynamic partial 

equilibrium model. Societal mean costs are EUR 24.2 mill. per epidemic 

(95% EUR 34.3 mill.). Measures to reduce infected farms, outbreak duration 

or loss of exports for fatty dairy product are central to reduce losses. 

4 - Simulating the shift from marine netcages to inland recirculating 

aquaculture systems 

Ilan Halachmi, Inst of Agricultural Engineering, Agricultural 

Research Organization, 50250, Bet Dagan, Israel, 

halachmi@volcani.agri.gov.il 

Recirculating Aquaculture System (RAS) is an intensive production-line implementing 

reused water, continuously produce around the year in its full capacity. 

The ’warmed-up’ (so called transient) period, until it runs to its full production 

capacity might take from few months up to two years. Shorten the transient 

period is curtail for the economic success of a new enterprise. The aim of the 

study was to develop a simulation model aiming at shortening the transient period. 

Gilthead Sea-bream (Sparus aurata) growth data was collected from 2003 

to 2005 from 22 marine net-cages located at the Red Sea. The model combines 

discrete-event as well as continuous-time stochastic variables; it is dynamic, 

i.e., the passage of time plays a crucial role. Model inputs: fish weight, number 

of fish in each batch and the fish batch arrival timing. Potential input combinations 

was 5,314,560,000. Simulation responses: monthly sales, fish stocking 

density (kg biomass/m3) and utilization of each culture tank at any given 

time, standing stock biomass (tons in each tank and in the entire system) and 

feed-load effects on the biofilters. A meta-model was developed, the optimal 

resource configuration maximized the annual profit subjected to constraints. 

The results from the meta-model fed back to the simulation for fine tuning and 

further scenario analysis. The aquaculture farm implemented the results. 

122 

� TA-33 


8.2.19 

Energy Pricing Models 



Chair: Mette Bjørndal, Department of Finance and Management 


Administration, NHH, Helleveien 30, N-5045, Bergen, Norway, 

mette.bjorndal@nhh.no 

1 - Renewable Technologies, Optimal Generation Mix and 

Price Volatility in Competitive Electricity Markets 

Irena Milstein, Faculty of Management of Technology, Holon 

Institute of Technology, 52 Golomb St., 58102, Holon, Israel, 

irenam@hit.ac.il, Asher Tishler 

We develop a two-stage decision model (game) with fossil-using and with 

weather-dependent renewable generating technologies. In the first stage of the 

game, when only the distribution function of demand is known firms choose capacities 

(by maximizing their expected profits). In the second stage, each firm 

selects its daily production of each technology subject to its available capacities 

(the availability of renewable technology depends on the random weather 

conditions). We show that total generation capacity is higher and price spikes 

are substantially higher in the presence of renewable technologies than in their 

absence. We demonstrate these results by applying the model to real world 

data. 

2 - Strategic Biding of GENCOs in Competitive Power Market 

Through Q-Learning Algorithm 

Hasan Rastegar, Electrical Eng., Amirkabir University of 

Technology, AUT,424 Hafez Ave., Tehran, Iran, 424, Tehran, 

Iran, Islamic Republic Of, rastegar@aut.ac.ir, Masoud Rahmani 

Electrical energy market is a multi-agent system that operates in uncertain and 

changing environment with very limited feedback. In this paper is shown by using 

Q-Learning method with applying minimum available information to model 

a supplier in the market, it is possible to distinguish the market participators 

behavior and bid an appropriate price to achieve maximum benefit. First it is 

tried to modify the convergence speed of algorithm concerning the same level 

benefit gradient power plants. Second a navel algorithm based on signing the 

price selection space of each generator is proposed to increase the sensitivity of 

Q-Learning algorithm and stability of proficiency in unexpected load changing. 

3 - Aggregation Choices in Zonal Pricing Algorithms for 

Managing Transmission Congestion in Electricity Markets 

Mette Bjørndal, Department of Finance and Management 


Administration, NHH, Helleveien 30, N-5045, Bergen, Norway, 

mette.bjorndal@nhh.no 

Locational marginal prices constitute a well known benchmark for managing 

transmission capacity constraints in electricity markets. We study aggregation 

choices when simplifying nodal prices into zonal or area prices. We discuss two 

different aggregation concepts, which we call economic and physical aggregation, 

and their relation to optimal nodal prices. In addition to balancing supply 

and demand, the prices derived in a zonal pricing scheme should be consistent 

with the physical constraints of the transmission network. As an illustration we 

consider the present Nord Pool spotprice algorithm.

� TA-34 


8.2.23 


Applications of Lot Sizing and Scheduling 

III 


Quantity 


Chair: Bernardo Almada-Lobo, Industrial Engineering and 

Management, Faculty of Engineering of Porto University, Rua Dr. 

Roberto Frias s/n DEIG, 4200-465, Porto, Portugal, 


1 - A single-stage ATSP-type model for lot sizing and sequencing 

in a beverage production process 

Deisemara Ferreira, Production Engineering, Universidade 

Federal de São Carlos, São Carlos, São Paulo, Brazil, 

deise@dep.ufscar.br, Alistair Clark, Bernardo Almada-Lobo, 

Reinaldo Morabito 

Soft-drink production often involves sequence-dependent changeover times 

and costs and requires synchronisation between production stages. The research 

literature proposes models with constraints for syrup preparation and 

bottling stages. However, this presentation proposes a one-stage formulation 

with ATSP-type lot-sequencing constraints. Different sets of constraints to 

eliminate subtours are tested and compared using instances from a soft-drink 

bottling plant. Moreover, a solution strategy that iteratively patches any subtours 

and provides upper bounds for the model is developed and tested 

2 - The block planning approach for lot-sizing and scheduling 

in make-and-pack production systems 

Hans-Otto Guenther, Production Management, TU Berlin, 

Strasse des 17. Juni 135, 10623, Berlin, Germany, 

hans-otto.guenther@tu-berlin.de 

In the consumer goods, industry make-and-pack production systems can be 

found which consist of a single bottleneck stage after which final products are 

packed and shipped to distribution centres or individual customers. To support 

lot sizing and scheduling in this type of industry the block planning concept 

is proposed. A novel MIP optimization model is presented that determines the 

size and the time phasing of the production lots. Numerical results demonstrate 

the practicability of this approach under experimental conditions which reflect 

typical settings from an industrial case study. 

3 - Flow Shop Production Scheduling Problem with Compressible 

Processing Times 

M.b. Aryanezhad, IE, IUST, Narmak, 1684613114, Tehran, 

Tehran, Iran, Islamic Republic Of, talebi.sahar@gmail.com, S. 

Talebi, M. Karimi-Nasab 

A new mathematical formulation for a production scheduling problem with 

compressible processing times is proposed. The model determines both lot 

sizing and lot streaming of multiple products in a flow shop and takes the undesirability 

of idle times into account. Compressibility of process times in an 

arbitrary time interval leads to nonlinearity. A genetic algorithm (GA) is developed 

to solve the model. Experiments with both real case parameters and 

random test problems shows that the model can consider a variety of real cases. 

The GA gives better solutions than analytical prediction. 

� TA-35 


6.2.46 

Facilitated Decision Analysis II 



Chair: Gilberto Montibeller, Dept. of Management (OR Group), 


London, United Kingdom, g.montibeller@lse.ac.uk 


1 - Evaluating the use of Scenarios and Multi-Criteria Decision 

Analysis (MCDA) as a Framework for Decisionmaking 

under Strategic Uncertainty 

Camelia Ram, Management, London School of Economics, 

Houghton Street, WC2A 2AE, London, United Kingdom, 

camelia.ram@gmail.com, Gilberto Montibeller 

Scenarios provide alternative frames through which the implications of a strategic 

decision can be viewed, so can be very useful when evaluating strategic 

options. Developing scenarios may be too time-consuming to incorporate into 

an evaluation exercise. This paper seeks to explore the effects of applying a 

method of options evaluation based on combining scenario generation rooted 

in Morphological Analysis with Multi-Criteria Decision Analysis. This paper 

will examine how it performs with respect to time; development of new 

options; and consideration of strategic priorities. Three cases relating to the 

strategic vision in agriculture, port services and land use Trinidad & Tobago 

are used. 

2 - Proposal of use the Multicriteria Decision Aid for improve 

the use of assessment tools with focus on performance 

indicators 

Edilson Giffhorn, Production Engineering, Santa Catarina 

Federal University, Rod. Amaro Antônio Vieira, 655, apt. 106-E, 

Bairro Itacorubi, 88034-101, Florianópolis, Santa Catarina, 

Brazil, edilson.giffhorn@gmail.com, Leonardo Ensslin, Sandra 

Ensslin, William Vianna 

The aim of this paper is to present a proposal for structuring indicators to improve 

the use of different approaches of performance evaluation and its applications. 

For this will be used a constructivist Multicriteria Decision Aid approach 

in order to identify, organize and measure ordinally the performance indicators. 

The result is a guide that helps in the identification and construction of performance 

indicators that allow better accuracy to the use of different assessment 

methodologies. 

3 - Stimulating creative thinking with MCDA: modelling the 

market space and exploring alternative perspectives 

Santiago Castro, Research, Cogentus, Soane Point„ 6-8 Market 

Place, RG1 2EG, Reading, Berkshire, United Kingdom, 

santiagocastro@yahoo.com 

This paper shows that, when combining with advanced software technologies, 

Multi-Criteria Decision Analysis (MCDA) becomes a very powerful methodology 

to represent strategic situations, comparing various competitors’ performances 

in all key market factors. Furthermore, modern software functionalities 

currently enable to capture different points of views of stakeholders and 

even go beyond, exploring scenarios, simulating alternative perspectives to redefine 

the company’s offering. Based on practical cases taken from consulting 

work implemented in various industries, this paper illustrates how combining 

facilitation skills and enhanced scenario simulation capabilities can stimulate 

imagination and creativity to finally produce strategic innovation. In that sense 

this article suggests a new scenario approach, which is more about creative 

thinking rather than purely normative assessment of options, or future oriented 

forecasting of scenario planning. 

� TA-36 


3.1.05 

Fuzzy Optimization and Decision Analysis 2 


Intelligence 


Chair: Jaroslav Ramik, Dept. of Math. Methods in Economics, 

Silesian University, School of Business, University Sq. 1934/3, 

73340, Karvina, Czech Republic, ramik@opf.slu.cz 

Chair: Josef Vícha, Mathematical Institute in Opava, Silesian 

University in Opava, Na rybničku 1, 74601, Opava, Czech Republic, 

Josef.Vicha@math.slu.cz 

1 - Fuzzy Bi-matrix Games — Optimistic and Pessimistic 

Approach 

Josef Vícha, Mathematical Institute in Opava, Silesian University 

in Opava, Na rybničku 1, 74601, Opava, Czech Republic, 

Josef.Vicha@math.slu.cz, Jaroslav Ramik 

123


It has been shown that a bi-matrix game with fuzzy goal is equivalent to a crisp 

non-linear programming problem in which the objective as well as all constraint 

functions are linear except two constraint functions, which are quadratic. This 

equivalence is further extended to bi-matrix games with fuzzy pay-offs. In this 

paper we introduce an optimistic and pessimistic approach to this problem. An 

illustrative example is presented and discussed. 

2 - Duality in Fuzzy Linear Optimization 

Milan Vlach, Theoretical Computer Science and Mathematical 

Logic, Charles University, Malostranske namesti 25, 118 00 

Prague 1, 118 00, Prague, Czech Republic, 

milan.vlach@mff.cuni.cz, Jaroslav Ramik 

In the classical linear optimization, the notion of a dual problem to a given 

problem is well understood. To devise a dual problem to a linear optimization 

problem that contains fuzzy data, one has to specify in advance a number of 

fuzzy objects. There are several concepts of duality for fuzzy linear optimization 

leading to results which resemble some of the results established in the 

classical case. We show that a number of duality models and results on fuzzy 

linear optimization from the literature can be obtained as special cases of a 

natural general model. 

3 - A Fuzzy TOPSIS Methodology for the Supplier Selection 

Problem 

Shabnam Mojtahedzadeh Sarjami, Mathematics and Statistics, 

Curtin University of Technology, Kent St, Bentley WA, 6102, 

Perth, Western Australia, Australia, 

Shabnam.mojtahed@postgrad.curtin.edu.au, Louis Caccetta 

The supplier selection problem is to determine a portfolio of suppliers from a 

set of candidates that best meets the requirement of an organisation. In this 

paper, we propose a fuzzy TOPSIS method to select the suppliers. Fuzzy logic 

is applied to cope with uncertainties of criteria weights and supplier’s performance 

ratings while TOPSIS ensures that the ranking order of all suppliers 

has the shortest distances to the positive-ideal solution and farthest distance to 

the negative-ideal solution simultaneously. The proposed method will be illustrated 

through a numerical example. 

� TA-37 


3.1.09 

Mathematical models for energy and 

environment 

Stream: Long Term Planning in Energy, Environment 

and Climate 


Chair: Edi Assoumou, Centre de Mathematiques Appliquees, Mines 

ParisTech, Sophia Antipolis, France, edi.assoumou@cma.ensmp.fr 

1 - ETSAP modeling tools: a bottom-up framework for energy 

issues 

Edi Assoumou, Centre de Mathematiques Appliquees, Mines 

ParisTech, Sophia Antipolis, France, 

edi.assoumou@cma.ensmp.fr, Gilles Guerassimoff, Sandrine 

Selosse, Nadia Maïzi 

The Energy Technology System Analysis Program has been building and 

promoting the use of bottom-up energy models for several decades. Today 

MARKAL/TIMES approaches are used in several countries worldwide. This 

paper is dedicated to the presentation of the optimal approach on which the 

TIMES-FR and TIAM-FR models are relying. Some specific points related 

to its implementation at country level in the French framework and for multiregional 

assessment are presented. 

2 - Mescalito : A simplified approach for global long term 

energy scenarios 

Francois Cattier, R&D, EDF, Site des Renardières, 77818 , Moret 

sur Loing Cedex, France, francois.cattier@edf.fr, Jerome Wirth 

Mescalito is a tool to analyse: -World energy prospects to 2100 : trends in 

demand, supply availability and constraints -Effects of policy actions or technological 

changes -CO2 emissions The Mescalito approach is characterised 

by: -Supply functions taking into account fossil fuel resources but also the rate 

of after-peak production decline -An adjustment of supply and demand made 

directly through quantities, though most of other models use prices -The estimation 

of unsatisfied demand This paper describes Mescalito and highlights its 

outcomes through different scenarios. 

124 

3 - Decision-helping tools for long term investment in energy 

storage systems 

Nicolas Omont, Artelys, 12 rue du 4 septembre, 75002, Paris, 

France, nicolas.omont@artelys.com, Florent Cadoux, Nicolas 

Bonnard, Arnaud Renaud 

The energy sector is characterized by very long asset life spans. Consequently, 

investment decisions are taken with a time horizon ranging up to several 

decades. Associated quantitative decision helping methods therefore face 

uncertainties as their major difficulty: models and optimization methods are 

designed to deal with them. We propose a survey of such methods in order to 

draw their strength and weaknesses from technical and decision maker point 

of views. We will focus on structural versus cyclical uncertainties and on their 

impact on risk and benefit functions on energy storage systems. 

4 - Comparison of statistical and engineering models to 

simulate long-term behavioural changes 

Jean-Michel Cayla, EDF/Mines ParisTech, Paris, France, 

jeanmichel.cayla@gmail.com, Benoit Allibe 

Conservation behaviour in residential space heating have shown a significant 

reduction potential of energy consumption but rarely appear in long term models. 

Several methods allow the simulation of long term behavioural changes, 

leading to various potential estimates. To investigate divergences we used engineering 

and statistical models using a survey on 2012 French households. Our 

results confirm the importance of behaviour in the current energy consumption 

as well as in long term planning. We also highlight sources of uncertainty in 

models and their impact on behavioural foresight results. 

� TA-38 


6.2.44 

Rating and Valuation of Credits 

Stream: Stochastic Valuation for Financial Markets 





Chair: Ömür Ugur, Institute of Applied Mathematics, Middle East 

Technical University, 06531, Ankara, Çankaya, Turkey, 

ougur@metu.edu.tr 

1 - LOTUS-based segmentation in credit scoring 

Katarzyna Bijak, School of Management, University of 

Southampton, Highfield, SO17 1BJ, Southampton, United 

Kingdom, k.bijak@soton.ac.uk 

Credit scoring is used to assess risk of bank customers. A one-scorecard model 

can be built for the customer population. Dividing the population into groups 

(segmentation) and developing separate scorecards for them is believed to improve 

performance of the whole model. In this research, a segmentation method 

is suggested which is based on the Logistic Tree with Unbiased Selection (LO- 

TUS) algorithm that is derived from data mining. It is applied to data provided 

by two of the major UK banks and one of European credit bureaus. As reference, 

logistic regression and classification trees are used. 

2 - Quantitative Analysis on Japanese Corporate Ratings 

with Artificial Neural Network? 

Katsuaki Tanaka, Faculty of Business Administration, Setsunan 

university, 17-8 Ikedanakamachi, 572-8508, Neyagawa, Osaka, 

Japan, k-tanaka@kjo.setsunan.ac.jp, Hideki Katsuda, Motohiro 

Hagiwara, Susumu Saito 

In Japan investors have begun to use corporate ratings given by 4 major rating 

agencies as an index to measure the credit risk of each company. Corporate 

ratings are based on quantitative data and qualitative information that investors 

do not obtain it completely. We propose the artificial neural networks method 

to evaluate the determinant structure of corporate ratings given by rating agencies 

in the USA and Japan with the quantitative data of Japanese manufacturing 

companies only. We also provide the sensitivity analysis about the number of 

explanatory variables.

3 - Predictive abilities of credit spread curves to the industrial 

production in the United States 

Petr Jablonsky, Faculty of finance and accounting, University of 

Economics, Prague, nam. W. Churchilla 4, 130 67 Praha 3, Praha 

3, Czech Republic, Petr.Jablonsky@seznam.cz 

The study examines predictive abilities of term structure of credit spread curves 

to changes in industrial production in the United States. Usually, the term structure 

is represented in models by arbitrary selected tenors. We suggest to parameterize 

the spread curves by a Nelson-Siegel model and utilize the whole term 

structure captured by three parameters — level, slope and curvature. Our following 

empirical analysis is based on US industrial spread curves of different 

credit quality in April 2002 to July 2009. 

4 - Predictive value of different levels of information in 

credit scoring: cross-country comparison 

Galina Andreeva, Business School, University of Edinburgh, 50 

George Sq, EH8 9JY, Edinburgh, United Kingdom, 

Galina.Andreeva@ed.ac.uk 

Credit scoring is a collection of decision support techniques used in consumer 

credit risk management. Models predicting consumer delinquency use different 

types of information: personal characteristics of borrowers, characteristics 

about the product(s) purchased, credit bureau data and, recently, macroeconomic 

variables. The paper analyses the value of different levels of information 

for predicting delinquency and purchase propensity using store card data from 

Belgium, the Netherlands and Germany. Comparisons are made across these 

three countries and with a generic cross-country model. 

� TA-39 


6.2.45 

Limit Behaviour and Approximations II 



Chair: Vladimir Veliov, Institute of Mathematical Methods in 


Argentinierstr. 8/119, 1040, Vienna, Austria, veliov@tuwien.ac.at 

1 - Infinite horizon optimal control problems - existence 

and applications 

Sabine Pickenhain, Mathematics, BTU Cottbus, 

Konrad-Wachsmann-Allee 1, 03046, Cottbus, Germany, 

sabine.pickenhain@tu-cottbus.de 

Control problems with infinite horizon have been investigated since the 1970s 

in the context of models from economics, as well as from natural sciences. 

The need to make a systematic distinction between the different interpretations 

of an objective as either Lebesgue integral or improper Riemann integral has 

been pointed out by the author and her co-workers. The simplest way to solve 

optimal control problems with infinite horizon is to find a solution on a finite 

interval and try to extend the solution onto the whole half-axis. But there is 

no guarantee for the extended solution to be admissible or optimal on an infinite 

interval. For that reason the proof of existence of optimal solutions is very 

important. 

Considering a weight function in the integrand of the objective we propose to 

choose Weighted Sobolev and Weighted Lebesgue-Spaces as state and control 

spaces,respectively and proof existence results for infinite horizon optimal control 

problem. Applications from economics and biology are given. 

2 - On HJB equations associated to the Optimal Control of 

DDE’s: Regularity and Optimal Feedbacks 

Salvatore Federico, Luiss, Rome, via Sant’Andrea, 13, 56125, 

Pisa, Italy, s.federico@sns.it, Fausto Gozzi 

I present a couple of papers dealing with the optimal control problem of delay 

differential equations and state constraint. The main contributions are the proof 

of a regularity result for the viscosity solutions of the associated HJB equation 

and a verification theorem yielding the construction and the characterization of 

optimal controls in some special cases. The class of problems includes some 

problems arising in economics, in particular the so-called models with time to 

build. 


� TA-40 


6.2.52 


Stream: Network Optimization [c] 


Chair: Bessaih Fawzi, Operational research, Université d’Avignon, 

44, chemin des roseaux, 84310, Morires, France, France, 

fawzi@hotmail.fr 

1 - Algorithm for designing survivable TDM networks with 

general failure scenarios 

Pavel Skums, Mechanics and Mathematics, Belarus State 

University, Nezavisimosti av., 4, 220030, Minsk, Belarus, 

skumsp@gmail.com, Yury Orlovich, Yufei Wang 

We consider a problem of designing survivable TDM network in which traffic 

demands of different protection types are routed over the SDH facilities subject 

to given capacity restrictions and prescribed set of failure scenarios. A common 

optimization objective of the problem is to minimize the total cost of the 

network. For this problem we developed an algorithm based on a combination 

of Lagrangian and surrogate Lagrangian relaxations approaches. The computational 

results show that our approach allows obtaining good near optimal 

solutions. 

2 - Lower bounds for the Weight-constrained Minimum 

Spanning Tree problem 

Eulália Santos, University of Aveiro, 3810-448, Aveiro, 

eulalia.santos@sapo.pt, Agostinho Agra, Adelaide Cerveira, 

Cristina Requejo 

We consider the Weight-constrained Minimum Spanning Tree problem 

(WMST). In order to obtain good lower bounds for the WMST we discuss 

families of valid inequalities for the corresponding set of feasible solutions 

and discuss the separation algorithms associated to each family of inequalities. 

Based on these separation algorithms we propose a cutting plane algorithm. 

Finally a computational study based on random generated data is presented. 

3 - Switch Matrix Validation Design on the Telecommuncation 

Satellites 

Bessaih Fawzi, Operational research, Université d’Avignon, 44, 

chemin des roseaux, 84310, Morires, France, France, 

fawzi@hotmail.fr, Philippe Michelon, Dominique Feillet 

The main mission of a telecommunications satellite is to receive a set of signals, 

amplify them, and retransmit them. The tube amplifiers used are complex 

and costly. To over come possible failures, and to ensure the long term success 

of the mission (15 to 20 years), these amplifiers must be duplicated. Several 

billion cases are analysed to confirm that the satellite is robust in all failure scenarios. 

This paper presents a method of resolution based on enumeration and 

systematic research, and stresses the need to reduce the combinatorics. 

4 - Packing edge-disjoint cycles in graphs 

Peter Recht, TU Dortmund, Germany, 

peter.recht@tu-dortmund.de 

Let G = (V,E) be a graph. A cycle packing Z = C1, . . . ,Cl of G is a collection 

of pairwise edge-disjoint cycles Ci of G (i = 1, . . . , l). This talk deals with 

ν(G), themaximumcardinalityofacyclepackingZ.Boundsonν(G)aregiven.Mo 

� TA-41 


3.1.06 

Methodological Aspects of System 

Dynamics Modeling 

Stream: System Dynamics Modeling 


Chair: Jim Duggan, Information Technology, NUI, Galway, 

University Road, Galway, Galway, Ireland, jim.duggan@nuigalway.ie 

125


1 - Risk cost modeling for transportation infrastructure 

project investment 

Steve Jang, Logistics Management, National Defense University, 

3, Ming-you 11th Street, Lane 112, Tao-yuan, 330, Taipei, 

Taiwan, jang_steve@yahoo.com 

A transportation infrastructure project is generally very complex that has highly 

dynamic and interdependent risks. It is principally the cause of cost overrun 

in almost 9 out of 10 transportation infrastructure projects around the world. 

Many studies have revealed that the current risk cost estimating methods do 

not address dynamic and non-linear risk interactions. This paper introduced an 

approach which applied System Dynamics modeling and the statistical techniques 

to sort out this issue. The model validation testing concludes that the 

proposed approach is robust on risk cost estimating. 

2 - The interpretation of system dynamics models 

Martin Rafferty, Accounting and Finance, London South bank 

University, Room L340c, London South Bank University, 

London Road, SE1 6LN, London, United Kingdom, 

rafferm@lsbu.ac.uk 

It is the object of this paper to re-examine some of the less well trodden paths 

within system dynamics. The initial vehicle for the analysis is an examination 

of mass balance, unit consistency, dimensional consistency and determination 

of initial states. The applicability of these concepts to the social sciences is 

analysed. Several cases are examined as system dynamics simulations; two 

from the domain of the physical sciences and two from the social sciences. The 

conclusions of this research are that all models are approximate and therefore 

can only be understood from an interpretive perspective. This is not to say 

that all simulation models are wrong or right; the judgement depends on the 

observer’s frame of reference. 

3 - Modelling Healthcare Processes with System Dynamics 

and Discrete Event Simulation: a Critical Realist Perspective 

Kristian Rotaru, Accounting and Finance, Monash University, 

Australia, 900 Dandenong Rd, 3145, Caulfield East, Victoria, 

Australia, kristian.rotaru@buseco.monash.edu.au, Leonid 

Churilov, Andrew Flitman 

Both System Dynamics (SD) and Discrete Event Simulation (DES) are actively 

used for modelling healthcare processes. Most of the previously published 

comparative SD and DES studies have been a priori adopting either DES or 

SD worldview as a starting point for their comparison, depending on the expertise 

of the authors. The objective of this paper is to investigate how Critical 

Realist philosophy of science can help better modelling in healthcare by facilitating 

explicit articulation of the fundamental assumptions underlying SD and 

DES modelling philosophies. 

4 - An Integrated Methodology for System Dynamics and 

Agent Based Simulation 

Jim Duggan, Information Technology, NUI, Galway, University 

Road, Galway, Galway, Ireland, jim.duggan@nuigalway.ie 

Simulation provides a means to gain insight into the past behaviour and future 

trajectories of complex systems. There are two recognised approaches to social 

simulation: system dynamics (SD), centred on the feedback perspective, 

and agent based simulation (ABS), which uses the individual and their interactions 

as unit of modelling. This paper presents an approach where the two 

methods can be viewed as complimentary rather than conflicting. It focuses 

on the strengths of each method and proposes an iterative cycle — using an 

epidemic model - where SD is used to capture the relationships between the 

target system and other systems, and ABS concentrates on the individual, or 

disaggregate, perspective. 

� TA-42 


3.1.07 

Optimization and Data Mining II 

Stream: Optimization and Data Mining 


Chair: Roberto Santana, Universidad Politecnica de Madrid, 0000, 

Madrid, Spain, roberto.santana@upm.es 

Chair: Concha Bielza, Inteligencia Artificial, Universidad Politécnica 

de Madrid, Campus de Montegancedo, 28660, Madrid, Spain, 

mcbielza@fi.upm.es 

126 

1 - A kernel-based model for spectral clustering with pairwise 

constraints 

Carlos Alzate, ESAT-SCD-SISTA, K. U. Leuven, Kasteelpark 

Arenberg 10, 3001, Leuven, Belgium, 

carlos.alzate@esat.kuleuven.be, Johan Suykens 

A spectral clustering method for incorporating prior knowledge on the cluster 

assignment is presented. This method fits in an optimization framework with 

primal and dual model representations. The primal problem is expressed in 

terms of high dimensional feature maps and it is extended by adding equality 

constraints which represent the prior knowledge. The dual problem is an 

eigenvalue decomposition of a modified kernel matrix. The formulation allows 

out-of-sample extensions and model selection in a learning setting. Simulations 

with toy examples and image processing problems are presented. 

2 - Combining discrete SVM and fixed cardinality warping 

distances for multivariate time series classification 

Carlotta Orsenigo, Management, Economics and Industrial 

Engineering, Politecnico di Milano, p.zza leonardo da vinci 32, 

20133, milan, italy, Italy, carlotta.orsenigo@polimi.it, Carlo 

Vercellis 

Time series classification is a supervised learning problem aimed at labeling 

temporal multivariate sequences of variable length. We propose a new classification 

method, based on an extension of discrete support vector machines, that 

benefits from the notions of warping distance and softened variable margin. 

Furthermore, to transform a dataset into a rectangular shape, we also develop 

a new method based on fixed cardinality warping distances. Computational 

tests on benchmark and real marketing datasets indicate the effectiveness of the 

proposed method in comparison to other techniques. 

3 - Graphical reports for distribution problems 

Jörn Grahl, Information Systems & Business Administration, 

University of Mainz, Jakob Welder-Weg 9, 55099, Mainz, 

grahl@uni-mainz.de 

Empirical studies show that decision makers more likely accept results from 

Decision Support Systems (DSS), when their mental model is aligned with the 

DSS model. A mental model summarizes experiences about the structure of solutions 

as well as the impact of decision variables on the solutions. DSS models 

are hard to understand and might not be directly accessible. Simple graphical 

aggregation schemes are proposed that visualize the impact of decisions and the 

structure of solutions. The current state of research for distribution problems 

and example visualizations are presented. 

4 - Data and graph mining algorithms in estimation of distribution 

algorithms applied to feature subset selection 

problems 

Roberto Santana, Universidad Politecnica de Madrid, 0000, 

Madrid, Spain, roberto.santana@upm.es, Concha Bielza, Pedro 

Larrañaga 

We investigate different facets of the relationship between the data generated by 

estimation of distribution algorithms (EDAs) in feature subset selection (FSS) 

problems, the improvement of these solutions and their interpretation. Different 

examples show the importance of taking into account the relationship between 

the probabilistic models used by EDAs, the choice of the classifier and the representation 

of the selected features. This type of analysis can help to achieve a 

better understanding of the obtained solutions and in some cases improve the 

classification results obtained by FSS. 

� TA-43 


8.2.02 

Revenue Management II 



Chair: Jochen Goensch, Department of Analytics & Optimization, 

University of Augsburg, Universitätsstraße 16, D-86159, Augsburg, 

Germany, jochen.goensch@wiwi.uni-augsburg.de

1 - Performance measures in nested capacity allocation 

mechanisms 

Behzad Samii, Operations and Technology Management Center, 

Vlerick Leuven Gent Management School, Reep 1, 9000, Gent, 

Belgium, behzad.samii@vlerick.be 

In an uncertain demand environment, one approach to provide differentiated 

service levels is to reserve some portion of the available capacity exclusively 

for specific high priority customer classes. In this research, we provide closed 

form solutions for service level measures in a two-class single-period resource 

reservation problem using both Standard and Theft Nesting allocation mechanisms. 

We show that under certain capacity and demand rate conditions one 

mechanism dominates the other and consequently the optimal reserved quantity 

should be decided according to the preferred mechanism. 

2 - Managing Consumer Price Sensitivity 

Sri Devi Duvvuri, Marketing, SUNY at Buffalo, 215 F Jacobs, 

14260, Buffalo, New York, United States, sduvvuri@buffalo.edu 

Effective customer relationship management requires robust measurement of 

consumers’ responses to various marketing stimuli. To test the existence of 

consumer-specific price sensitivity, we estimate two competing models – Thurstone 

and Hierarchical model. The former assumes the existence of only groupspecific 

factors (e.g., factors for related products) for explaining the variation 

in the price sensitivities; the latter posits an additional general or intrinsic price 

sensitivity factor. We use scanner panel data from several categories to test 

these models. 

3 - Optimal Pricing Policy for the Recycable Short Life- 

Cycle Products in a Duopoly Market 

Hsiao-Fan Wang, IEEM, Natl Tsing Hua University, 101 Sec.2, 

Kuang-Fu Rd.„ 30043, Hsinchu, Taiwan, 

hsiaofanwang@gmail.com 

In recent years, the cycle of releasing products is shortened, and many countries 

have set up the barriers of the strict regulation to ensure the green products 

being imported. Since green products are market driven, proposing a method 

for pricing on a green product for its short life cycle in the market is our primal 

intention. This study particularly concerns about the optimal pricing policy 

where there are two kinds of products having different degrees of greenness 

from different manufacturers sold by one retailer. The objective is to maximize 

the retailers’ profit. 

4 - On Retail Assortment, Pricing, and Return Policies 

Alex Grasas, Economics and Business, Universitat Pompeu 

Fabra, Ramon Trias Fargas 25-27, 08005, Barcelona, Spain, 

alex.grasas@upf.edu, Aydin Alptekinoglu 

Using a nested-MNL-based consumer choice model, we study a retailer’s assortment, 

pricing and return policy (fraction of price refunded upon return) 

decisions. Practical circumstances that render prices and refunds exogenous 

to the problem have structural consequences for optimal assortment. When 

all variables are endogenous, the retailer carries some number of most popular 

products; whereas, when prices and refunds are exogenous, it is optimal for the 

retailer to carry a mix of most popular and most eccentric products if the return 

policy is sufficiently strict. 

� TA-44 


8.2.03 

Modelling energy systems 



Chair: Steven De Schepper, VRIJE UNIVERSITEIT BRUSSEL, 

BE-1050, BRUSSELS, Belgium, Steven.De.Schepper@vub.ac.be 

Chair: Jean-Pierre Brans, MOSI (CSOO), V.U.B., Pleinlaan,2, 1050, 

Brussels, Belgium, jpbrans@vub.ac.be 

1 - Using Biodiesel in Belgium 

Steven De Schepper, VRIJE UNIVERSITEIT BRUSSEL, 

BE-1050, BRUSSELS, Belgium, 

Steven.De.Schepper@vub.ac.be 


The paper investigates the use of biodiesel as a substitute for petroleum diesel 

in Belgium by means of system dynamics modelling and statistical techniques. 

The potential advantages and drawbacks with respect to the Belgian economy 

are analysed. Important aspects such as safety of supply, balance of payment, 

prices, impacts on local employment and incomes, environmental and food issues, 

etc. are discussed. Promotion policies implemented by the national and 

regional authorities are also investigated. 

2 - Using EV batteries for grid regulation — A system dynamics 

approach 

Robert Hein, Logistics Management, WHU, Burgplatz 2, 56179, 

Vallendar, Germany, robert.hein@whu.edu, Stefan Spinler 

There are two ways of providing grid regulation by electric vehicle (EV) batteries: 

either by EVs being parked and connected to the grid (vehicle-to-grid) or 

by stationary grid regulation units with discarded battery packs formerly used 

in EVs (battery-to-grid). We evaluate both potential alternatives simultaneously 

using system dynamics to determine how prices for grid regulation and resulting 

salvage values for discarded battery packs evolve over time. Since salvage 

values affect the total costs of ownership, they are likely to have a strong impact 

on EV purchase decisions. 

3 - Improving the performance of a photovoltaic solar 

panel production line 

Ana Raquel Xambre, DEGEI, Universidade de Aveiro, Campus 


raquelx@ua.pt, João Nicolau, Ana Luísa Ramos, Helena Alvelos 

The work refers to a study carried out in order to solve a problem of manufacturing 

process reengineering, in a manufacturer of photovoltaic solar panels 

that wanted to increase its production output by introducing some changes in 

the manufacturing line. The decision maker wanted to analyse the performance 

of the current production system and the impact of the proposed modifications 

so simulation was used in order to assist the decision making process. The use 

of simulation proved to be the correct approach since it allowed the modelling 

of the system dynamics as close to reality as possible. 

4 - Characterisation and classification of Energy Systems 

Models 

Christos Ioakimidis, MECHANICAL ENGINEERING, 

INSTITUTO SUPERIOR TECNICO (IST) - UNIVERSIDAD 

DE LISBOA (UTL), Av. Rovisco Pais, 1049-001, LISBON, 

Portugal, christos.ioakimidis@dem.ist.utl.pt, Paulo Ferrao 

The use of energy models nowdays is one of the basic tools to approach succesfully 

a problem related with the climate change and even more for the energyenvironment-economy 

of a place. A literature review of the use of these modelling 

tools with the characterisation and classification of these energy systems 

models through the creation of a mapping scheme that provides information - 

via indicative clusters - of the most appropriate energy model tool(s) that have 

to be used according to the case study are presented in this work. 

� TA-45 


8.2.12 

Nonconvex Programming approaches for 

Machine Learning and Data Mining 


Approaches 


Chair: Tao Pham Dinh, INSA Rouen, 76131, Rouen, France, 


1 - Sparse Fisher Discriminant Analysis by exact penalty 

techniques in DC Programming 

Mamadou Thiao, LMI INSA of ROUEN, National Institute for 

Applied Sciences of Rouen (INSA), LMI - Avenue de 

l’Université, 76801 Saint-Etienne-du-Rouvray Cedex, 

Saint-Etienne du Rouvray, France, 

mamadou.thiao@insa-rouen.fr, Tao Pham Dinh, Hoai An Le Thi 

127


We investigate the sparse Fisher discriminant analysis which arises in various 

fields such as machine learning and statistics. We propose a new solution for 

the sparse Fisher dicriminant analysis by directly adding to the Fisher discriminant 

problem a constraint for sparsity requirements. We formulate equivalently 

the problem as a DC program (minimization of a DC function over a closed 

convex set) by exact penalty techniques in DC programming and then use the 

local DCA algorithm for the computations. The method is demonstrated with 

some numerical simulations on some real-world data sets. 

2 - Rare event simulations applied to learn an optimal Intelligence 

policy in a hierarchical context 

Frédéric Dambreville, DGA, 16bis Av. Prieur de la Cote d’Or, 

94110, Arcueil, France, submit@fredericdambreville.com, 

Francis Celeste, Fabienne Ealet 

Resource optimization for information collection in the context of Intelligence 

process is a challenging application of military operation research. The main 

issue is to satisfy the prior constraints of many sensors at different operational 

and hierarchical levels, while maximizing the planned information requests. 

Our approach is constraint oriented, and is based on Rare Event simulation. 

This technique (Cross Entropy) is supported by the definition of a family of 

sampling laws, which models the structures (e.g. hierarchical context of the 

decision process) and constraints of the problem. 

� TA-46 


8.2.14 

OR in Fisheries, Maritime Sciences and 

Related Aspects 

Stream: OR in Fisheries, Maritime Sciences and Related 

Aspects 


Chair: Pall Jensson, Engineering Faculty, University of Iceland, 

Hjardarhagi 4, 107, Reykjavik, Iceland, pall@hi.is 




1 - A stochastic optimization model for obtaining a total allowable 

catch quota in a multispecies Chilean fishery 

Victor Albornoz, Departamento de Industrias, Universidad 

Tecnica Federico Santa Maria, Av. Santa Maria 6400, 6671219, 

Santiago, Chile, victor.albornoz@usm.cl, Cristian Canales 

In this contribution the authors present a two-stage stochastic nonlinear optimization 

model proposed as a planning tool in the common sardine and anchovy 

multispecies fishery. The stochastic program is an extension of a nonlinear 

programming one and allows a better representation of the problem, by 

including the uncertainty in the recruitment of new individuals through a finite 

set of possible scenarios, that according to the sensitivity analysis done seems 

an interesting aspect to be considered. Moreover, we are proposing a joint annual 

quota for the multispecies fishery, incorporated as a decision variable in 

the proposed model. 

2 - A Nonlinear Mixed-Integer Stochastic Programming 

Formulation for Fish Processed Production Planning 

Herman Mawengkang, Mathematics, The University of Sumatera 

Utara, FMIPA USU, KAMPUS USU, 20155, Medan, Indonesia, 

hmawengkang@yahoo.com 

Fish and its processed products are the most affordable source of animal protein 

in the diet of most people in Indonesia. The goal in production planning 

is to meet customer demand over a fixed time horizon divided into planning 

periods by optimizing the trade-off between economic objectives such as production 

cost and customer satisfaction level. We use scenario generation based 

approach and feasible neighborhood search for solving the model. The results 

which show the amount of each fish processed product and the number of workforce 

needed in each horizon planning are presented. 

3 - Optimal Control Path of Bangladesh Trawl Shrimp Fishery 

- A continuous Time Analysis 

128 

Mohammed Khan, Finance and Banking, University of 

Chittagong, Professor, Department of Finance and Banking, 

University of Chittagong, Bangladesh, Chittagong, Bangladesh, 

shamimukhan@gmail.com 

The optimum allocation of dynamic system of the renewable resource can only 

provide optimal schedule or time-path indicating optimal amount to be harvested 

in each time period. This time schedule along with optimal rate of harvest 

is the optimal control path, which can manage the resource perpetually. 

The paper attempts to formulate a non-linear dynamic model of Bangladesh 

trawl shrimp fishery for optimal control with continuous-time horizon. An optimal 

control path of Bangladesh marine shrimp fishery is determined using the 

model, which is observed as not managed and utilized optimally. 

4 - A Spatial Resource Allocation Model for Marine Protected 

Area (MPA) Planning 

Yang-Chi Chang, Marine Environment & Engineering, National 

Sun Yat-sen University, 70 Lain-Hae Road, 804, Kaohsiung, 

Taiwan, changyc@mail.nsysu.edu.tw 

MPA is an important measure for maintaining biodiversity and rescuing endangered 

species through effective inhibition of human interferences. Thus how to 

design an effective MPA is an important issue to be explored. The current study 

developed a spatial resource allocation model based on integer linear programming 

for no-take reserve planning. The MPA partitions suggested by the model 

are able to preserve the maximum ecological resources under the limited spatial 

area. Besides, decision makers have the options of choosing from various 

trade-off solutions to fit their decision preferences. 

� TA-47 


8.2.16 

Advances in Discrete-Continuous Optimal 

Control 2 

Stream: Discrete Optimal Control 





Chair: Dmitrii Lozovanu, Institute of Mathematics and Computer 

Science, Academy of Sciences of Moldova, Academiei 5, IMI, 

MD-2020, Chisinau, Moldova, Moldova, Republic Of, 


1 - Collective Taxis in Cities: A Simulation Tool for Optimal 

Real Time Operation and Design 

Jennie Lioris, CERMICS_IMARA, ENPC-INRIA, Domaine de 

Voluceau Rocquencourt, 78150, Le Chesnay, France, 

jennie.lioris@inria.fr, Guy Cohen, Arnaud de La Fortelle 

A discrete event simulation is coupled with real time control algorithms in order 

to study collective taxi systems operating in large cities with or without a 

dispatching center. Repeated simulations and a methodology of analysis of the 

results allow to tune those algorithms but also to dimension resources (number 

of taxis in service, their capacity, etc.) and evaluate the influence of demand 

(level, geometry) on performances. This is an invaluable tool to assess the potentials 

of such a system and choose the best adapted operation mode prior to 

any experiment in the field. 

2 - Trade discounts and pass-through as control tools in a 

vertical distribution channel 

Igor Bykadorov, Sobolev Institute of Mathematics, Siberian 

Branch Russian Academy of Sciences, Acad. Koptyug Av. 4, 

630090, Novosibirsk, Russian Federation, bykad@math.nsc.ru, 

Andrea Ellero, Elena Moretti 

We consider a distribution channel where a manufacturer sells a product to a 

single retailer. Retailer’s motivation and skills are explicitly considered in an 

optimal control framework where the manufacturer controls wholesale price 

and maximizes profit. Channel pricing policies are then considered in a differential 

game framework allowing the retailer to control the mark-up amount. 

For piecewise constant controls we compare manufacturer’s and retailer’s profits 

at Nash and Stackelberg equilibria, considering both the manufacturer and 

the retailer as possible leaders of the channel.

3 - Combined Dynamic Traffic Assignment and Optimal 

Traffic Control models 

Evangelos Mitsakis, Department of Civil Engineering, Aristotle 

University of Thessaloniki, Ploutonos 14, 54655, Thessaloniki, 

Thessaloniki, Greece, emit@civil.auth.gr 

The paper aims to present the major findings of an extensive critical review 

of existing research and literature in the fields of modelling dynamic traffic assignment 

and optimal traffic control. Issues related to the combined use of such 

models are being discussed, while demand uncertainty is also considered. The 

models are presented and analysed both in terms of the underlying mathematical 

formulations, as well as in terms of algorithmic solutions, in order to better 

evaluate quantitative and qualitative results for their practical applicability in 

large scale networks. 

� TA-48 


8.2.04 

Heuristics 1 

Stream: Nonlinear Programming 


Chair: Ana Maria A.C. Rocha, Production and Systems, University of 

Minho, Campus de Gualtar, 4710-057, Braga, Portugal, 

arocha@dps.uminho.pt 

1 - Packing rectangular items in convex regions: an 

heuristic MINLP approach 

Andrea Cassioli, Dipartimento di sistemi e informatica, 

Universita’ di Firenze, Via di S.Marta 3, 50139, Firenze, Italy, 

cassioli@dsi.unifi.it, Marco Locatelli 

The packing of rectangular items in arbitrary convex regions, allowing only 

orthogonal rotations, can be cast as a MINLP unconstrained feasibility problem. 

A bi-level heuristic approach, based on a variable neighbourhood search 

framework, is devised: global exploration is performed perturbing both continuous 

and integer variables, and then, fixed the latter, standard continuous local 

searches are used to find optimal configurations. Numerical results confirm that 

the proposed strategy is effective and robust. 

2 - Heuristics for QFT loop shaping by evolutionary optimization 

Joaquín Cervera, Informática y Sistemas, Universidad de Murcia, 

Facultad de Informática., Campus de Espinardo., 30100, Murcia, 

Murcia, Spain, jcervera@um.es, Alfonso Baños 

This work presents ad-hoc heuristics for Quantitative Feedback Theory (QFT) 

loop shaping, a non convex and nonlinear optimization problem. The authors 

already proposed a suboptimal solution using a fractional structure in the compensator 

and evolutionary optimization. Fractional compensators reduce the 

set of to-be-optimized parameters, which is crucial for evolutionary search success, 

but adequate heuristics are still needed to avoid local optima. This work 

shows considered heuristics for that purpose together with a comparison of the 

results obtained by different heuristics for a case study. 

3 - New concave penalty functions for improving the feasibility 

pump 

Marianna De Santis, DIS, Sapienza, University of Rome, Via 

Ariosto, 25, 00185 , Roma, Italy, mdesantis@dis.uniroma1.it, 

Stefano Lucidi, Francesco Rinaldi 

Mixed-Integer optimization represents a powerful tool for modeling many optimization 

problems arising from real-world applications. The Feasibility pump 

is a heuristic for finding feasible solutions to mixed integer linear problems. 

In this work, we propose a new feasibility pump approach using concave nondifferentiable 

penalty functions for measuring solution integrality. We present 

extensive computational results on binary MILP problems from the MIPLIB 

library showing the effectiveness of our approach. 

4 - Heuristic strategies for optimal design of quality control 

tools 

Vicent Giner-Bosch, Departament d’Estadística i Investigació 

Operativa Aplicades i Qualitat, Universitat Politècnica de 


València, Camí de Vera, s/n, Edifici 7A, 46022, VALENCIA, 

Spain, vigibos@eio.upv.es, Susana San Matías 

Heuristics have been successfully applied in the last decades to solving complex 

decision and design problems in an industrial context. Particularly, optimization 

is being used in process control design, and evolutionary techniques 

such as Genetic Algorithms have proven to be effective in fine tuning quality 

tools such as control charts. Here we propose a heuristic approach for designing 

quality control tools based on exploiting the particular features of the 

functions being involved in each case and we present a concrete application to 

the optimal determination of the Pre-control parameters. 

129

TB-01 EURO 24 - Lisbon 2010 

Tuesday, 10:40-12h00 

� TB-01 


Aula Magna 




Chair: Valerie Belton, Dept. Management Science, University of 

Strathclyde, 40 George Street, G1 1QE, Glasgow, United Kingdom, 

val.belton@strath.ac.uk 

1 - Problem Structuring Methods ’in the Dock’! : Arguing 

the case for Soft OR 

Fran Ackermann, Management Science, Strathclyde University, 

40 George Street, G1 1QE, Glasgow, United Kingdom, 

fran.ackermann@strath.ac.uk 

Problem Structuring Methods (or Soft OR) have been around for nearly 40 

years and yet these methods are still very much overlooked in the OR world. 

Whilst there is almost certainly a number of explanations for this, two key 

stumbling blocks are 1) the subjective nature of the modelling yielding insights 

rather than testable results, and 2) the demand on users to both manage content 

(through modelling) and manage processes (work WITH rather than on behalf 

of groups). This keynote presentation aims to put a case forward to support an 

increase in the use of these methods, either on their own to support clients with 

messy complex problems or in combination with more mathematical methods 

thus facilitating models that address a shared well understood objective, provide 

testable results, and are negotiated and thus owned by key stakeholders. 

� TB-02 


3.2.14 




Chair: J. M. Valério de Carvalho, Departamento de Produção e 


vc@dps.uminho.pt 

1 - Exact Integer Programming in SCIP 

Kati Wolter, Optimization, Zuse Institute Berlin, Takustrasse 7, 

14195, Berlin, Berlin, Germany, wolter@zib.de 

Most MIP solvers focus on quickly finding solutions that are accurate with 

respect to numerical tolerances. There are, however, applications, e.g., chip 

verification, for which this slight inaccuracy is not acceptable. We introduce 

an approach for the exact solution of MIPs in SCIP. It combines inefficient but 

always applicable rational computations with a safe floating-point approach, 

which is efficient but of limited applicability. Preliminary computational results 

will be presented 

2 - The Inverse 1-Median Problem in the d-dimensional 

space with the Chebyshev-Norm 

Johannes Hatzl, Department of Optimization and Discrete 

Mathematics, Graz University of Technology, Steyrergasse 30, 

8010, Graz, Austria, hatzl@opt.math.tugraz.at 

In this talk, we consider the 1-median problem in the d-dimensional space with 

the Chebyshev-norm. We give an optimality criterion for this problem which 

enables us to solve the following inverse location problem in polynomial time: 

Given n points P_1,..., P_n with non-negative weights and a point P_0 the task 

is to change the weights at minimum cost such that P_0 is a 1-median with respect 

to the modified weights. In fact, this problem reduces to a balancing flow 

problem for which an optimal solution can be obtained in polynomial time. 

3 - Stabilization procedures based on dual feasible functions 

130 

Cláudio Alves, Produção e Sistemas, Escola de Engenharia, 

Universidade do Minho, Campus de Gualtar, 4710-057 Braga, 

Braga, Portugal, claudio@dps.uminho.pt, François Clautiaux, J. 

M. Valério de Carvalho, Jürgen Rietz 

Convergence of column generation for the cutting stock problem is addressed. 

We introduce a general framework for deriving dual cuts, and describe a new 

type of cuts, which exclude solutions that are linear combinations of some other 

known solutions. New lower and upper bounds for the dual variables are discussed. 

We also show how the prior knowledge of a good dual solution helps 

improving the results. It tightens the bounds around the dual values, and makes 

the search converge faster if a solution is sought in its neighborhood first. Computational 

experiments on hard instances are reported. 

4 - Improving a lagrangean decomposition for the unconstrained 

binary quadratic programming problem 

Luiz A. N. Lorena, LAC - Lab. Assoc. Computação e Mat. 

Aplicada, INPE - Brazilian Space Research Institute, Av. dos 

Astronautas - 1758, Caixa Postal 515, 12243-970, São José dos 

Campos, São Paulo, Brazil, lorena@lac.inpe.br, Geraldo Mauri 

We present a lagrangean decomposition based on column generation techniques 

to solve the unconstrained binary quadratic programming problem. We use a 

mixed binary linear version of the original quadratic problem with constraints 

represented by a graph. This graph is partitioned in clusters of vertices forming 

sub-problems whose solutions use the dual variables obtained by a coordinator 

problem. Computational experiments consider a set of difficult instances 

and the results show the efficiency of the proposed method over traditional lagrangean 

relaxations and other methods found in the literature. 

� TB-03 


3.2.15 

Routing problems 



Chair: Jose Brandao, Management, University of Minho, Largo do 

Paço, 4704 -553 , Braga, Portugal, sbrandao@eeg.uminho.pt 

Chair: Elisabeth Gussmagg-Pfliegl, Faculty of Business, Economics 

and Statistics, University of Vienna, Chair for Production and 

Opterations Management, Bruennerstrasse 72, 1210, Vienna, Austria, 

elisabeth.gussmagg-pfliegl@univie.ac.at 

1 - A hybrid approach for real world postman problems 

Elisabeth Gussmagg-Pfliegl, Faculty of Business, Economics and 

Statistics, University of Vienna, Chair for Production and 

Opterations Management, Bruennerstrasse 72, 1210, Vienna, 

Austria, elisabeth.gussmagg-pfliegl@univie.ac.at, Karl Doerner, 

Richard Hartl, Stefan Irnich, Fabien Tricoire 

We are solving a rich postman problem using a hybrid approach. Due to the 

complexity of the problem, and the size of the (real world) instances, we tackle 

the problem in two steps: first, we generate routes heuristically, secondly, a 

set covering problem is solved using an exact method. After each step, a local 

search is performed to improve the solution. 

2 - Iterated tabu search algorithm for the open vehicle routing 

problem with time windows 

Jose Brandao, Management, University of Minho, Largo do 

Paço, 4704 -553 , Braga, Portugal, sbrandao@eeg.uminho.pt 

The problem studied here, the open vehicle routing problem with time windows 

(OVRPTW), is different from the vehicle routing problem with time windows 

in that the vehicles do not return to the distribution depot after delivering the 

goods to the customers. We have solved the OVRPTW using iterated tabu 

search. The performance of the algorithm is tested using a large set of benchmark 

problems. 

3 - A heuristic approach for the CVRP with open routes 

R. Aykut Arapoglu, Industrial Engineering Department, 

Eskisehir Osmangazi University, Meselik Campus, 26480, 

Eskisehir, Turkey, arapoglu@ogu.edu.tr, Abdurrahman Yildiz

We present a variant of the classical capacitated VRP that allows "closed’ routes 

ending at the depot as well as "open’ routes that end at customer node(s). The 

problem arises when the company has the option of outsourcing extra vehicles 

in addition to its own fleet of vehicles. The objective is to find a route for 

each vehicle used, minimizing the total distance traveled. Only small sized instances 

of the problem can be solved to optimality using integer programming 

techniques and therefore, a simulated annealing based heuristic procedure is 

proposed for larger problems. 

4 - Multi-Vehicle One-to-One Pickup and Delivery Problem 

with Split Loads 

Mustafa Sahin, Faculty of Engineering and Natural Sciences, 

Sabanci University, 34956, Istanbul, 

mustafasahin@sabanciuniv.edu, Gizem Cavuslar, Temel Öncan, 

Guvenc Sahin, Dilek Tuzun Aksu 

The multi-vehicle one-to-one pickup and delivery problem determines a set of 

least cost vehicle routes to satisfy a set of pickup and delivery requests between 

location pairs. With split loads, a delivery request may be satisfied with more 

than a single vehicle service. We develop a tabu-search algorithm. The improvements 

are obtained through insert, split and swap moves in addition to an 

optimal-split neighborhood considering the optimality conditions that prohibit 

some moves. The algorithm is tested on data sets from the literature; it provides 

good-quality solutions in reasonable time. 

� TB-04 


3.2.13 

Multi-objective scheduling with 




Chair: Juan Carlos Rivera Agudelo, Ciencias Básicas, Universidad 

Eafit, Cr 49 No. 7 Sur - 50. Bloque 38. Oficina 424., 057, Medellín, 

Antioquia, Colombia, jkrivera@gmail.com 




1 - L-ant to solve a multiobjective complex scheduling 

problem 

Frédéric Dugardin, LOSI, University of Technology of Troyes, 

12, rue Marie Curie, 10010, Troyes, France, 

frederic.dugardin@utt.fr, Farouk Yalaoui, Lionel Amodeo 

This work deals with multiobjective scheduling of a reentrant hybrid flow shop 

problem. The objectives are the minimization of the makespan and the minimization 

of the sum of total tardiness. The method employed here is a new 

meta-heuristic called L-ant. This method involves a multi-objective ant colony 

system algorithm which uses the Lorenz dominance relationship. The latter 

provides a different selection from the Pareto one. Then the Lorenz dominance 

increases the convergence speed of the algorithm. This algorithm has been 

compared with other metaheuristics on multiple instances. 

2 - A Genetic Algorithm to Solve the Bi-Objective 

Resource-Constrained Project Scheduling Problem 

Juan Carlos Rivera, Mathematical Engineering, Eafit University, 

Medellín, Antioquia, Colombia, jrivera6@eafit.edu.co 

Resource-Constrained Project Scheduling Problem (RCPSP), is a NP-Hard 

combinatorial optimization problem, which consists of assigning start times 

to a series of activities with fixed duration. Activities have precedence and 

resource constraints for their execution. To solve the bi-objective RCPSP, in 

which minimization of the makespan and maximization of the robustness are 

considered, a Genetic Algorithm based method is proposed. An alternative 

random key representation is used, which allow reducing feasible space. Comparisons 

are made using benchmark instances from the PSPLIB library. 

3 - Genetic Algorithm with Parameter Design and Variable 

Weighting to Bicriteria Open Shop Scheduling Problems 

Hong Tau Lee, Industrial Engineering and Management, 

National Chin-Yi University of Technology, Taipin, Taichung, 

EURO 24 - Lisbon 2010 TB-05 

Taiwan, ROC, 411, Taichung County, Taiwan, 

leeht@ncut.edu.tw, Sheu-hua Chen 

An experimental design is employed to determine the best combination of parameter 

levels that can then be adopted in the genetic algorithm for an openshop 

scheduling problem. The fitness function comprises both the modified 

makespan and total tardiness, thus avoiding the effect of a dominating criterion. 

A variable weighting approach for the two criteria is utilized to alter the 

search directions of each generation and accelerate convergence of the algorithm. 

The proposed genetic algorithm is implemented using the data of a case 

company. The results show that the schedule generated by the proposed genetic 

algorithm outperforms the company’s current FCFS with SPT scheduling 

approach in terms of makespan and total tardiness as well as number of tardy 

jobs. 

4 - Multiobjective Simulation Optimization for Operating 

Room Scheduling 

Felipe Baesler, Industrial Engineering, Universidad del 

Desarrollo, Ainavillo 456, Concepcion, Chile, fbaesler@udd.cl, 

Juan Pedro Sepúlveda 

This paper presents an approach for operating room scheduling considering 

two main conflicting objectives. The approach uses discrete event simulation 

to capture the resources randomness involved in the patient flow process. The 

model interacts with a metaheuristic that searches for better schedules. Pareto 

frontiers for different experiments were constructed using the e-constraint technique 

and showed that the scheduling approach improves the hospital performance 

in comparison to the current method. 

� TB-05 


3.2.16 

EURO Excellent in Practice Award 2010 

Stream: EURO Excellence in Practice Award 2010 





1 - Inventory management of multiple items with irregular 

demand: A case study 

George Nenes, Mechanical Engineering, University of Western 

Macedonia, Bakola & Sialvera, 50100, Kozani, Greece, 

gnenes@uowm.gr, Sofia Panagiotidou, George Tagaras 

We present the case of a Greek enterprise managing inventories of thousands of 

different items with intermittent and lumpy demand. We describe the solutions 

given to several practical problems in the course of devising an easy-to-use yet 

effective and all-encompassing procurement and inventory control system. Using 

simple quantitative tools we develop an efficient procedure for the determination 

of the base stock levels that achieve the target fill rates in the proposed 

periodic review regime. We outline the computerized implementation of the 

new system and the very encouraging results. 

2 - Modelling Performance Appraisal and Career Development 

in the European Commission 

Michael Pidd, The Management School, Lancaster University, 

Dept. Of Management Science, LA1 4YX, Lancaster, 

Lancashire, United Kingdom, m.pidd@lancaster.ac.uk, Dave 

Worthington, Stephan Onggo, Didier Soopramanien 

The European Commission employs over 22,000 officials, with a performance 

appraisal and promotion system based on points that officials earn each year. 

A Lancaster University team worked closely with EC officials to develop a 

new system. At the core of the work was a new manpower planning problem 

in which the appraisal system was a major focus. We combined multivariate 

analysis and simulation techniques in a novel way to produce a transparent and 

easy-to-use model. Much of the model use was as a ’tool for thinking’, and it 

was a crucial part of negotiating a new system implemented in 2009. 

3 - Quantitative Methods for a New Configuration of Territorial 

Units in a Chilean Government Agency Tender 

Process 

Guillermo Duran, Ingenieria Industrial, University of Chile, 

Republica 701, 1000, Santiago, Chile, gduran@dii.uchile.cl, 

Rafael Epstein, Cristian Martinez, Gonzalo Zamorano 

131


JUNAEB is a Chilean agency responsible for the retention of children in the 

school system. The agency manages a school meals program under which firms 

bid on supply contracts for different Territorial Units (TU). Before 2007 TUs 

attractiveness to suppliers was highly variable. This led a series of problems for 

the agency in the tender process. This work uses OR methodologies to determine 

a new TU configuration which ensures a similar attractiveness. The new 

configuration added challenges to the combinatorial tender process. In 2007 

the agency increased the service by 40%, using our proposal. 

� TB-06 


8.2.30 

DEA Application I - Banking 



Chair: Joseph Paradi, Chemical Engineering and Applied Chemistry, 

Univeresity of Toronto, 200 College Street, M5S3E5, Toronto, 

Ontario, Canada, paradi@mie.utoronto.ca 

1 - Efficiency analysis of the banks and insurance companies 

operating in Poland using the DEA method 

Anna Matuszyk, Finance and Banking Collegium, Warsaw 

School of Economics, Niepodleglosci 162, 02-554, Warsaw, 

Poland, amatuszyk@matuszyk.com, Agnieszka K. Nowak 

The purpose of this article was to present the results of a study to determine the 

relative effectiveness of selected groups of commercial banks and insurance 

companies. The period of study covered years 2004 - 2008. The models were 

input-oriented. This enabled authors to obtain results, with which the strategy 

has been determined for the selected input factors, in order the analysed entities 

could become more efficient, as well as it helped to develop benchmarks 

for them. 

2 - A DEA Incentive Plan for Branch Managers 

Yossi Hadad, Industrial Engineering and Management, SCE - 

Shamoon College of Engineering, Balik Bazel, 8 Peretz Lasker 

St., Beer Sheva 84519, Israel, 84100, Beer Sheva, Israel, 

yossi@sce.ac.il, Baruch Keren, Lea Fridman 

We propose a DEA incentive plan for branch managers. Each branch has predetermined 

values of regional variables that may affect its performance. The 

method utilizes two DEA models: the CCR model and the super-efficiency 

model. The regional variables are considered here as inputs (resources). The 

measured efficiencies are then translated into a wage incentive plan. The proposed 

plan has more fairness and it can increase the flexibility of the top management 

to swap between branch managers. The method also enables to evaluate 

the performance of each branch over periods of time. 

3 - A study on profitability and marketability of Taiwanese 

bank firms before and after the Financial Holding Company 

Act 

Yi-Kang Chen, Department of Business Administration, National 

Dong Hwa University, 1, Sec. 2, Da Hsueh Rd., ShouFeng, 

Hualien, 97401, Taiwan, R.O.C., Taiwan, 

d9532006@ems.ndhu.edu.tw, Dauw-Song Zhu, Yuang-sung 

Chen 

The purpose of this paper is to determine whether subordinate subsidiaries to 

financial institutions can improve their operating performance by establishing 

financial holding companies (FHCs) in Taiwan. Specifically, this paper uses 

data envelopment analysis (DEA) to measure profitability and marketability 

changes among 14 banks with subsidiary FHCs. Results show that the efficiency 

scores of 2003—2007 were lower than those of 1997—2001; a bilateral 

model shows that prior to the implementation of the FHC Act, banks demonstrated 

higher levels of efficiency than banks subsidiaries to FHCs. 

4 - Bank Branch Human Resource Allocation Efficiency 

Using DEA — Validated Against Real Management 

132 

Joseph Paradi, Chemical Engineering and Applied Chemistry, 

Univeresity of Toronto, 200 College Street, M5S3E5, Toronto, 

Ontario, Canada, paradi@mie.utoronto.ca 

The objective of this study is to develop an intelligent staffing allocation model 

in DEA context to evaluate the bank branch’s efficiency and identify the best 

and worst branches in terms of staffing levels. This proposed model is applied 

in a big Canadian bank with 977 branches. The results obtained from the proposed 

DEA model are further compared with that obtained from the Bank’s 

current internal resource optimization system. These comparison results help 

the Bank’s managers and analysts monitor and enhance their branch staff allocation 

strategy. 

� TB-07 


8.2.47 

DEA - General topics I 



Chair: Vania Sena, Aston University, B4 7ET, Birmingham, United 

Kingdom, v.sena@aston.ac.uk 

1 - An Assessment of International Chain-owned Hotels in 

Taiwan Using DEA 

Hsiu-Feng Yong, Department of Management, Fo Guang 

University, 26247, Yilan, shachita.lu@gmail.com, Shinn Sun 

The purpose of this study is to assess of operating performance of international 

chain-owned tourist hotels in Taiwan over the years of 1997- 2007. Data Envelopment 

Analysis and Balanced Scorecard are used to assess hotel performance 

from four perspectives in terms of financial, customer, internal processes, grow 

and learn. The important remarks are concluded. 

2 - A Productivity Study on Taiwanese Pharmaceutical Industry 

Yan Chen Li, Department of Management, Fo Guang University, 

26247, Yilan, 974428@mail.fgu.edu.tw, Shinn Sun 

The purpose of this study is to assess the productivity, productivity change of 

18 pharmaceutical firms in Taiwan over 2001-2009 and examine the effects of 

environmental factors on productivity. The important findings are presented. 

3 - A Performance Study on Energy Use of International 

Tourism Hotels in Taiwan 

Tz-Chiang Li, Department of Management, Fo Guang 

University, 26247, Yilan, 974417@gmail.com, Shinn Sun 

The purpose of this paper is to assess the performance of energy use, productivity 

growth of 58 international tourist hotels in Taiwan from 1997 to 2007; and 

to examine effects of environmental variables on the performance of energy 

use. Tbe important conclusions are presented. 

4 - A Performance Study on The Public Junior High 

Schools in Taipei County 

Chin Fang Chien, Department of Management, Fo Guang 

University, 26247, Yilan, rtes02@gmail.com, Shinn Sun 

The purpose of this study is to assess educational performance of elementary 

schools in Taipei County over the years of 2005-2008. This study examined 

school perforamce in terms of overall perforamce, individual perforamce, and 

productivity growth over time. The important remarks regrading school perforamce 

are concluded. 

� TB-08 


6.1.36 

Airport operations scheduling 



Chair: Florian Jaehn, Business Administration, Management 

Information Science, Hoelderlinstrasse 3, 57068, Siegen, Germany, 

florian.jaehn@uni-siegen.de 

Chair: Erwin Pesch, FB 5, University of Siegen, Hoelderlinstr. 3, 

57068, Siegen, Germany, erwin.pesch@uni-siegen.de

1 - Integer solutions of ground holding models for air traffic 

Kevin White, Mathematics and Statistics, University of South 

Australia, U of SA, Mawson Lakes, 5095, SA, Australia, 

kevin.white@unisa.edu.au, Jerzy Filar 

When the capacity of an airport is limited due to bad weather or other emergencies, 

there are various coping strategies available to air traffic authorities, and 

various mathematical models for optimising these strategies. We consider the 

properties of the system of inequalities that arises from a popular ground holding 

binary integer linear programming model, and present some results from 

studies of air traffic in Australia. 

2 - Optimization of Runway Utilization and Communication 

Complexity 

Andrea Peter, Department of Mathematics, Research Group 

Optimization, Technische Universität Darmstadt, 

Schlossgartenstrasse 7, 64289, Darmstadt, Germany, 

apeter@mathematik.tu-darmstadt.de, Alexander Martin, 

Sebastian Pokutta 

Runways are a bottleneck in aviation. We compute an optimal schedule for 

arrivals and departures with the help of sophisticated MIP techniques in one 

single step. Up to now, a schedule for the arrivals is computed and afterwards 

the departures are filled into the gaps. Besides, we analyze an elaborated way 

of combining the independent computations of the schedules of the arrivals and 

the departures. Is it possible to link these scheduling procedures to obtain an 

optimal utilization of the runway with the help of a communication tool? We 

are investigating the communication complexity. 

3 - A Scheduling Approach to Allocate Capacity Efficiently 

at Congested Airports 

Yiannis Salouras, Department of Management Science and 


Evelpidon 47A and Lefkados 33, 11362, Athens, Greece, 

yiannis_salouras@yahoo.com, Konstantinos Zografos, Michael 

Madas 

A mathematical programming methodology is proposed for initial slot allocation 

at schedule coordinated airports. Two models are developed and assessed: 

i) one directly implementing the EU / IATA rules with the aim to model the decision 

making process of a rational schedule coordinator, and ii) another model 

taking also into account specific airside elements affecting capacity, rather than 

the abstract declared capacity notion. A comparison between the two models 

shows that the latter allows better schedules to be constructed from both the 

airlines’ and the airports’ perspective. 

4 - Aircrafts scheduling on ground: a case study 

Ludovica Adacher, DIA, Roma Tre University, via della Vasca 

Navale 79, 00146, Roma, Italy, Italy, adacher@dia.uniroma3.it, 

Marta Flamini 

In this paper we deal with the problem of scheduling aircrafts moving on 

ground, subject to safety constraints and minimizing the following three objective 

functions in lexicographical order: (i) the number of late airplanes; (ii) 

the mean waiting time at the stop bars, regulating the crossing of a runway; (iii) 

pollution and noise, in terms of the total time the aeromotors are kept on. We 

model the problem with the alternative graph, we develop several heuristic algorithms 

and we compare the preliminary results obtained by considering real 

data related to a specific case study. 

� TB-09 


6.2.53 

Recent Advances in the Theory of 




Chair: Sorin-Mihai Grad, Faculty of Mathematics, Chemnitz 

University of Technology, 09107, Chemnitz, Sachsen, Germany, 

grad@mathematik.tu-chemnitz.de 




EURO 24 - Lisbon 2010 TB-10 

1 - On Solving a Problem of Non-Smooth Function Minimization 

over a Convex Set 

Elena Morozova, Applied Mathematics Department, Herzen 

State Pedagogical University of Russia, 48, Moika Emb., 

191186, St.-Petersburg, Russian Federation, melena65@mail.ru 

A new multidimensional bisection method for minimizing non-smooth function 

over simplex is proposed for solving nonlinear constrained global quasiconcave 

minimization problem. The method does not require a differentiability of function, 

and is guaranteed to converge to the global minimum. The computational 

results are presented for a set of test problems. 

2 - Nonlinear Constrained Optimization Methods Newly 

Compared and Penalty/Barrier Methods Applied 

Teófilo Melo, Mathematics, ESTGF-IPP, Rua da Cachada - 

Margaride, 4610-250, Felgueiras, Porto, Portugal, 

tmm@estgf.ipp.pt 

In this work, nonlinear constrained optimization methods, with and without 

derivatives, are compared. Some penalty/barrier and merit functions are implemented 

to evaluate the progress of the objective function and constraints 

violation values. To promote global convergence, linesearch techniques are 

used and compared, testing Armijo, Wolfe-Powell and Goldstein criteria. 

3 - A Cross Entropy Method for Mixed Integer Programming 

Ali Eshragh Jahromi, School of Mathematics and Statistics, 

University of South Australia, Mawson Lakes Campus, 5095, 


Ali.EshraghJahromi@unisa.edu.au, Asef Nazari, Jerzy Filar 

We propose a version of the Cross Entropy (CE) method to solve a transmission 

expansion problem arising in management of national and provincial electricity 

grids. The aim is to find an expansion policy that is economical and operational. 

Often, this problem is formulated as a mixed integer nonlinear program that is 

challenging because of the presence of possibly many local optima. CE method 

shows promise in solving global optimization problems. We sample the integer 

variables using CE and solve LPs to obtain matching continuous variables. 

Numerical results demonstrate the potential. 

� TB-10 


6.2.56 

Graphs and Networks VII 



Chair: A. Ridha Mahjoub, Mathematics and Computer Science, 

LAMSADE, Universit, Place du Mar, 75775, Paris Cedex 16, 

mahjoub@lamsade.dauphine.fr 

1 - Optimizing the edge-weight over vertex-induced bipartite 

subgraphs 

Denis Cornaz, LAMSADE, Universite Paris-Dauphine, Pce Mal 

Lattre de Tasigny, 75016, Paris, France, 

cornaz@lamsade.dauphine.fr, A. Ridha Mahjoub 

Naturally, 0-1 vertex-variables allow to modelizing vertex-induced subgraphs. 

However, with these variables, a linear edge-weight objective function becomes 

quadratic. We propose a sophisticated modelization using only edge-variables 

for finding a vertex-induced bipartite subgraph with maximum edge-weight, 

and we show that this approach is competitive. 

2 - Structural Analysis for Differential-Algebraic Systems : 

Formulation and Facets. 

Sébastien Martin, LAMSADE, Université Paris-Dauphine, 

Université Paris-Dauphine, Place du Maréchal de Lattre de 

Tassigny, 75775, PARIS Cedex 16, France, 

martin@lamsade.dauphine.fr, Mathieu Lacroix, A. Ridha 

Mahjoub 

133


We consider the structural analysis problem for differential-algebraic systems 

with conditional equations. This consists, given a conditional differential algebraic 

system, in verifying if the system is well-constrained for every state 

and if not in finding a state for which the system is bad-constrained. We first 

show that the problem reduces to the perfect matching free subgraph problem 

in a bipartite graph. We give a formulation as an integer linear program of this 

problem. We also discuss the polytope of the solutions of this problem and 

propose a Branch-and-Cut algorithm. 

3 - Design of Multilayer Survivable Optical Networks 

Raouia Taktak, Lamsade, Université Paris Dauphine, 94 rue de 

Charonne, 75011, Paris, France, taktak@lamsade.dauphine.fr, 

Virginie Gabrel, A. Ridha Mahjoub 

We consider the following survivable IP-over-optical network design problem. 

Given a set of demands and two node-disjoint paths routing each demand in the 

IP layer, the problem is to find for each demand two node-disjoint paths in the 

optical layer going through the optical switches corresponding to the routers 

visited in the IP layer paths and respecting the same order. We give two integer 

programming formulations: a cut formulation and a path formulation. We 

discuss the polyhedron associated with the first one and the pricing problem of 

the latter and present some computational results. 

4 - Integer Programming Formulations for the k-Edge- 

Connected 3-Hop-Constrained Network Design Problem 

A. Ridha Mahjoub, Mathematics and Computer Science, 

LAMSADE, Universit, Place du Mar, 75775, Paris Cedex 16, 

mahjoub@lamsade.dauphine.fr, Ibrahima Diarrassouba, Virginie 

Gabrel 

In this paper, we study the k-edge-connected network design problem with 

bounded paths. We introduce new integer programming formulations for this 

problem. Then we study the polytopes associated with these formulations and 

introduce some classes of valid inequalities as well as conditions for these inequalities 

to define facets. Using these results, we devise Branch-and-Cut and 

Branch-and-Cut-and-Price algorithms for the problem and give some computational 

results. 

� TB-11 


8.2.38 

Advances in the Use of Information 

Technology I 









1 - Innovation Projects Realization Using Elements of Information 

Systems Support 

Biljana Stosic, Operations Management, Faculty of 

Organizational Sciences, Jove Ilica 154, 11000 Beograd, Serbia, 

Europe, 11000, Belgrade, Republic of Serbia, Serbia, 

biljst@fon.bg.ac.rs, Sonja Isljamović 

The paper is about innovation management supported by information system 

for continuous monitoring the realization. Starting from a concept, technical 

inventiveness and commercial exploitation of innovation can be considered as 

process/output of the process. Information and Communication Technology 

provides the possibility to gain competitive advantage. An example is given for 

project realization monitoring to enable the up-to-date insight. Developing the 

system, also the opportunity arises to have information about other phases and 

attributes of the whole project. 

2 - The Impact of Information Sharing on System Dynamics 

in a Hybrid System with a Pull-based Remanufacturing 

Process 

134 

Li Zhou, SMS, Greenwich Business School, Park Row, 

Greenwich, SE10 9LS, London, United Kingdom, 

zl14@gre.ac.uk 

This paper studies a hybrid system with joint manufacturing and remanufacturing 

process. It investigates how information sharing influences on the system 

dynamics performance. Especially, it focuses on the system capability of coping 

with uncertainties occurring in the remanufacturing/reverse process. A well 

known APIOBPCS is adopted as our inventory control strategy. The research 

methodology includes system dynamics and control theory. It concludes that 

in a pull-based remanufacturing process, information sharing does not always 

contribute to improve dynamics performance. 

� TB-12 


8.2.39 

ANP 02 


Processes 


Chair: Monica Garcia-Melon, Engineering Projects, Universidad 

Politecnica de Valencia, Camino de Vera s/n, 46022, Valencia, Spain, 

mgarciam@dpi.upv.es 

1 - Manufacturing Systems Performance Evaluation Using 

Holonic Architecture and the Analytic Network Process 

(ANP) 

M Reza Abdi, School of Management, Bradford University, 

Emm Lane, BD9 4JL, Bradford, Wet Yorkshire, United 

Kingdom, r.abdi@bradford.ac.uk, Ashraf Labib 

The paper develops holonic architecture for reconfigurable manufacturing systems 

(RMSs), which are capable of adapting to predictable and unpredictable 

changes in demands. The holonic architecture is then linked to an Analytical 

Network Process (ANP) model as a multi-criteria approach in order to evaluate 

RMSs performance. Accordingly, the complex holonic architecture is transformed 

into the fundamental components, and then grouped into a few clusters 

in accordance with their similarities or self-similarities. The model is generic 

in structure and was examined through an industrial case study in UK by using 

the SuperDecision software. 

2 - An ANP-based framework for revenue management 

Petr Fiala, Dept. of Econometrics, University of Economics 

Prague, W.Churchill Sq. 4, 13067, Prague 3, Czech Republic, 

pfiala@vse.cz 

Revenue management (RM) is the process of understanding, anticipating, and 

influencing consumer behavior in order to maximize revenue from resources. 

The challenge is to sell the right resources to the right customer at the right time 

for the right price through the right channel. RM addresses three categories of 

demand-management decisions: structural, price, and quantity decisions. In the 

paper an Analytic Network Process (ANP) framework for revenue management 

is presented.The Dynamic Network Process (DNP) is a dynamic extension of 

ANP that can deal with time dependent priorities in RM. 

3 - Selection of a Vineyard to Produce a High Quality Wine 

Using Analytic Network Process 

Monica Garcia-Melon, Engineering Projects, Universidad 

Politecnica de Valencia, Camino de Vera s/n, 46022, Valencia, 

Spain, mgarciam@dpi.upv.es, Pablo Aragones-Beltran, Alberto 

Escardino-Malva, Alfonso Porcar-Ramos, Santiago Leon-Rubio 

In this paper an application of the multicriteria decision making technique ANP 

(Analytic Network Process) is presented to select a vineyard to produce a high 

quality wine. The work has been done with the help of a wine expert, who 

acted as decision maker, and an ANP decision making specialist team. The 

whole process includes: the vineyards selection, the decision making criteria 

selection and analysis and its weighting process, the valuation of the alternatives 

and the aggregating priorities. A sensitivity analysis of the results is also 

carried out. 

4 - Prioritizing Actions for Local Suistanable Development 

with Analytic Network Process 

Monica Garcia-Melon, Engineering Projects, Universidad 

Politecnica de Valencia, Camino de Vera s/n, 46022, Valencia,

Spain, mgarciam@dpi.upv.es, Jordi Peris-Blanes, Carola 

Calabuig-Tormo, Tomas Gomez-Navarro 

Local Agenda 21 aims to institutionalize in a progressive and cross-cutting way 

the sustainable development principles and bring about change in the ways in 

which people think and value urban issues and the ways in which people, organizations 

and institutions behave. Thus, programs and actions proposed within 

LA21 have to be considered as outcomes whose goals are not only to directly 

impact local sustainable development issues but to support institutional changes 

at different levels ranging from values and behaviors of people, to organizational 

and structural change (UN-HABITAT, 2004c). In the last years many 

Spanish municipalities have started developing their own LA21 programs. Due 

to budget restrictions they have to devote time and effort to prioritize the proposed 

programs. In this paper a new tool to prioritize programs within LA21 

processes is presented. It will contribute to provide greater consistency and 

transparency when it comes to select and publicly justify the actions to be undertaken. 

The approach is based on a multiexpert multicriteria Decision Aid 

Technique, such as Analytic Network Process (ANP). Development planning 

involves several criteria, since it includes a broad range of social, economic 

and environmental goals and incorporates the complex and diverse interactions 

amongst all the elements of the problem. On the other hand, sustainable development 

planning depends upon how policymakers and other stakeholders 

involved understand and interpret the process. Gathering and considering their 

different opinions and judgments is a difficult task intrinsic to these processes. 

ANP provides a more truthful approach for modelling complex situations such 

as making decisions about sustainability actions because it allows the general 

study of the quantitative and qualitative explanatory variables and the incorporation 

of feedback and interdependence relationships among variables. 

To demonstrate the goodness of the proposed methodology it has been applied 

to the LA21 municipality of Benetússer (Valencia, Spain). The participating 

experts coincided in appreciating the procedure proposed in this paper is useful 

and an improvement from traditional techniques 

� TB-13 


2.2.21 

Continuous Location II 






1 - Lower Bounding Approaches for the Multi-commodity 

Capacitated Multi-facility Weber Problem 

Temel Öncan, Industrial Engineering, Galatasaray University, 

Ciragan Cad., Ortaköy, 34357, ISTANBUL, Turkey, 

ytoncan@gsu.edu.tr, Mehmet Hakan Akyüz, I. Kuban Altinel 

The Multi-Commodity Capacitated Multi-facility Weber Problem (MCMWP) 

is concerned with locating I capacitated facilities in the plane satisfying the 

demands of J customers for K types of commodities subject to capacity constraints 

on commodity flows between facilities and customers. We propose 

a Lagrangean relaxation scheme where the subproblem is solved by the column 

generation approach on an equivalent Set Covering problem. Efficient 

strategies are developed to accelerate the subgradient optimization algorithm. 

Discretization strategies based on block norm approximations are also devised. 

2 - Upper Bounding Methods for the Multi-commodity Capacitated 

Multi-facility Weber Problem 

Mehmet Hakan Akyüz, Industrial Engineering, Galatasaray 

University, Ciragan Caddesi No:36 Ortakoy, 34357, Istanbul, 

Turkey, mhakyuz@gsu.edu.tr, Temel Öncan, I. Kuban Altinel 

The Multi-commodity Capacitated Multi-facility Weber Problem is concerned 

with locating I capacitated facilities on the plane and satisfying the demand 

of J customers for multiple commodities subject to facility capacity, demand 

requirement and bundle constraints while minimizing the total transportation 

costs which are assumed to be directly proportional to the distance between 

facilities and customers. We propose upper bounding methods based on alternate 

location allocation and discrete approximation strategies. Computational 

experiments are reported on randomly generated test instances. 

3 - On the ordered anti-Weber problem for any norm in R2 

Carlos Guerrero, Applied Mathematics, University of Malaga, 

ETSI Informatica y Telecomunicaciones, Campus de Teatinos, 


29071, Malaga, Spain, cguerrerog@uma.es, Juan José Saameño, 

Jose Muñoz 

A family of single obnoxious facility location problems is modelled here by 

considering the same objective function that is used in the ordered median location 

problem. They involve distances defined by any arbitrary norm and hence 

it can be used in a general framework. We prove that the solutions to these obnoxious 

location problems, restricted to a polygonal region with m vertices and 

considering n existing population centers, can be found in a set defined in terms 

of the weighted equidistant points. For many usual norms, this dominating set 

is finite and can be constructed in O(mn2+n4) 

4 - A continuous competitive facility location (and design) 

problem with variable demand 

Pilar M. Ortigosa, Computer Architecture and Electronics, 

University of Almería, Ctra. Sacramento s/n, La Cañada de San 

Urbano, 04120, Almería, Spain, ortigosa@ual.es, Juana 

López-Redondo, Aranzazu Gila-Arrondo, Jose Fernandez, I. 

Garcia 

In most competitive location models available in literature, it is assumed that 

the demand is fixed independently of the conditions of the market. However, 

demand may vary depending on prices, distances to the facilities, etc. In this 

work, a new planar competitive location and design problem with variable demand 

is presented. Using it, it is shown numerically for the first time in literature 

that the assumption of fixed demand influences the location decision 

very much. Two methods are presented to cope with the new model, an exact 

interval Branch-and-Bound method and an evolutionary algorithm 

� TB-14 


2.2.15 

Supply Chain Coordination 



Chair: Wenyih Lee, Department of Business Administration, Chang 

Gung University, 259 Wenhua 1st Rd., Kweishan, 333, Taoyuan, 

Taiwan, leewe@mail.cgu.edu.tw 

1 - The application of co-opetition in the creation of supply 

contracts between haulage companies and 3pl companies 

Panagiotis Kyriazis, operations management and ERP center, 

Athens University of Economics and Business, aiginis 28 kifisia, 

14564, Athens, Greece, kyriazisp@aueb.gr, George Ioannou 

In supply chain co-opetition, firms simultaneously compete and co-operate. We 

consider the nature of co-opetition between two firms: the haulage company is 

the supplier and invests to buy new trucks, and the buyer is the 3PL company 

who invests to develop the market before uncertainty in demand is resolved. 

We consider four different decision making structures for each company and 

discuss the optimal decision We will show that the level of investment by the 

firms depends on the nature of co-opetition between them and the level of uncertainty 

demand. 

2 - A collaborative decentralized distribution system with 

demand updates 

Ulas Ozen, Bell Labs, Alcatel-Lucent, Alcatel-Lucent„ 

Blanchardstown Industrial Park, 15, Dublin, Ireland, 

ulas.ozen@alcatel-lucent.com, Greys Sosic, Marco Slikker 

This paper studies inventory pooling coalitions within a decentralized distribution 

system consisting of a manufacturer, a warehouse, and n retailers. Two 

types of cooperation are considered: with forecast sharing and with joint forecasting. 

We show that the associated cooperative games have non-emptiness 

cores. However, in two examples, we illustrate that collaboration with forecast 

sharing might lead to bad performance, and higher forecasting accuracy might 

harm the cooperation. Finally, we focus on coordination of the entire supply 

chain. 

3 - Buyback Contracts in a Dual Channel Environment 

Murat Kaya, Sabanci Universitesi, Sabanci University, MDBF 

Orhanli Tuzla, 94305, Istanbul, Turkey, mkaya@sabanciuniv.edu 

135


We compare the coordination performances of the wholesale price and buyback 

contracts in a dual-channel environment. We consider a manufacturer selling 

through both its direct channel and also through an independent retail channel. 

The consumers consider the delivery lead time in the direct channel and 

the product availability level in the retail channel in their channel choice. We 

determine the optimal contract parameters and the optimal channel mix for the 

manufacturer under different market and product characteristics. 

4 - Coordinating the consignment policy of deteriorating 

inventory with buyer’s space restrictions 

Wenyih Lee, Department of Business Administration, Chang 

Gung University, 259 Wenhua 1st Rd., Kweishan, 333, Taoyuan, 

Taiwan, leewe@mail.cgu.edu.tw, Sheng-Pen Wang 

We consider a single-manufacturer, single-buyer supply chain problem in 

which the manufacturer produces a single deteriorating product and delivers 

it to the buyer on the basis of a consignment policy. An integrated inventory 

control model, jointly determining the manufacturer’s production batch and the 

replenishment lot size, subject to the buyer’s warehouse capacity constraint, is 

proposed to minimize the system total cost. The characteristics of the model 

and the impact of warehouse capacity on the total costs, production batch, and 

replenishment lot sizes are also discussed. 

� TB-15 


2.2.12 

Polyhedral Approaches to Routing 

Problems 



Chair: Angel Corberan, Estadistica e Investigacion Operativa, 

Universitat de Valencia, Facultat de Matematiques, Avda. Dr. 

Moliner, 50, 46100, Burjasot, Valencia, Spain, angel.corberan@uv.es 

1 - A polyhedral model for the windy clustered prizecollecting 

arc routing problem 

Carles Franquesa, Estadística Investigació Operativa, Universitat 

Politècnica de Catalunya, Barcelona, Catalunya, Spain, 

carlesfranquesa@gmail.com, Angel Corberan, Elena Fernandez, 

Jose Maria Sanchis 

Among Arc Routing Problems (ARPs), the Prize-collecting ones (PARPs) are 

those in which input data graphs have profits in some edges, beside their costs, 

that are collected only once at most, if the demand edge is serviced. In the Clustered 

Prize-collecting Arc Routing Problem (CPARP), furthermore, the whole 

connected components, clusters, of the demand subset must be completely serviced 

to get the profit. In its windy version, as usual, costs associated with the 

edges are not symmetrical. In this work a polyhedral model for the WCPARP 

is presented, including new facet defining inequalities. 

2 - Hop-indexed Circuit-based formulations for the Travelling 

Salesman Problem 

Maria Teresa Godinho, Mathematics, ESTIG-IPBEJA, Campus 

do IPBeja, Rua Pedro Soares, 7800-295, Beja, Portugal, 

mtgodinho@ipbeja.pt, Luis Gouveia, Pierre Pesneau 

We introduce a new Hop-indexed Circuit-based formulation for the TSP in 

which we consider the non necessarily simple circuit associated to each node 

as a subproblem. We discuss model enhancements and contextualize the linear 

programming (LP) relaxation of the new enhanced model with the LP relaxation 

of some of the strongest formulations know from the literature. We will 

show that among known compact formulations from the literature, the proposed 

formulation is the one with the tightest LP bound. Finally, we will show that the 

proposed formulation is quite interesting for the related and so-called cumulative 

travelling salesman problem. Computational results taken from instances 

with up to 40 nodes show that the proposed formulation provides LP gaps that 

are within 1% of the optimum. 

3 - On the Maximum Benefit Chinese Postman Problem 

136 

Angel Corberan, Estadistica e Investigacion Operativa, 

Universitat de Valencia, Facultat de Matematiques, Avda. Dr. 

Moliner, 50, 46100, Burjasot, Valencia, Spain, 

angel.corberan@uv.es, Isaac Plana, Antonio Manuel 

Rodríguez-Chía, Jose Maria Sanchis 

Here we study the Maximum Benefit Chinese Postman Problem (MBCPP). It 

is a generalization of the CPP in which, associated with each edge, a cost for 

its traversal with service, a deadhead cost for its traversal with no service and 

a set of benefits are considered. A benefit is derived from every traversal with 

service of an edge. The objective is to find a closed walk starting and ending 

at the depot with maximum net benefit. We have studied the polyhedron associated 

with the MBCPP and present a branch-and-cut algorithm for its exact 

resolution. 

� TB-16 


2.2.14 

Public Bus Transportation 



Chair: Marjan van den Akker, Information and Computing Sciences, 

Utrecht University, POBox 80089, 3508TB, Utrecht, 

marjan@cs.uu.nl 

1 - Resource scheduling in public transport - scheduling 

with similarity aspects 

Boris Amberg, Decision Support & Operations Research Lab, 

University of Paderborn, Warburger Str. 100, 33098, Paderborn, 

Germany, boris.amberg@dsor.de, Natalia Kliewer 

In public bus transport timetables usually consist of many trips that are serviced 

every day. But there are also some trips that do not repeat daily. Solving the 

corresponding resource scheduling problems day after day at minimum costs 

may then produce schedules that are completely different. As most companies 

prefer similar schedules, scheduling approaches should also consider similarity 

aspects: E.g. the similarity can be improved by using reference schedules or by 

tackling the scheduling problems of various days simultaneously. We present 

different approaches and compare their results. 

2 - Optimizing depot locations based on a public transportation 

timetable 

Marjan van den Akker, Information and Computing Sciences, 

Utrecht University, POBox 80089, 3508TB, Utrecht, 

marjan@cs.uu.nl, Han Hoogeveen, Marcel van Kooten Niekerk 

When a bus company is going to start a new area, a totally new timetable, or 

wants to reorganize, a decision may have to be made on the depot locations. 

In this talk, we consider different approaches to determine depot locations for 

a given timetable. We used a clustering heuristic as well as integer linear programming 

models and a combination of these two. We present computational 

results to compare our approaches. 

3 - Modifying Timetables for integrated schedules 

Neele Hansen, Mathematics, TU Kaiserslautern, 

Paul-Ehrlich-Str. 14, 67653, Kaiserslautern, Germany, 

neele.hansen@itwm.fraunhofer.de, Sven Krumke 

Timetabling and scheduling are two classical aspects of planning in public 

transport. While historically these problems have been considered separately, 

recently there have been a number of integrated models. We study this integrated 

situation in the following context. We are given a timetable and a feasible 

schedule. One is allowed to modify (within certain bounds) the timetable 

in order to improve upon the routing of the vehicles and the transfer quality. 

We provide hardness and algorithmic results to various versions of the problem 

(discrete, continuous modifications). 

4 - Multicriteria Optimization in Public Transportation 

Steffen Weider, Optimization, Zuse Institute Berlin, Takustr. 7, 

14195, Berlin, weider@zib.de 

Costs, operational stability, and employee satisfaction are typical objectives in 

optimization problems in public transportation. These criteria are traditionally 

simply merged into a single objective. In order to study the tradeoffs between 

competing goals, however, one needs to compute the entire Pareto curve. The 

talk discusses extensions of Lagrangean relaxation and column generation approaches 

to compute such Pareto curves for vehicle and crew scheduling problems 

in public transit.

� TB-17 


1.3.14 

Container Terminal Planning I 



Chair: Christian Bierwirth, Martin-Luther-University 

Halle-Wittenberg, 06108, Halle, Germany, 

christian.bierwirth@wiwi.uni-halle.de 

Chair: Frank Meisel, Martin-Luther-University Halle-Wittenberg, Gr. 

Steinstr. 73, 06108, Halle, Germany, frank.meisel@wiwi.uni-halle.de 

1 - Models for the Discrete Berth Allocation Problem: A 

Computational Comparison 

Jesper Larsen, Department of Management Engineering, 

Technical University of Denmark, Produktionstorvet, Building 

426, 2800, Kgs. Lyngby, Denmark, jesla@man.dtu.dk, Katja 

Buhrkal, Sara Zuglian, Stefan Ropke, Richard Lusby 

We will consider the problem of allocating arriving ships to discrete berth locations 

at container terminals. This problem is recognized as one of the most 

important processes for any container terminal. We review and describe the 

three main models of the discrete dynamic berth allocation problem, improve 

the performance of one model, and, through extensive numerical tests, compare 

all models from a computational perspective. The results indicate that a 

generalized set-partitioning model outperforms all other existing models. 

2 - How distant berth allocation models are from practice? 

In praise of human planners 

Luigi Moccia, Istituto di Calcolo e Reti ad Alte Prestazioni - 

ICAR-CNR, Consiglio Nazionale delle Ricerche, Via P. Bucci 

41C, 87036, Rende, Cosenza, Italy, moccia@icar.cnr.it 

The berth allocation models recently proposed in the literature are considerably 

richer than earlier ones. However, modeling some additional features of 

the berth planning process is very important in order to succeed in supporting 

human planners by optimization algorithms. In this talk the following features 

will be discussed: more extensive integration toward other terminal resources; 

realistic cost functions, usually non-linear, reflecting the yard-side effects of 

the berth-side choices; specific transshipment related issues; multiple objectives 

induced by the integration, etc. 

3 - Storage arrangement on the yard for transshipment 

containers by meta-heuristics 

Etsuko Nishimura, Graduate School of Maritime Sciences, Kobe 

University, Fukae, Higashinada, 658-0022, Kobe, Japan, 

e-nisi@maritime.kobe-u.ac.jp, Akio Imai 

This study is concerned with the container storage arrangement on the container 

yard for the transshipment containers, in order to carry out the ship handling 

operations efficiently. Objective function is assumed that the total service time 

from the mega-containership to feeders and the waiting time for feeder ships. 

To facilitate the solution procedure we employ the heuristics based by Genetic 

Algorithm(GA) and Lagrangian relaxation(LR). As our computational results, 

at the mega-containership early arrival, the total service times obtained by LR 

are smaller than those done by GA. 

4 - Determining the quay crane capacity for a seaport container 

terminal 

Frank Meisel, Martin-Luther-University Halle-Wittenberg, Gr. 

Steinstr. 73, 06108, Halle, Germany, 

frank.meisel@wiwi.uni-halle.de, Christian Bierwirth 

The number of quay cranes to deploy at a container terminal is a crucial decision 

regarding the achievable vessel handling times. Terminals in operation 

must take into account the liner services contracted with vessel operators when 

adjusting their crane capacity. In this talk, we present a model for identifying 

a deployment of cranes that fits best to the requirements of a given set of 

liner services. A computational study is conducted to investigate the tradeoffs 

between service quality and cost that result from alternative crane deployment 

policies. 


� TB-18 


1.3.15 

Simulation and Optimization of Networks 

under Uncertainty 











1 - N-Dimensional Volume Estimation of Convex Bodies: 

Algorithms and Applications 

Mamta Sharma, International Institute of Information 

Technology, 26/C, Hosur Road, Opposite Infosys„ Electronics 

City, Bangalore, 560100, Bangalore, Karnataka, 

mamta.sharma@iiitb.ac.in, G. N. Srinivasa Prasanna, Abhilasha 

Aswal 

Current polynomial time algorithms (e.g. Markov Sampling - Lovaz-Vempala 

2004, Laplace Transform - Lassere 2001) for estimating volume of convex bodies 

are complex and difficult to implement. We present practical approaches towards 

volume computation, for problems with moderate number of dimensions 

(10-20). Our methods use polyhedral decomposition, as well as fast sampling. 

Our methods can be used to quantify information content and uncertainty, in 

constraint regions, in a robust optimization framework. We show applications 

in supply chain management, under conditions of future uncertainty. 

2 - Extinction probabilities of decomposable branching 

processes 

Sophie Hautphenne, Computer science, Université Libre de 

Bruxelles, Boulevard du Triomphe, CP 212, 1050, Brussels, 

Belgium, shautphe@ulb.ac.be 

Many systems in biology and telecommunication may be modeled by multitype 

branching processes. When some types of individuals are unable to generate 

other types, we talk about decomposable branching processes. Types are then 

partioned into irreducible equivalence classes, and partial extinction of some 

classes is possible without the whole process becoming extinct. We establish 

criteria for partial and total extinctions, and we show that the partial extinction 

probability is the minimal nonnegative solution of a fixed point equation with 

constraints or equals the total extinction probability of a modified branching 

process. 

3 - Quantitative release planning in Extreme Programming 

Tommi Tervonen, Faculty of Economics and Business, 

University of Groningen, PO Box 800, 9700AV, Groningen, 

Netherlands, t.p.tervonen@rug.nl, Gert van Valkenhoef, Bert de 

Brock, Douwe Postmus 

Extreme Programming is an agile software development methodology defined 

through a set of practices and values. Its value is established through many 

real-life case studies, but it lacks practices for project planning. Therefore we 

provide a multiple knapsack model to assist in release planning. Our model 

gives a plan that maximizes expected business value. The plan consists of several 

sets of stories that can be implemented with decreasing probability given 

the predicted velocity distribution. We evaluate parameterization of the model 

with simulation and its application in a real-life case. 

4 - Search of the best alliance composition in a interdomain 

network 

Daniel Villa Monteiro, Labo. PRiSM, Universite de Versailles - 

Saint Quentin, 45, Av des Etats Unies, 78000, Versailles, 

Yvelines, France, dvm@prism.uvsq.fr, Thierry Mautor, 

Dominique Barth, Thierry Mautor 

137


In an interdomain network where the domains have to satisfy service requests 

with QoS guarantees, an alliance is a subset of domains where members share 

their knowledge and provide to the other members a routing (stopover) service. 

The optimization problem we focus on consists in finding the best alliance that 

gives to the members the highest increase in their service satisfaction rate and 

benefit. 

Several solutions are proposed based on heuristics or exact solution methods. 

These solutions have been tested and analyzed by simulations on realistic generated 

topologies. 

� TB-19 



Networks and Industrial Organization 



Chair: Telmo Parreira, Universidade do Porto, 4169-007 , Porto, 

Portugal, c0570091@alunos.fc.up.pt 

Chair: Fernanda A. Ferreira, Mathematics, ESEIG - Instituto 

Politécnico do Porto, R. D. Sancho I, 981, 4480-876, Vila do Conde, 

Portugal, fernandaamelia@eu.ipp.pt 

1 - Checking capacity of gas transmission network 

Dan Gugenheim, Departement of Research and Innovation, 

GDFSUEZ, 361, avenue du Président Wilson, 93211, 

Saint-Denis La Plaine, France, 

dan-externe.gugenheim@gdfsuez.com, Fabrice Chauvet 

Given a gas transmission network owned by GRTgaz, it is necessary to check 

its capability to deliver gas within a set of supply and demand constraints. The 

pressure drop resulting from the gas flow through underground pipelines has 

a non linear, non convex mathematical formulation. Furthermore, the French 

network is difficult to handle: it is a meshed network with several supply and 

demand locations and interconnexion stations. We discuss different models of 

this problem. We experiment them on small instances. Eventually, we use the 

most reliable model to solve our realistic problem. 

2 - Single-Commodity Network Design with random edge 

capacities 

Biju K. Thapalia, Department of Economics,Informatics and 

Social Science, MOLDE UNIVERSITY COLLEGE, Britvegen 

2, MOLDE, P.O. Box 2110„ N-6402 , MOLDE„ Norway, 

biju.k.thapalia@himolde.no, Teodor Gabriel Crainic, Michal 

Kaut, Stein W. Wallace 

A good network for stochastic environment requires anticipation of the uncertainty 

while designing it. This paper examines the single-commodity network 

design problem with stochastic edge capacities. The paper characterizes the 

structures of the optimal designs and compare with the deterministic counterparts. 

The paper presents new approaches to compare between the deterministic 

and stochastic design. Also suggest an alternate to achieve good design which 

works well in stochastic settings without solving a full stochastic problem. 

3 - A Hotelling Network 

Telmo Parreira, Universidade do Porto, 4169-007 , Porto, 

Portugal, c0570091@alunos.fc.up.pt, Alberto A. Pinto 

This paper develops a theoretical framework to study spatial price competition 

in a Hotelling-type network game. Each firm i is represented by a node of degree 

ki, where ki is the number of firm i’s direct competitors (neighbors). We 

investigate price competition á la Hotelling with complete and incomplete information 

about the network structure. The goal is to investigate the effects 

of the network structure and of the uncertainty on firms’ prices and profits. 

We first analyse the benchmark case where each firm knows its own degree as 

well as the rivals’ degree. Then, in order to understand the role of information 

in the price competition network, we also analyse the incomplete information 

case where each firm knows its type (i.e. number of connections) but not the 

competitors’ type. 

4 - The licensing of patents 

138 

Fernanda A. Ferreira, Mathematics, ESEIG - Instituto 

Politécnico do Porto, R. D. Sancho I, 981, 4480-876, Vila do 

Conde, Portugal, fernandaamelia@eu.ipp.pt 

We study the optimal patent licensing under Cournot duopoly where the technology 

transfer takes place from an innovative firm, which is relatively costinefficient 

in the pre-innovation stage compared to the recipient firm. We determine 

the output levels at the Nash equilibrium and the corresponding profits 

of the firms. We found that the optimal licensing arrangement often involves a 

two part tariff, fixed fee plus a linear per unit output royalty 

� TB-20 


1.3.33A 




Chair: Gleb Belov, Numerical Mathematics, TU Dresden, 01062 , 

Dresden, Germany, bg37@gmx.net 

1 - A hybrid algorithm for Constraint Order Packing 

Nikolaus Furian, Engineering- and Business Informatics, Graz 

University of Technology, Kopernikusgasse 24/III, 8010, Graz, 

Austria, Austria, nikolaus.furian@tugraz.at, Siegfried Voessner 

Constraint Order Packing extends known Bin Packing problems by new additional 

placement and order constraints. While metaheuristics produce good results 

for common Bin Packing, they are not able to take advantage of the special 

structures resulting from these constraints. We introduce a hybrid algorithm 

that is based on greedy search and nested within a network search algorithm 

with dynamic node expansion and meta logic, inspired by human intuition, to 

overrule decisions. Further we discuss bounds and show on representative test 

cases that our algorithm outperforms common approaches. 

2 - A biobjective bin packing problem with conflicts 

Ali Khanafer, INRIA Lille, France, khanafer.aly@gmail.com, 

François Clautiaux, El-Ghazali Talbi 

We consider the bin packing problem with conflicts from a multiobjective point 

of view. This problem consists in minimizing the number of conflicts violated 

and the number of bins used. The Pareto set is generated and approximated 

using an epsilon-constraint strategy by means of lower and upper bounds. The 

upper bounds are based on methods dedicated to the graph coloring problem 

and to the classical bin packing problem. The lower bounding procedure is 

based on linear programming and column generation. Our methods are validated 

experimentally on a large amount of instances. 

3 - An Efficient Algorithm for a Real-Life Case of the Variable 

Size Bin Packing Problem 

Rune Larsen, Department of Mathematics and Computer 

Science, University of Southern Denmark, Lathyrusvej 87 4. th, 

5000, Odense C, Denmark, enuren@gmail.com, Joergen 

Bang-Jensen 

Real-life problems often differ from academic problems by details that make 

the standard solution methods infeasible. In our case these details include: a 

real-time environment, the possibility of ignoring waste over a threshold on a 

single bin; and the presence of an unusually high number of possible assignments 

of items to bins. The proposed solution method is a local search algorithm 

based on a destruction-reconstruction neighborhood that recomputes to 

optimality assignments of multiple bins by a dynamic programming approach. 

4 - An algorithm for orthogonal packing using 

consecutive-ones matrices 

Gleb Belov, Numerical Mathematics, TU Dresden, 01062 , 

Dresden, Germany, bg37@gmx.net 

Given a feasible layout of orthogonal packing, it automatically represents 

a non-preemptive cumulative-resource schedule along each coordinate axis. 

Such a schedule can be modeled by a 1D cutting matrix with the additional 

consecutive-ones property (the 1’s in each row are consecutive). We discuss 

pure-combinatorial and LP-based enumeration strategies to obtain such matrices.

� TB-21 


6.2.47 

Optimization Modeling V 



Chair: Alexandra Newman, Economics and Business, Colorado 

School of Mines, 80401, Golden, Colorado, United States, 

anewman@mines.edu 

1 - A Hub Location Model for Network Design of Wagonload 

Traffic 

Julia Sender, Faculty of Mechanical Engineering, TU Dortmund, 

Leonhard-Euler-Str. 2, 44227, Dortmund, NRW, Germany, 

julia.sender@tu-dortmund.de, Uwe Clausen 

In railway logistics a special network for producing wagonload traffic is 

needed. Here a hub location model for wagonload traffic is presented. The 

aim is to determine the location, size and function of marshalling yards and 

links. This network design has special characteristics: two hub types, cost and 

capacity are considered. A stepwise cost-structure reflects the specific characteristics. 

The presented problem is formulated as a multi commodity flow 

problem. The resulting Mixed Integer Program is solved with the software 

CPlex. Several test scenarios are developed and results are presented. 

2 - A collaborative optimization approach to plant-wide 

scheduling problems 

Sleman Saliba, Corporate Research Germany, ABB AG, 

Wallstadter Str. 59, 68526, Ladenburg, Germany, 

sleman.saliba@de.abb.com, Guido Sand, Xu Chaojun 

We discuss the requirements of industrial solutions to plant-wide planning and 

scheduling problems. Subsystems of plants are often optimized locally without 

an overall coordination on plant level. We present a new optimization method 

that utilizes local scheduling algorithms and optimizes the overall schedule 

by coordinating coupling parameters. The performance is evaluated on a real 

steel plant coupling melt shops and hot rolling mills showing the coordination 

scheme’s advantage to fully decentralized or centralized scheduling algorithms 

in terms of solution quality and computational effort 

3 - Efficient Mixed Integer Programming Formulations in 

Energy and Mining 

Alexandra Newman, Economics and Business, Colorado School 

of Mines, 80401, Golden, Colorado, United States, 

anewman@mines.edu 

Instances of mixed integer programming models can be (nearly) intractable 

and/or can exhibit wildly different performance depending on the model instance. 

We discuss types of formulations that have proven to be relatively 

efficient, in particular, by using preprocessing, alternate variable definitions, 

and cuts, inter alia. We draw on models from the mining sector (both open pit 

and underground production scheduling) as well as from energy applications 

(the unit commitment model, building design). Our models are implemented in 

AMPL and solved with CPLEX. 

� TB-22 



Health Care Scheduling II 



Chair: Jens Brunner, TUM School of Management, Technische 

Universität München, Arcisstr. 21, 80333, München, Germany, 

jens.brunner@wi.tum.de 

1 - Fair optimization of the fortnightly physician schedules 

with flexible shifts 

Jens Brunner, TUM School of Management, Technische 

Universität München, Arcisstr. 21, 80333, München, Germany, 

jens.brunner@wi.tum.de, Raik Stolletz 


This research addresses a shift scheduling problem for physicians. We test a 

reduced set covering approach that requires shift templates to be generated for 

a single day and compare it to an implicit modeling technique. The objective 

is to minimize the paid out hours under the restrictions given by the labor 

agreement. Furthermore, we extend the basic flexible shift scheduling model 

by introducing preference and fairness aspects. Computational results show the 

efficiency of the reduced set covering formulation in comparison to the implicit 

modeling. 

2 - Blood ordering quantities: A model to calculate ideal 

order quantities in hospital blood banks to minimize 

wastage 

Sebastian Stanger, Chair of Business Administration and 

Logistics, University of Erlangen-Nuremberg, Lange Gasse 20, 

90403, Nuremberg, BY, Germany, 

sebastian.stanger@wiso.uni-erlangen.de, Marina Gebhard, 

Gernot Kaiser 

Human blood is a scarce and perishable resource. Outdated units of blood impose 

significant costs on the health system. Efficient management of blood 

inventories is a difficult task and order quantities in hospitals are a significant 

driver for good inventory performance and hence low wastage levels. This paper 

develops a new mathematical programming model to determine ideal order 

quantities for hospitals. The model takes account of varying remaining shelf 

lives in the hospitals’ inventories and for units delivered from suppliers as well 

as the opportunity to substitute blood groups. 

3 - Materials Handling in Hospitals: A Case Study 

Christiane Reichart, Universidade Nova de Lisboa, Rua 

Fernando Cabral N12/3esq, 1750-329, Lisbon, Portugal, 

christiane.reichart@gmx.at, Amílcar Arantes 

Warehousing costs and especially order picking costs account for a considerable 

amount of total logistics costs, what leads to the purpose of this paper 

which is to elaborate means by which the materials handling process in hospitals 

can be improved. A detailed analysis of the material storage and picking 

process at Hospital Santa Maria, in Portugal, is carried out in order to demonstrate 

best practices as well as main limitations for which solutions are developed. 

The findings are that introduction of within-aisle storage, electric 

foldable picking vehicles and pick-by-voice system in combination with the 

Japanese concept Kaizen can improve warehouse operations drastically. 

� TB-23 


6.2.49 

MOO: Network Optimization and 




Chair: Marta Pascoal, Departamento de Matemática, Universidade de 

Coimbra, INESC-Coimbra, Largo D. Dinis - Apartado 3008, 

3001-454 , Coimbra, Portugal, marta@mat.uc.pt 

1 - A reference point approach for multicriteria dial-a-ride 

problems 

Julie Paquette, Operations Management and Logistics, HEC 

Montréal, 3000, chemin de la côte-sainte-catherine, H3T 2A7, 

Montreal, Quebec, Canada, julie.2.paquette@hec.ca, 

Jean-François Cordeau, Gilbert Laporte 

Three service quality criteria are incorporated as objective, in addition to the 

cost, within a tabu search algorithm for the dial-a-ride problem. This reference 

point multicriteria solution approach is used to propose a set of non-dominated 

solutions to the manager of the service. Computational results on real-life and 

randomly generated data will be presented. 

2 - A hybrid algorithm for a service technician routing and 


Sophie Parragh, Department of Business Administration, 

University of Vienna, 1210, Vienna, Austria, Austria, 

sophie.parragh@univie.ac.at 

139


The real-world service technician routing and scheduling problem we consider 

consists in routing a given number of technicians in order to complete a given 

set of tasks within a given planning horizon. Each task demands a technician 

that disposes of the appropriate skills of at least the demanded level. Time windows, 

validity periods, and maximum shift lengths have to be respected and 

breaks have to be scheduled. In addition, two technicians’ tours may have to be 

synchronized to complete those tasks that demand two technicians. We present 

a hybrid solution algorithm for this problem. 

3 - The pipeline and valve location problem 

Marta Pascoal, Departamento de Matemática, Universidade de 

Coimbra, INESC-Coimbra, Largo D. Dinis - Apartado 3008, 

3001-454 , Coimbra, Portugal, marta@mat.uc.pt, Gilbert Laporte 

We deal with finding a location for a pipeline between 2 points of a network 

and for a set of security valves, which control the damage provoked by possible 

spills, in order to minimize the environmental impact. A labeling approach 

is proposed to determine the pipeline and valve locations simultaneously, 

optimizing an impact measure that depends on the average number of 

accidents, damages and the flow in each pipe, and restricting the number of 

valves. Computational experiments on random instances are presented to evaluate 

the method’s performance and to compare it to sequential approaches. 

4 - A new hybrid evolutionary heuristic for the bi-objective 

bus driver rostering problem 

Ana Respicio, DI - CIO, Faculdade de Ciências da Universidade 

de Lisboa, Campo Grande, Bloco C5, Piso 1, 1749-016, Lisboa, 

Portugal, respicio@di.fc.ul.pt, Margarida Moz, Margarida Pato 

The bus driver rostering problem consists of assigning to each bus driver a sequence 

of duties and days-off for a given horizon, satisfying rules imposed by 

contracts and company’s norms. Two objectives are considered: minimizing 

the maximum overtime and minimizing the roster cost. For solving this highly 

complex problem we propose a bi-objective evolutionary algorithm integrating 

a constructive random heuristic at each iteration. Computational results show 

that this algorithm outperforms previously developed evolutionary heuristics, 

producing a spreader approximation of the Pareto frontier. 

� TB-24 


6.2.50 

Algorithms in Computational Biology 


Medicine 


Chair: Paola Festa, Dept. of Mathematics and Applications, 

University of Napoli Federico II, Compl. MSA - Via Cintia, 80126, 

Napoli, Italy, paola.festa@unina.it 

1 - Cliques and related cliques: models and applications 

Pedro Martins, ISCAC, IPC and CIO, Quinta Agrícola - 

Bencanta, 3040-316, Coimbra, Portugal, pmartins@iscac.pt 

A recently proposed discretized formulation for the maximum clique problem 

revealed to be more efficient on sparse graphs than known models in the literature. 

Furthermore, the new discretized information can also be used to characterize 

other clique related problems, answering and discovering important properties 

on Sociological and Biological networks. This communication presents 

modeling aspects for those clique related problems and presents applied results 

addressing Sociological networks, Molecular Biology networks and Networks 

of Interacting Pathways. 

2 - Modeling of epidemic transmission and predicting the 

spread of pathogenic 

Tertia Delia Nova, Envir managementonment, University of 

Sumatera Utara, Jl. Ampera, Jl. Dr. Mansur, 20155, Medan, 

Indonesia, nt.delia@yahoo.com, Herman Mawengkang 

Bird flu, or sometimes called avian flu, is an epidemic caused by H5N1 virus 

that primarily affects birds such as chickens, wild water birds, etc. These infuenza 

viruses occur naturally among birds. The transmission mode of avian 

flu can occur due to the spread from one farm to another farm of chickens or 

birds. This paper addresses a transmission model of avian flu taking into account 

the factors that affect the epidemic transmission such as source of infection, 

social and natural factors and various control measures. From the model 

we estimate key parameters determining the spread of highly transmissible animal 

diseases between farms. We implement the model aimed at controlling 

such transmission between chicken farms at Padang city, the capital of West 

Sumatra Province, Indonesia. 

140 

3 - Microarray data reduction and classification 

Sílvia Pedro, School of Management, Hotel and Tourism, 

University of Algarve & CEAUL, Portugal, 

smdpedro@gmail.com, Lisete Sousa, Ana Pires 

One of the central features of microarray data is their high-dimensionality, 

being the number of variables greater than the number of observations. This 

naturally calls for the use of a dimension reduction method together with the 

classification one. In this presentation we compare the classification accuracy 

obtained applying classical principal components analysis (PCA) and robust 

PCA followed by three supervised classification methods: classification trees, 

neural networks and nearest neighbours. We applied these methods to leukemia 

and colon cancer data sets available in literature. 

� TB-25 


6.2.48 


Optimization II 



Chair: Michael Markov, Markov Processes International, 25 Maple 

Str, Suite 200, 07901, Summit, NJ, United States, 

michael.markov@markovprocesses.com 

1 - Portfolio Calibration Approach for Asset Allocation and 

Financial Optimizations 

Michael Markov, Markov Processes International, 25 Maple Str, 

Suite 200, 07901, Summit, NJ, United States, 

michael.markov@markovprocesses.com, Evgeny Bauman 

We introduce the Portfolio Calibration method in Financial Optimizations that 

results in optimal portfolios preserving their efficiency in different market scenarios. 

For each scenario the portfolio is projected onto the efficient portfolio 

set and then aggregated. Different types of projections are introduced. A robust 

efficient set Calibrated Efficient Frontier is defined. A measure of stability of 

portfolio efficiency is suggested. Using Markowitz MVO as an example, we 

show that the Calibration method produces more stable results than both the 

original MVO and other approaches. 

2 - Minimum Risk Methodology for Portfolio Selection 

Konstantinos Kiriakopoulos, DEPARTMENT OF REGIONAL 

ECONOMICS, UNIVERSITY OF CENTRAL GREECE, 

Greece, k_kiriak@otenet.gr, George Kaimakamis, Theodoros 

Mavralexakis 

This paper aims to propose a new methodology for selecting ex-ante minimum 

risk portfolios. We consider different selection criteria such as the risk measure, 

rebalancing frequency, historical window and the inclusion or not of short 

sales. The methodology finds optimal portfolios for each selection criterion in 

separate and in total. The testing period is the recent financial crisis period. 

The portfolio comparison and the analysis performed gave as some interesting 

results regarding the optimal measures and we found that the portfolio constructed 

using these measures outperforms in all our data sets the indexing and 

equal weight strategy. 

3 - Multiperiod Portfolio Optimization: Comparing Approximate 

Dynamic Programming with Stochastic Programming 

Methods 

Dimitrios Karamanis, Operational Research, London School of 

Economics and Political Science, 50 Crispin Street, Lilian 

Knowles House, Room D2B3, E1 6HQ, London, United 

Kingdom, d.karamanis@lse.ac.uk, Katerina Papadaki 

In this paper we present a multiperiod portfolio optimization model that includes 

both risk and transaction costs. The objective is to find how much the 

investor should buy and sell from each security in order to maximize expected 

terminal wealth. Scenario generation and scenario reduction techniques are being 

used to represent the stochastic process of the returns with a scenario tree. 

We use approximate dynamic programming and stochastic programming methods 

to solve the model. A comparison between the two approaches is presented.

4 - Multi-Criteria Decision Making Techniques: A Stock Selection 

Application 

M.Fatih Bayramoglu, Bahcesehir University, Turkey, 

mfbayram@gmail.com, Coskun Hamzaçebi 

Stock selection is a crucial decision for investors. Investors aim to maximize 

their return while minimizing the risk. In order to make more accurate decisions, 

multi-criteria decision making (MCDM) techniques may be used. In this 

study, the usage of MCDM techniques as a stock selection tool is examined. 

� TB-26 


3.1.11 

Games solutions 



Chair: Encarnación Algaba, Applied Mathematics II, Seville 

University, C/Camino de los Descubrimientos s/n, Isla de la Cartuja., 

41092, Sevilla, Spain, ealgaba@us.es 

1 - A Non-cooperative and an Axiomatic Characterizations 

of the AL-values 

Yukihiko Funaki, Economics, Waseda University, 1-6-1 

Nishi-waseda, Shinjuku-ku, 1698050, Tokyo, Japan, 

funaki@waseda.jp, Takumi Kongo, Mariana Rodica Branzei, 

Stef Tijs 

We give a non-cooperative and an axiomatic characterizations of the AL-value 

on the class of balanced games. Here the AL-value is the weighted average of 

some vertices of the core. In both characterizations, average consistency property 

plays an essential role, and the property is obtained by the Davis-Maschler 

consistency property of the leximals. 

2 - The restricted prenucleolus 

Ilya Katsev, St. Petersburg Institute for Economics and 

Mathematics, Russian Academy of Sciences, Tchaikovsky st. 1, 

191187 St. Petersburg, Russia, 195067, Saint-Petersburg, 

Russian Federation, katsev@yandex.ru, Elena Yanovskaya 

In this talk we deal with a generalization of the prenucleolus for games with 

restricted cooperation. The restricted prenucleolus for the class of games with 

restricted cooperation is defined by a analogous way as the prenucleolus for 

classical TU games. The restricted prenucleolus may be neither single-valued, 

nor symmetric. Necessary and sufficient conditions for the collection of feasible 

coalitions providing single-valuedness of the restricted prenucleolus and 

its symmetry are given. For some particular collections axiomatic characterizations 

of the restricted prenucleolus are proposed. 

3 - Cooperative solutions on ranking situations 

Manuel A. Pulido Cayuela, Estadística e Investigación Operativa, 

University of Murcia, Campus de Espinardo. Facultad de 

Matemáticas, 30100, Murcia, Murcia, Spain, mpulido@um.es, 

Joaquin Sánchez-Soriano, Natividad Llorca, Juan Aparicio, Julia 

Sancho 

In this paper, we deal with cooperative situations arising from markets where 

an Internet search service provider offers a service listing firms in decreasing 

order according to what they bid. We introduce the corresponding TU-game. 

The core, as well as the two friendly solutions for the corners of the market, are 

described using a related assignment game. We study also the Alexia value and 

the Shapley value of this type of games. 

4 - Properties of The Myerson value in communication 

structures 

Encarnación Algaba, Applied Mathematics II, Seville University, 

C/Camino de los Descubrimientos s/n, Isla de la Cartuja., 41092, 

Sevilla, Spain, ealgaba@us.es, Jesus-Mario Bilbao, Rene van 

den Brink, Jorge López 

This paper deals with cooperative games in which partial cooperation is based 

on union stable systems. These systems have the communication situations, 

permission structures and augmenting systems as particular cases. We analyze 

the relation between the restricted game and the conference game to establish 

later which effects a union stable system has on certain desirable properties of 

these games. Basing on the properties of the position value two new characterizations 

are given for the Myerson value in this context. 


� TB-27 


8.2.06 

AIR TRANSPORTATION AND LOGISTICS 

Stream: Transportation and Logistics [c] 


Chair: Kerem Akartunali, Management Science, University of 

Strathclyde, University of Strathclyde, Dept. of Management 

Science, G1 1QE, Glasgow, United Kingdom, 

kerem.akartunali@strath.ac.uk 

1 - Efficient Delay Distribution in Air Transportation Networks 

Claus Gwiggner, ATM Modeling, Electronic Navigation 

Research Institute, 7-42-23 Jindaiji Higashi-machi, Chofu-shi, 

182-0012, Chofu, Tokyo, Japan, claus@enri.go.jp, Sakae 

Nagaoka 

A new problem in air traffic flow optimization is queue management, where 

delays have to be distributed efficiently among the network. Minimizing fuel 

depends on error propagation due to trajectory prediction uncertainties. The 

optimal distribution strategy is known in a few simple cases. 

This talk extends our results on single-server queues to small networks, where 

we analyze delay propagation as a function of control policies. We identify experimentally 

the feasibility of the problem and show our first analytical results. 

2 - Optimally Allocating Boat Resources at the United 

States Coast Guard 

Michael Wagner, Saint Mary’s College of California, United 

States, mrw2@stmarys-ca.edu, Zinovy Radovilsky 

We detail the results of an applied research project where we optimally allocate 

a fleet of boats for the United States Coast Guard. The fleet consists of hundreds 

of boats of different types, which needs to be allocated to approximately 

200 stations on both coasts of the United States, as well as Hawaii and Alaska. 

The basic model is a mixed integer program with novel constraints; for example, 

the model determines the best way to "share’ certain limited boat resources 

under geographical distance restrictions. Probabilistic enhancements result in 

an integer second order cone program. 

3 - Flow Variation with Savings in Average Cost 

Sonia , Decision Sciences Group, Indian Institute of 

Management, Prabandh Nagar, Off Sitapur Road, 226013, 

Lucknow, India, sonia@iitdalumni.com, Ankit Khandelwal 

The paper discusses the cases of flow enhancement and curtailment in a capacitated 

minimum cost network flow problem (MCFP) along with a saving in 

average shipment cost. These cases are termed as pseudo paradox and reverse 

pseudo paradox respectively. Conditions governing the existence of both types 

of paradoxes are identified and illustrated with the help of numerical examples. 

An equivalent standard MCFP is also formulated whose optimal solution 

provides the best paradoxical solution pertaining to each case. 

4 - Airline Schedule Design: Network Design Optimization 

and Heuristics Ideas 

Kerem Akartunali, Management Science, University of 

Strathclyde, University of Strathclyde, Dept. of Management 

Science, G1 1QE, Glasgow, United Kingdom, 

kerem.akartunali@strath.ac.uk, Natashia Boland, Ian Evans, 

Mark Wallace, Hamish Waterer, Olivia Smith 

Airlines plan and schedule their activities by solving a sequence of separate 

problems, where generating a timetable of flights is the first step in the process. 

Therefore, schedule design is very crucial for any airline with significant savings 

potential. We study an integrated model of the schedule design and fleet 

assignment problems for creating efficient timetables. We propose a multicommodity 

network formulation, as well as a variety of exact and heuristics 

methods to solve it. We conclude with computational results and a discussion 

of future directions. 

141


� TB-28 



Stochastic Programming Models 2 



Chair: David Wozabal, Business Adminstration, University of 

Vienna, Bruenner Str. 72, A-1210, Vienna, Austria, 

david.wozabal@univie.ac.at 

Chair: John Tomlin, Yahoo! Research, 2821 Mission College Blvd, 

95054, Santa Clara, CA, United States, tomlin@yahoo-inc.com 

1 - Stochastic project scheduling: a financial viewpoint for 

the gate setting problem 

Bernardo Pagnoncelli, Escuela de Negocios, Universidad Adolfo 

Ibañez, DIAGONAL LAS TORRES 2640 PEÑALOLÉN Oficina 

533 C Santiago, Chile, 75000000, Santiago, Chile, 

bernardo.pagnoncelli@uai.cl, Nicole Suclla Fernandez, 

Alexandre Street 

We study the problem of setting gates for activities in stochastic project 

scheduling, where each activity has random duration. We propose a two-stage 

risk-averse stochastic mixed-integer programming model where gates play the 

role of first stage decisions. The formulation goes beyond the standard riskneutral 

maximization, using the Conditional Value-at-Risk as a risk-adjusted 

objective function. A discussion over the applicability of such model as a strategic 

bidding tool for sealed-bid auctions will be provided based on the financial 

interpretations of the proposed objective function. 

2 - Medium-term Planning of a Multi-product Batch Plant 

under Evolving Multi-period Multi-uncertainty by Means 

of a Moving Horizon Strategy 

Jian Cui, Biochemical and Chemical Engineering, TU 

Dortmund, Emil-Figge-Str. 70, 44227, Dortmund, Germany, 

rossicui@hotmail.com, Sebastian Engell 

In this contribution, two-stage stochastic mixed integer programming with recourse 

(2S-MILP) is employed for medium-term planning of a multi-product 

batch plant under evolving multi-period multi-uncertainty (MPMU) with discrete 

probability distributions. A dynamic 2S-MILP formulation is developed 

for these evolving MPMU and the corresponding method, a moving horizon 

strategy is proposed. In order to reduce the computation effort, the near future 

is modeled by a tree of scenarios of future uncertainties whereas the remote 

future is represented by the expected values (EVs). 

3 - A Probabilistic Provider Selection and Task Allocation 

Model in Telecommunication Networks with Stochastic 

QoS Guaranties 

Hasan H. Turan, Industrial Engineering, University of Yalova, 

77100, Yalova, Turkey, hasanturan@sabanciuniv.edu, Nihat 

Kasap 

In this study, two important parameters of Quality of Service, delay and jitter 

are considered as random variables to capture the nature of stochastic network 

environment. We model provider selection and task allocation problem as an 

expected cost minimization problem with stochastic chance constraints. Since 

the model is nonlinear mixed integer problem, first stochastic model is converted 

into its deterministic equivalent and then a heuristic algorithm is proposed 

to solve. Moreover, we present a simulation study to check performance 

of solution procedure. 

4 - Sensitivity of an Inventory Allocation Model for Online 

Advertising with Stochastic Supply 

John Tomlin, Yahoo! Research, 2821 Mission College Blvd, 

95054, Santa Clara, CA, United States, tomlin@yahoo-inc.com, 

Vijay Bharadwaj 

We describe a stochastic programming variant of an inventory allocation model 

for online advertising and investigate the sensitivity of the solutions to the variance 

of the underlying supply distributions, when these are (realistically) assumed 

to be log-normal. We also describe a medium to large scale implementation 

and give computational results. 

142 

� TB-29 


8.2.11 

Applications of Boolean Functions 



Chair: Utz-Uwe Haus, FMA-IMO and MaCS, Otto-von-Guericke 

Universitaet Magdeburg, Universitaetsplatz 2, 39106, Magdeburg, 

Germany, haus@imo.math.uni-magdeburg.de 

1 - Minimal Conflicting Sets in ancestral genome reconstruction 

Tamon Stephen, Department of Mathematics, Simon Fraser 

University, 14th Floor Central City Tower, 250-13450 102nd 

Ave., V3T 0A3, Surrey, British Columbia, Canada, 

tamon@sfu.ca 

We consider problems of generating minimal obstacles (Conflicting Sets) to 

the consecutive-ones property for binary matrices used in ancestral genome reconstruction. 

We show that this problem can be reduced to a joint generation 

problem for boolean functions, and that this strategy can be helpful in discriminating 

between true and false positive ancestral syntenies in simulated and real 

data sets. 

This is joint work with Cedric Chauve, Utz-Uwe Haus and Vivija You. 

2 - MDD-Based Propagation of Among Constraints 

Willem-Jan van Hoeve, Tepper School of Business, Carnegie 

Mellon University, 5000 Forbes Avenue, 15213, Pittsburgh, PA, 

United States, vanhoeve@andrew.cmu.edu, Samid Hoda, John N. 

Hooker 

Fixed-width MDDs were introduced recently by Hooker et al. as a more refined 

alternative for the domain store to represent partial solutions to CSPs. In this 

work, we study MDD-based propagation algorithms for ‘among’ constraints. 

We show that MDD-based filtering for ‘among’ can be done in polynomial 

time, and we present an efficient heuristic propagation algorithm that can be 

applied to any ordering of the MDD. Our experimental results demonstrate that 

MDD-based propagation can dramatically reduce the search tree size and computation 

time when compared to the traditional domain store. 

3 - Discovering All Associations in Discrete Data using 

Frequent Minimally Infrequent Attribute Sets 

Utz-Uwe Haus, FMA-IMO and MaCS, Otto-von-Guericke 

Universitaet Magdeburg, Universitaetsplatz 2, 39106, 

Magdeburg, Germany, haus@imo.math.uni-magdeburg.de, Elke 

Eisenschmidt 

Associating biological categories with measured or observed attributes is a central 

challenge for discrete mathematics in life sciences. We propose a new concept 

to formalize this question: Given a binary matrix of objects and attributes, 

determine all attribute sets characterizing object sets of cardinality t that do not 

characterize any object set of size s>t. We determine how many such attribute 

sets exist, give an output-sensitive quasi-polynomial time algorithm to determine 

them, and show that k-sum matrix decompositions known from matroid 

theory are compatible with the characterization. 

4 - Computing exact solutions for row layout problems 

Miguel Anjos, Management Sciences, University of Waterloo, 

200 University Avenue West, N2L 3G1, Waterloo, Ontario, 

Canada, manjos@uwaterloo.ca, Frauke Liers 

Row layout problems occur in several engineering applications such as flexible 

manufacturing and circuit design. These problems can be formulated using 

Boolean variables within either a linear or a semidefinite programming framework. 

Recently there has been significant progress in both directions with 

respect to the special case of single row layout. We will survey these developments 

and present new results in computing exact solutions for row layout 

problems.

� TB-30 


8.2.13 

MCDA II: Group Decision 


other Issues 


Chair: Juan Carlos Leyva-Lopez, Departamento de Investigación y 

Posgrado, Universidad de Occidente, Carr. a Culiacancito Km. 1.5, 

80020, Culiacan, Sinaloa, Mexico, jleyva@culiacan.udo.mx 

1 - A Model of Consensus in Group Multicriteria Decision 

Aiding 

Juan Carlos Leyva-Lopez, Departamento de Investigación y 

Posgrado, Universidad de Occidente, Carr. a Culiacancito Km. 

1.5, 80020, Culiacan, Sinaloa, Mexico, jleyva@culiacan.udo.mx 

This paper presents an order-based consensus model in group multicriteria decision 

aiding that proceeds from consistency to consensus. It is based on the 

use of valued outranking relations to model individual and group preferences. 

A consensus measure and proximity measure are defined, the first one, guides 

the consensus process and the second one supports the group discussion phase 

of consensus process. The consensus degrees indicate how far a group of individuals 

is from the maximum consensus, and the proximity measure indicates 

how far each individual is from current consensus. 

2 - An approach for group decision with incomplete information 

Paula Sarabando, ESTGV and INESC Coimbra, Rua Antero de 

Quental, N o 199, 3000 - 033, Coimbra, Portugal, 

psarabando@mat.estv.ipv.pt, Luis C. Dias 

We address the problem in which a group of decision makers is not able to provide 

complete information about their individual preferences. In the context of 

additive multi-criterion aggregation, we consider problems with ordinal information,namely 

considering a ranking of the criterion weights. The approach 

we presented has as objective to inform decision-making by checking if there 

are optimal alternatives and if there are alternatives that can be eliminated due 

to dominance 

3 - A Group Multicriteria Decision Support System for 

Rank a Finite Set of Alternatives 

Pavel Alvarez Carrillo, Universidad de Occidente, 80020, 

Culiacán, pavel@uas.uasnet.mx, Alfonso Duarte, Juan Carlos 

Leyva-Lopez 

This paper presents an Internet-based Group Decision Support System (GDSS) 

prototype built around MCDA models, which provides support to collaborative 

decision makers in reaching a consensus when they try to solve a ranking problem 

in an asynchronous and distributed environment. It includes the sequential 

and parallel coordination modes. The first is processed with the ELECTRE III 

— MOEA methodology and the second is processed with the ELECTRE GD 

— MOEA methodology. 

� TB-31 


8.2.15 

OR and Ethics II 

Stream: OR and Ethics 


Chair: Haavard Koppang, Gimle Terrasse 3„ 0264, Oslo, Norway, 

haavard.koppang@bi.no 

1 - Modeling Distributive Justice 

John N. Hooker, Tepper School of Business, Carnegie Mellon 

University, 15213, Pittsburgh, PA, United States, 

john@hooker.tepper.cmu.edu 


This talk describes mathematical programming models for distributive justice. 

It derives sometimes surprising structural properties of optimal distributions 

that result from utilitarian and Rawlsian (lexmax) objectives. It investigates the 

extent to which they require egalitarianism when some individuals are more 

productive than others. It shows that mixed integer modeling of combined utilitarian 

and Rawlsian criteria can be highly nontrivial, and indicates how models 

can nonetheless be constructed using disjunctive modeling principles. 

2 - The compromise efficiency vs. egalitarianism among 

generations with an infinite horizon 

José Carlos R. Alcantud, Universidad de Salamanca, Facultad de 

Economía y Empresa, Campus Unamuno, E37007, Salamanca, 

Spain, jcr@usal.es 

This work concerns ethical aggregation of infinite utility streams. Position i is 

typically interpreted as the endowment of generation i. We analyze the broad 

question: In order for the social welfare to increase, the interest of how many 

generations can be respected if we intend to be ethical? Here "ethical" refers to 

verifying adequate equity axioms, and case-studies cover: Hammond Equityrelated 

principles; or extensions of restricted non-substitution; together with 

the usual Anonymity axiom. 

3 - In bed with the military: Ethics dilemmas for humanitarian 

organizations 

Bent Erik Bakken, St. Georgs vei 4, 0280 OSLO, 0280 OSLO, 

Norway, Oslo, beerikba@online.no 

Humanitarian organizations and military organizations sometimes cooperate, 

such as in current Afghanistan. A System Dynamics model is built to investigate 

the interrealtionships between such cooperation and the degree to which 

humanitarian organizations might cross boundaries for ethical behavior. It is 

shown that using a consequentialist framework for ethics, humanitarian organizations 

should cooperate with military organizations. Using other frameworks 

lead to different conclusions. 

4 - A Reflection on OR, Freedom and Responsibility 

João Clímaco, Faculdade de Economia da Universidade de 

Coimbra and INESC-Coimbra, 3004-512, Coimbra, Portugal, 

jclimaco@fe.uc.pt, José Antonio Peixoto 

There is an increasing interest in integrating OR in global and local sustainable 

development directives for assessments. This demands a public/democratic 

validation of the ways relations between Humanity and Nature are realized, 

namely, to model the interaction between theoretical and practical knowledge. 

We discuss the ethical tradition of the classical OR paradigm and the actual 

demands, arguing that, constrained by the "imperative of performance’, the analysts 

should critically update references and criteria for evaluations from a 

new perspective of freedom and responsibility. 

� TB-32 


8.2.17 

OR in Agriculture 



Chair: Marcela Gonzalez, Departamento de Modelación y Gestión 

Industrial, Universidad de Talca, Merced 437, s/n, Curicó, Región del 

Maule, Chile, mgonzalez@utalca.cl 

1 - Effects of Education and Investment in Research in 

Agricultural Production in Brazil 

Eliane Gomes, Brazilian Agricultural Research Corporation, 

Parque Estação Biológica, W3 Norte final, Asa Norte, 70770901, 

Brasília, DF, Brazil, eggomes@yahoo.com.br, Geraldo Souza, 

Rosaura Gazzola, Antonio Flavio Avila 

We use census data (1996 and 2006) to model the agricultural production at 

state level in Brazil. Cost efficiency measurements are computed using data envelopment 

analysis and the response is assessed via censored regressions. We 

study the effects of region, education and investment in agricultural research 

on the economic efficiency. Education has a strong significant effect, as well 

as investment in research. The intensity of the effects varies over regions for 

education, and is statistically the same for investment in research. 

143


2 - Optimal Multiobjective Crop Rotation Planning Using 

Evolutionary Algorithms 

Ruth Pavón Mendoza, Facultad Politécnica, Universidad 

Nacional de Asunción, sn, Asunción, Paraguay, 

rpavon@pol.una.py, Ricardo Brunelli, Christian von Lucken 

Crop rotation is the agricultural practice of planting different types of crops in 

the same land area in sequential seasons. The use of crop rotations is an effective 

practice for disease, insect and weeds control. When planning a rotation, 

a farmer wants to choose those crops that best fit the soil characteristics in order 

to obtain the maximum benefit with the least possible cost and risks. This 

work presents the crop rotation problem in a multiobjective context considering 

simultaneously economic and ecological factors. 

3 - A linear optimization approach for vegetable crop production 

planning 

Marcos Arenales, Dept of Applied Mathematic and Statistics, 

universidade de São Paulo, Av. do Trabalhador São-Carlense, 

400 - Centro - CP 668, 13560-970 , Sao Carlos, SP, Brazil, 

arenales@icmc.usp.br, Lana Mara Santos, Alysson Costa, 

Ricardo Henrique Santos 

We deal with crop rotation scheduling in the context of vegetable crop production 

under some ecological criteria. Three planning problems based on different 

practical situations are presented and a core mathematical model called the 

crop rotation scheduling model is developed. We present a general modeling 

framework, a solution methodology based on column generation technique and 

a set of computational experiments using instances based on real-world data. 

4 - A Model for Balancing Production Lines in a Packing 

Plant 

Marcela Gonzalez, Departamento de Modelación y Gestión 

Industrial, Universidad de Talca, Merced 437, s/n, Curicó, 

Región del Maule, Chile, mgonzalez@utalca.cl, Hugo Santelices 

We have developed a mixed integer programming model for balancing workers 

allocation in different stages of the production process in a packing plant. The 

results were implemented in a packing plant located in Chile. 

� TB-33 


8.2.19 

Renewable Energy Production 



Chair: Gonçalo Cardoso, Instituto Superior Técnico, Av. Rovisco 

Pais, 1041-009, Lisbon, Portugal, goncalo.cardoso@ist.utl.pt 

1 - Impact of bio-fuels in the adoption of small CHP units 

in the Portuguese building sector 

Gonçalo Cardoso, Instituto Superior Técnico, Av. Rovisco Pais, 

1041-009, Lisbon, Portugal, goncalo.cardoso@ist.utl.pt, Paulo 

Ferrao, Ana Paula Barbósa-Póvoa 

Efficient and sustainable energy supply solutions for buildings are nowadays 

an emerging problem. DER-CAM is a MILP model that addresses this problem 

and provides optimal energy supply solutions for buildings or groups of 

buildings. These include optimal selection of supply technologies, installed capacity 

and hourly dispatch. This paper introduces bio-fuels to DER-CAM and 

analyzes how its use may influence the adoption of small combined heat and 

power (CHP) units in the Portuguese building sector. Sensitivity analyses are 

made considering CO2 emissions, financial incentives and fuel pricing. 

2 - A Network Integration Oriented Model to Evaluate New 

Wind Power Generation Projects 

144 

Halil Cobuloglu, Industrial Engineering, University of Yalova, 

Suleymanbey Mah. Tukenmez Sk., Feribot Iskelesi Karsisi, 

77100, Yalova, Turkey, halil.cobuloglu@gmail.com, Ilhan Or, 

Gürkan Kumbaroglu 

In this paper, a series of optimization models oriented on the network integration 

of new wind power projects is proposed. In one group of models, the 

objective is cost minimization, while desired minimum additional capacity is 

taken as a constraint. In another group of models, the aim is to maximize installed 

wind power capacity and delivered energy. The existing high voltage 

transmission network constitutes the primary constraints. The output gives the 

choice of wind power projects to be invested and place of network connection. 

The models are run with data representing the case of Turkey. Results will be 

discussed. 

3 - Investment scenarios in new generation power grids 

across the mediterranean 

Elena Claire Ricci, Università degli Studi di Milano and 

Fondazione Eni Enrico Mattei, Italy, elenaclaire.ricci@unimi.it, 

Emanuele Massetti 

The paper aims at evaluating changes in climate-stabilization policy costs and 

in the electricity mix when Super-Grid is added to the portfolio of available 

options to reduce the electricity sector carbon footprint.We extend the 

WITCH model including the possibility for Europe to import electricity generated 

by large scale concentrated solar power plants in Middle-East and North 

Africa.The aim is to compute the optimal timing and investments, and evaluate 

the economic attractiveness by means of a long-term optimization where 

economic resources are allocated efficiently across sectors and time. 

� TB-34 


8.2.23 


Applications of Lot Sizing and Scheduling 

IV 


Quantity 


Chair: Alistair Clark, Dept of Mathematics and Statistics, University 

of the West of England, Frenchay Campus, Coldharbour Lane, BS16 

1QY, Bristol, United Kingdom, Alistair.Clark@uwe.ac.uk 

1 - An integrated blending and lot-sizing problem in a cotton 

spinning industry 

Victor C. B. Camargo, Instituto de Ciências Matemáticas e de 

Computação, Universidade de São Paulo, 14010-070, São 

Carlos, SP, Brazil, victor@icmc.usp.br, Bernardo Almada-Lobo, 

Franklina Toledo 

The Brazilian textile sector is an important generator of jobs and regional development. 

Its supply chain is composed of seven steps, which are addressed in 

a disaggregated fashion. We propose to study the integration of two production 

planning problems in a cotton spinning industry, the blending of raw material 

and the sizing of yarns lots. The yarns must be produced respecting quality 

limits, while satisfying the demand without backlog. We present a mathematic 

model to the integrated problem and its results from computational tests. 

2 - Lot-sizing and Scheduling Problems with Small Time 

Buckets 

Waldemar Kaczmarczyk, Department of Operations Research & 

Information Technology, AGH University of Science & 

Technology, ul. Gramatyka 10, 30-067, Krakow, Poland, 

waldek@agh.edu.pl 

This paper deals with two modeling issues arising in short term manufacturing 

planning. Firstly, new MIP model formulations are proposed for the PLSP with 

set-up times longer than a period. Experimental results prove that they are easier 

to solve with standard MIP methods than already known models. Secondly, 

correct setting of inventory holding cost parameters is discussed. Sometimes 

real time periods, macro-periods, are subdivided into shorter fictitious microperiods. 

In such problems inventory holding costs may be accounted either in 

all micro-periods or at the end of every macro-period. 

3 - Planning printers production by a lot-sizing and 

scheduling model 

Joao Flavio F. Almeida, UFMG, Brazil, 

joaoflavio.ufmg@gmail.com, Marcio Mariano Junior, Magno 

Silverio Campos

A real problem of lot-sizing and scheduling printer’s production was found 

in an EMS company where the setup and inventory must be optimized over 

the horizon. Simultaneous classical and reformulated capacitated models were 

evaluated for this NP-hard problem and solved by CPLEX mixed-integer programming. 

Results were compared with the past practices and improvements 

in cost, lead time and planning’s quality were evidenced, reducing to zero the 

number of delays on product’s delivery. This model will be graphically interfaced 

once managers have validated this decision support tool. 

� TB-35 


6.2.46 

Facilitated Modelling in Action Research 



Chair: L. Alberto Franco, Warwick Business School, University of 

Warwick, ORIS Group, Gibbet Hill Road, CV4 7AL, Coventry, 

United Kingdom, alberto.franco@warwick.ac.uk 

1 - A Systematic Procedure to Evaluate the Impact of Multi- 

Methodology Interventions 

Felipe Henao, Faculty of Management, ICESI University, ICESI 

University, ICESI University, Calle 18# 122 - 135, Pance, Pance, 

Cali, Valle del Cauca, Colombia, jfhenao@icesi.edu.co, L. 

Alberto Franco 

This paper discusses a systematic procedure to assess the impacts of multimethodology 

(MM) in practice. The approach uses the principles of Critical 

Theory, particularly Habermas’ notions of human cognitive interests, as the basis 

for understanding MMs’ impacts in real situations. A case study in Colombia 

is employed in order to test the approach. The case shows how MMs’ 

benefits can be categorised in terms of intervention stages and Habermas’ three 

worlds. The evidence suggests that MM can improve team members’ interactions 

and the quality of their work. 

2 - Using Action Research on the Process of Decision Support 

with VIP Analysis Software 

Alecsandra Ventura, Faculty of Economics, University of 

Coimbra, Rua Paulina Maria de Mendonça, 771 - Mangabeiras, 

57037-110, Maceió, Alagoas, Brazil, 

alecsandra.ventura@gmail.com, Luis C. Dias, João Clímaco 

This work is a contribution to suggest an Implementation Model of Decision 

Support with the VIP Analysis Software including the proposal of using Cognitive 

Maps as a Problem Structuring Method (PSM) and using Multi-Attribute 

Utility Theory (MAUT) to elaborate Additive Value Functions. 

In a real-world intervention, Action Research was selected as a research method 

because it enables the assessment of acceptability of VIP Analysis Software in 

Organizations; at the same time it makes possible the implementation of a decision 

aiding process by the researcher. 

3 - Creating a Shared Language in Multi-Disciplinary 

Teams through Facilitated Modelling: A Case from the 

Innovation Sector 

Donna Champion, Business School, Loughborough University, 

Epinal Way, LE11 3TB, Loughborough, United Kingdom, 

d.champion@lboro.ac.uk 

Inventors of new high technology platforms often lack the resources to develop 

products with commercial appeal and so they work with other business managers 

to develop their products. Multi-disciplinary teams often have to bridge 

a language gap between the different professions. This paper will report on 

the process of an innovation team working to develop a shared language and 

creating a set of performance indicators to monitor relationship development 

supported by a facilitated modelling approach employing systems maps and 

the PEArL framework. 

4 - Methodological Criteria for the Internal Validity and Utility 

of Practice Oriented Research 

Hubert Korzilius, Institute for Management Research, Radboud 

University Nijmegen, Netherlands, h.korzilius@fm.ru.nl 


For conducting practice oriented research different research strategies are available. 

In this paper a typology is developed for differentiating between practice 

oriented research strategies that are data based and participatory research strategies. 

The former category involves quantitative and qualitative data based research 

strategies. The latter category, participatory research strategies, includes 

knowledge based research and practice based research. At present, methodological 

criteria for assessing the quality of practice oriented research heavily 

rely on those developed for theory oriented research, in particular internal and 

external validity and reliability. However, we argue that for assessing the results 

of practice oriented research other criteria are necessary. In this paper, 

methodological criteria are formulated for evaluating the internal validity and 

practical utility of practice oriented research with the help of a Delphi study 

using research methodologists as experts. They agree upon the criteria of verifiability, 

comprehensibility and acceptance of the results, as well as holism. 

Moreover, different categories of participatory and data based research strategies 

are compared to these criteria. Practice based research and qualitative data 

based research are best equipped to fulfill these criteria. These findings may 

enable researchers to make a more deliberate choice for a specific research 

strategy in practice oriented research. 

� TB-36 


3.1.05 

Forecast based on fuzzy logic or neural 

networks 


Intelligence [c] 


Chair: Heinrich Rommelfanger, Economics and Business 

Administration, Goethe University, Niebergallweg 16, 65824, 

Schwalbach a. Ts., Hessen, Germany, 

Rommelfanger@wiwi.uni-frankfurt.de 

1 - Market Share Forecasting by Using Fuzzy Multiple Objective 

Piecewise Logistic Model and Delphi Method 

Jing-Rung Yu, Information Management, National Chi-Nan 

Univ., 470 Univ. Road, 545, Pu-Li, Nan-Tau, Taiwan, 

jennifer@ncnu.edu.tw, Fang-Mei Tseng 

It is very important for a firm to be able to make a prediction on their sales. 

The piecewise concept is adopted to highlight the impact of a new generation 

to the market so that the effective length of sales forecasting intervals would be 

obtained. After collecting some period data for the future by Delphi method, an 

interpolation by the fuzzy multiple objective piecewise logistic growth model is 

performed. Then the result considers quantitative and qualitative criteria would 

be achieved. Four television technologies, CRT, RP, PDP and LCD TVs, are 

demonstrated by our proposed model. 

2 - A New Approach to Forecast Neural Network-Based 

Fuzzy Time Series 

Ozer Ozdemir, Statistics, Anadolu University, Anadolu 

University Faculty of Science, Department of Statistics, 

Eskisehir, Turkey, ozerozdemir@anadolu.edu.tr, Memmedaga 

Memmedli 

Neural network-based fuzzy time series models have been applied to improve 

forecasting of fuzzy time series. Hence, we proposed a new method to improve 

forecasting with adjustment all degrees of membership in establishing fuzzy 

relationships by using neural network. Differing from previous studies, we adjust 

not only three degrees of membership but also all degrees of membership. 

A well-known time series which are enrollment data for the University of Alabama 

is chosen to demonstrate comparison results. It is found that proposed 

method outperforms the other methods proposed in the literature. 

3 - Neural Network Approach for Product Return Rate 

Forecasting in Reverse Logistics 

Gül Tekin Temur, Management Engineering, Istanbul Technical 

University, ITU Isletme Fakultesi Macka Besiktas, 34267, 

Istanbul, Turkey, temurg@itu.edu.tr, Bersam Bolat, Lütfü 

Yakupo˘glu, Volkan Yılmaz 

145


The complexity of reverse logistics problems increases due to high uncertainty 

in quantity, quality and time of product returns. In this study, quantity of returned 

products that depends on numerous variables such as environmental regulations, 

lifecycle of products, technical innovation level, etc. is predicted by 

using Artificial Neural Network (ANN) method. In this pursuit, firstly a wide 

literature is searched for product returns in reverse logistics. Then, the proposed 

model is generated based on relevant effective factors and implemented 

to an electric-electronic firm. 

4 - Prediction of Exchange Rates with Parametric and Non 

Parametric Techniques: A Comparison Study 

Slim Chokri, EMQ, ISCAE, Manouba, TUNISIA, 2010, 

Manouba, TUNISIA, Tunisia, chokri.slim@iscae.rnu.tn 

Several forecasting techniques have been proposed in order to predict exchange 

rate. We propose a parametric techniques such as autoregressive integrated 

moving average (ARIMA), vector autoregressive (VAR) and co-integration 

techniques, and nonparametric techniques such as support vector regression 

(SVR) and artificial neural networks (ANN). These have been employed to 

predict currency exchange rate between USD/TND, EURO/TND, GPB/TND 

and JPY/TND. The performance of the proposed models have been evaluated 

and have been compared. It is observed that the ANN model performs the best. 

� TB-37 


3.1.09 

Meeting the targets: policy and measures 


and Climate 


Chair: Nadia Maïzi, Center for Applied mathematics, MINES 

ParisTech, Rue C. Daunesse, BP 207, 06904, Sophia-Antipolis, 

France, nadia.maizi@mines-paristech.fr 

1 - Regional impacts of post Copenhaguen emission reduction 

pledges using TIAM-FR 

Sandrine Selosse, Centre for Applied Mathematics, MINES 

ParisTech, Rue Claude Daunesse, 06904, Sophia Antipolis, 

France, sandrine.selosse@mines-paristech.fr, Edi Assoumou, 

Nadia Maïzi 

The aim of this paper is to analyze the outcomes of different coordination 

schemes, derived from the submitted pledges, and associated to intermediate 

targets levels. The study relies on the TIAM-FR approach performing the optimization 

of the energy system in the long-term with descriptions of the technologies. 

2 - Impact of the price of CO2 certificates on the energy 

system of the EU-27 

Markus Blesl, IER, University of Stuttgart, 70565, Stuttgart, 

Germany, Markus.Blesl@ier.uni-stuttgart.de 

TIMES PanEU is used through variation of CO2 prices (for ETS and Non-ETS 

sectors), and reduction potential curves are constructed. Based on two scenarios 

runs with a reduction target of -15 % and -40 % in 2020 compared to Kyoto 

base year and the resulting CO2 prices, the frame of CO2 prices variation is set. 

Using the results of the different scenarios, the emission reduction compared 

to the case of the lowest CO2 prices (10 EUR /t in 2020) are evaluated. The 

total reduction over all sectors is displayed, with a focus on the different energy 

sectors and the reduction technologies are analysed. 

3 - Modelling technology pathways for oil use reduction in 

Swedish passenger car transport 

Martin Börjesson, Energy and Environment, Chalmers 

University of Technology, 412 96, Göteborg, Sweden, 

martin.borjesson@chalmers.se, Erik Ahlgren 

A strongly decreasing oil dependency is a key energy policy objective in many 

countries. In this study, an application of the well-established, cost-optimizing 

MARKAL model is used in order to estimate energy and technology costs associated 

with a strong reduction of oil use in Swedish passenger car transport 

until 2030. The model application, which describes the entire energy system of 

Sweden, is used with a simulating approach in which potential transport technology 

pathways are evaluated for different levels of oil use reduction and in 

different energy market and climate policy scenarios. 

146 

4 - Theory versus reality in the residential buildings sector: 

analysis of a gap and its consequences. 

Benoit Allibe, EDF R&D - CIRED, France, 

benoit.allibe@gmail.com 

Modelling of dwelling stock thermal performance is of great interest for long 

term energy planning and policy-making. Recently, the number of energy standards 

has increased but empirical evidence shows discrepancy between normative 

energy calculations and reality. However, all actors increasingly use these 

standards as they provide a simple and common language. We show how normative 

calculations can be misleading in terms of effort intensity required to 

reach quantified objectives and highlight the need for calculations based on a 

double metrics including normative and observed consumptions. 

� TB-38 


6.2.44 

Stochastic Valuation of Derivatives and 

Commodities I 



Chair: Martin Rainer, Inst. Applied Mathematics, METU Ankara, 

ENAMEC Inst., Glockengasse 15, D-97070 Würzburg, 97070, 

Würzburg, Germany, martin.rainer@enamec.de 




Chair: Javier de Frutos, Matematica Aplicada, Universidad de 

Valladolid, Facultad de Ciencias, Prado de la Magdalena s/n, 47005, 

Valladolid, Spain, javier.defrutos@gerad.ca 

1 - An Empirical Study on R&D Value Based upon Modified 

Finite Compound Option Models 

Yu Wen Lan, Finance and Banking, Lunghwa University of 

Technology and Science, No.300,Sec.1,Wanshou 

Rd.,Guishan,Taoyuan County 33306,Taiwan (R.O.C.), 33306, 

Taoyuan County, Taiwan, Taiwan, 93441007@cc.ncu.edu.tw, 

Tze Chin Huang 

Bellalah (1999) firstly incorporate the factor of information cost into an real option 

model for R&D valuation. However, the Bellalah’s model has default on 

catching the change of R&D’s payoff due to undefined events within project’s 

lifetime. Lo and Lan (2009) partly solve the problems as to incorporate exponential 

decay and Poisson event, but it cannot deal with cases of investment 

opportunity which do not exist permanently. We develop new models based 

upon the finite compound option setting in Black-Scholes fashion as the solution. 

2 - Information Costs Impacts on Real Options Valuation 

Jean-Michel Sahut, Amiens School of Management, 18 place 

Saint Michel, 80000, Amiens, jmsahut@gmail.com 

We present a simple framework for the analysis, valuation and simulation of 

several real options in the presence of shadow costs of incomplete information. 

Information costs can be viewed as sunk costs in the spirit of Merton’s model of 

capital market equilibrium with incomplete information. We incorporate these 

costs in standard discounted cash flow techniques and present the basic concepts 

of real options. We present valuation procedures and simulations for the 

values of common real options in the presence of shadow costs of incomplete 

information. 

3 - Callable Russian options with the finite maturity 

Atsuo Suzuki, Meijo University, NIjigaoka 4-3-3, Kani, Gifu, 

5090261, Japan, atsuo@urban.meijo-u.ac.jp, Katsushige Sawaki 

We consider callable Russian options with the finite maturity. Callable Russian 

option is a contract that the seller and the buyer have the rights to cancel and to 

exercise it at any time, respectively. We discuss the pricing model of callable 

Russian options when the stock pays dividends continuously. We show that the 

pricing model can be formulated as a coupled optimal stopping problem which 

is analyzed as Dynkin game.

� TB-39 


6.2.45 

Advances in Control Problem 



Chair: Robert Baier, Department of Mathematics, University of 


Germany, robert.baier@uni-bayreuth.de 

1 - Virtual Control Concept for Linear Quadratic Optimal 

Control Problems with state constraints 

Bjoern Huepping, Institut für Mathematik, 

Julius-Maximilians-Universität Würzburg, Am Hubland, 97074, 

Würzburg, Germany, 

bjoern.huepping@mathematik.uni-wuerzburg.de, Matthias 

Gerdts 

In this talk, a regularization for Linear Quadratic Optimal Control Problems 

will be presented. This approach ensures the existence of feasible solutions 

without interfering with the original control. The regularized problems can 

be solved efficiently using a function space Newton’s method. Conditions for 

convergence are presented, and numerical examples illustrate the idea of the 

algorithm and its application in a model predictive controller. 

2 - On the optimal control of a cascade of hydro-electric 

power stations 

Maria do Carmo Guedes, Matematica, FCUP, Rua do Campo 

Alegre, 687, 4169-007, Porto, Portugal, mmguedes@fc.up.pt, 

Ana Ribeiro, Georgi Smirnov, Sonia Vilela 

In modern reversible hydroelectric power stations, associated with reservoirs 

along a river basin with a cascade structure, it is possible both to turbine water 

from upstream to produce electric power and to pump from downstream to 

help to refill an upstream reservoir. Our objective is to maximize the profit of 

producing power providing at the same time some guidelines on when, how 

much and in what direction to allow the water flow. This is modeled as an optimal 

control problem with mixed constraints, solved using penalty functions 

and derivative free methods. 

3 - Optimal control of a flexible server in a queueing network 

with operation costs 

Dimitrios Pandelis, Mechanical Engineering, University of 

Thessaly, Pedion Areos, 38334, Volos, Greece, 

d_pandelis@mie.uth.gr 

We consider queueing networks with Poisson arrivals at each node. Jobs move 

from node to node or exit the network according to certain probabilities. Service 

times are exponentially distributed and linear holding costs are incurred 

by the jobs during the time they remain in the network. There are dedicated 

servers, one for each node, and one flexible server that can serve in any node. 

This flexible server incurs linear operating costs during his busy time. We use 

a Markov Decision Process formulation to derive structural properties of an 

optimal allocation strategy for the flexible server. 

� TB-41 


3.1.06 

Applications of System Dynamics Modeling 

I 



Chair: Lukas Schmid, Institute for Modeling and Simulation, FHS 

St.Gallen, University of Applied Sciences, Poststrasse 28, 9000, 

St.Gallen, Switzerland, lukas.schmid@fhsg.ch 

Chair: Katrin Hügel, Institute for Modelling and Simulation, 

University of applied science St.Gallen, Poststrasse 28, 9001, 

St.Gallen, Switzerland, katrin.huegel@fhsg.ch 


1 - Quality Assurance in Education — How Does the Selection 

of Applicants Matter? 

Peter Bradl, Business Administration, Universtity of Applied 

Sciences Wuerzburg-Schweinfurt, Muenzstr. 12, 97070, 

Wuerzburg, Germany, bradl@fh-wuerzburg.de 

Education is a serious topic for countries and their public universities — and a 

big business for private businesses. The focus of that paper is to detect how the 

selection of the students should be undertaken to assure quality in schooling 

and finally of the graduates — in accordance with the strategy of the school. 

The paper deals with these issues by providing models that show dependencies 

between different stakeholders and tries to identify some key factors of success 

in education and training. Using CLD and SFD feedback behavior is modeled 

and critical impacts are described. 

2 - Animated visualizations of structure and behavior in 

System Dynamics 

Ricardo Sotaquira Gutierrez, Facultad de Ingenieria, Universidad 

de la Sabana, Campus Puente El Comun, Chia, Cundinamarca, 

Colombia, ricardo.sotaquira@unisabana.edu.co 

Traditional visual representations of structure and behavior in System Dynamics 

(SD) are static. Time doesn’t flow in a causal map, or even in a time graph. 

The user of the model has to imagine a living phenomena based on "frozen’ 

images of structure and behavior. Inspired by the field of information visualization, 

we propose a new type of visual tools: SD animated diagrams. They 

have key differences with previous uses of animation in SD.We tested these 

animated diagrams with novice and expert users, and observed an improved 

understanding of the relationship between structure and behavior 

3 - Spreading System Dynamics to SMEs by using generic 

structures 

Katrin Hügel, Institute for Modelling and Simulation, University 

of applied science St.Gallen, Poststrasse 28, 9001, St.Gallen, 

Switzerland, katrin.huegel@fhsg.ch, Lukas Schmid 

The implementation of SD models supporting strategic decisions in SMEs is 

often marred by validation problems: The availability of historical data may 

be limited and the views of experts often have to be challenged, as psychological 

studies indicate. Generic structures offer a promising solution to this, as 

they are capable of aggregating experience from diverse cases. The value of 

this approach is also suggested by evidence from engineering and the natural 

sciences. Moreover, practicability of generic structures is indicated by a case 

study conducted with 4 companies. 

4 - Success Dynamics: An Application of System Dynamics 

to Corporate Success Logic 

Lukas Schmid, Institute for Modeling and Simulation, FHS 

St.Gallen, University of Applied Sciences, Poststrasse 28, 9000, 

St.Gallen, Switzerland, lukas.schmid@fhsg.ch, Katrin Hügel, 

Marcel Loher 

The principle of bounded rationality postulates that decision makers rely on 

mental models which have serious limitations. However, simulation models 

may complement these mental models and provide decision support. This circumstance 

was explored in the context of strategic management by addressing 

the general issue of business success. Four individual system dynamic models 

were developed according to the manager’s conception about the corporate 

success logic. In order to design the modeling process more practicable an 

application-oriented concept was derived using generic building blocks. 

� TB-42 


3.1.07 

Data Analysis and Decision Making 

Stream: Data Mining and Applications [c] 


Chair: Vadim Strijov, Computing Center of the Russian Academy of 

Sciences, Klara Zetkin 13-79A, 127299, Moscow, Russian 

Federation, strijov@ccas.ru 

Chair: Michael Khachay, Ural Branch of RAS, Institute of 

Mathematics and Mechanics, S.Kovalevskoy, 16, 620990, 

Ekaterinburg, Russian Federation, mkhachay@imm.uran.ru 

147


1 - Stable Group Purchasing Organizations 

Greys Sosic, Marshall School of Business, University of 

Southern California, Bridge Hall 401, 90089, Los Angeles, CA, 

United States, sosic@marshall.usc.edu, Mahesh Nagarajan, Hao 

Zhang 

GPOs exist in several sectors and benefit its members through quantity discounts 

and negotiation power when dealing with suppliers. However, GPOs 

may suffer from member dissatisfaction due to unfair allocations of the savings. 

We consider several allocation mechanisms and identify stable buyer alliances 

for them by using a dynamic stability concept. We look at discount schedules 

that seem to encompass a large class of practical schedules and analyze both 

exogenous and endogenous purchasing requirements of the members. 

2 - Development of Competitive Marketing Strategy using 

Conjoint Analysis and Game Theory 

Marija Kuzmanovic, Department for Operations Research and 

Statistics, Faculty of Organizational Sciences, Jove Ilica 154, 

11000, Belgrade, Serbia, mari@fon.rs, Milan Martic 

This paper explores the possibility of combining two well-know concepts, Nash 

equilibrium and Conjoint analysis, in the procedure of competitive marketing 

strategy development. In order to overcome the shortcomings of each concept 

when used individually, directions in which it is possible to combine them will 

be proposed. It will be shown that Conjoint data can be used either for specifying 

demand function in the first stage of procedure, or for pre-selection of a 

set of acceptable strategies, taking into account consumer preferences and their 

impact on market share. 

3 - Ship’s trajectory analysis in Lisbon harbor 

João Paiva, Portuguese Naval School, Portugal, 

tremoceiro.paiva@marinha.pt 

The purpose of this paper is to create a tool that gives port authorities information 

about the pattern of the maritime traffic so that they develop activities 

with better planning and security. The algorithm for the detection of intensity 

traffic zones is based on density clustering techniques. The algorithm for the 

anomalous trajectory detection is based in a method developed by Piciarelli, 

Foresti and Snidaro , they use a dynamic method for the creation of clusters, 

clusters will change with the allocation of new trajectories as time pass. 

4 - A Neighborhood-based Clustering Validity Index 

Tulin Inkaya, Industrial Engineering, Middle East Technical 

University, METU Industrial Engineering Department, 06531, 

Ankara, Turkey, tulin@ie.metu.edu.tr, Cem Iyigun 

Cluster validity indices aim to find the quality of the clusters obtained by clustering 

algorithms. In this work, a new validity index is proposed to evaluate the 

clusters with arbitrary shapes. Considering the data set as a proximity graph, 

a neighborhood structure is defined for each point in the data set. Compactness 

and separation of the clusters are determined according to the dispersion 

of points in the neighborhoods. Performance of the proposed index is tested on 

various data sets. 

� TB-43 


8.2.02 

Revenue Management III 



Chair: Claudius Steinhardt, Department of Analytics & Optimization, 

University of Augsburg, Universitätsstraße 16, 86159, Augsburg, 

Germany, claudius.steinhardt@wiwi.uni-augsburg.de 

1 - Overbooking with last-hour voluntary standby 

Leonardo Lustosa, Engenharia Industrial, Pontifícia 

Universidade Católica do Rio de Janeiro, Rua Marquês de S. 

Vicente, 225, Gávea, 22451-900, Rio de Janeiro, RJ, Brazil, 

leonardo.lustosa@gmail.com, Cristina Araneda-Fuentes 

When an overbooked flight is likely to become overcrowded, some airlines 

offer a compensation for checking-in passengers to renounce their boarding 

rights and join the standby queue for a future fly. Having accepted the deal, 

the passenger will gain an additional compensation if not boarded on the original 

flight. This policy is analyzed from the airline viewpoint and compared 

with other frequent practices. Favorable and unfavorable conditions for adoption 

of this policy are commented and ways of dimensioning its parameters are 

presented. 

148 

2 - The pull-to-center effect in newsvendor problems — behavioral 

evidence from experiments with skewed demand 

distributions 

Christian Koester, Institute of Management and Economics, 

Clausthal University of Technology, Julius-Albert-Straße 2, 

38678, Clausthal, Niedersachsen, Germany, 

christian.koester@tu-clausthal.de, Heike Schenk-Mathes 

Empirical research has found ordering behavior in newsvendor problems to 

deviate from theoretical predictions. We present an experimental study of a 

simple supplier-retailer wholesale price contract facing a skewed distribution 

of market demand. This modification of related research, that almost exclusively 

uses uniformly distributed market demand, explores if the commonly 

found "pull-to-center’ effect is present in more complex ordering scenarios. We 

furthermore examine the validity of the currently most discussed explanatory 

theories of the observed inventory behavior. 

3 - Dynamic pricing and Efficient Management of Repairs 

and Provisions Resources 

Rupal Rana, Business School, Warwick University, University of 

Warwick, cv4 7al, coventry, United Kingdom, 

Rupal.rana06@phd.wbs.ac.uk, Fernando Oliveira 

We analyse a problem faced by a telecommunication company who manage a 

free repair service and paid for service provision service. We analyze the allocation 

of teams to Provisions and Repairs and its relationship with demand 

cycles. This is a problem in which two products are interdependent (as they 

share the same resource) and in which one of the products is free (repairs) 

but subject to regulatory constraints, and the other product (provisions) can be 

priced dynamically. The importance of this article arises from the ability of 

the proposed algorithm to address the interactions between regulation and dynamic 

pricing. Our main objective is to improve the allocation of resources 

so that we can increase profits and, at the same time, keep the lead times under 

acceptable levels of quality. We attempt to improve the management of 

demand through dynamic pricing using reinforcement and the management of 

resources through the schedule of work team members and overtime work in 

the most efficient way. We have shown that, even in such an industry, the pricing 

policy can be used both to reduce the consumption of resources, to improve 

consumers’ welfare and increase the companies revenue. 

� TB-44 


8.2.03 

Simulation Decision Support in Enterprises 

Stream: Simulation Based Decision Support 


Chair: Miroljub Kljajic, Faculty for organizational sciences, 

University of Maribor, Kidriceva cesta 55, 4000, Kranj, Slovenia, 

miroljub.kljajic@fov.uni-mb.si 

1 - Simulation Based Decision Support System for Inventory 

Control with Stochastic Lead Time and Demand 

Davorin Kofjac, Laboratory of Cybernetics and DSS, University 

of Maribor, Faculty of Organizational Sciences, Kidriceva cesta 

55a, 4000, Kranj, Slovenia, davorin.kofjac@fov.uni-mb.si, 

Miroljub Kljajic, Valter Rejec 

This article describes web based decision support system for inventory optimization 

for products with stochastic lead time and demand. To ensure the 

real time inventory optimization the optimization core was implemented with 

C# programming language while the web based graphical interface was implemented 

with PHP. Several inventory control algorithms were analyzed with a 

goal of producing lower total inventory control costs than the actual costs provided 

by the observed company. Web based solution was utilized for predictive 

validation of algorithms in the real environment. 

2 - Approach to determination of manpower transition 

strategies — modeling, problems and methods 

Andrej Skraba, Faculty of Organizational Sciences, University of 

Maribor, Kidriceva cesta 55a, 4000, Kranj, Slovenia, Slovenia, 

andrej.skraba@fov.uni-mb.si, Miroljub Kljajic, Davorin Kofjac

Proposed paper will addresses modeling of strictly hierarchical transition system 

of human resources, which is generally applicable for transitions between 

particular job ranks, where only sequential transitions are allowed. The model 

will be represented in the form of state space and system dynamics. The problem 

of the oscillations in several optimization approaches will be presented. 

Determination of optimal strategies will be addressed, which define the transitions 

between particular ranks. 

3 - The simulation model for agricultural policy decision 

making with respect to organic agriculture 

Črtomir Rozman, Chair of Agricultural Economics and Rural 

Development, University of Maribor, Faculty of Agriculture and 

Life Sciences, Pivola 10, SI 2311 Hoce, -, Slovenia, 

crtomirrozman1@gmail.com, Pazek Karmen, Miroljub Kljajic, 

Andrej Skraba 

Proposed paper will address the development of a system dynamic simulation 

model. Model will be used by agricultural policy makers for organic farming 

development. The model will incorporate main variables and feedback loops 

influencing the organic farming system relevant for strategic plan of sustainable 

development of organic farming. Simulation model will be used for simulation 

of different policy scenarios for organic farming and their impact on economic 

and environmental parameters of organic production at an aggregate level. Developed 

platform will also include sensitivity analysis. 

� TB-45 


8.2.12 

Restricted and Unrestricted Clustering 

Stream: Geometric Clustering 


Chair: Peter Gritzmann, Mathematics, TU München, Arcisstr. 21, 

D-80290, Munich, Germany, gritzman@ma.tum.de 

Chair: Andreas Brieden, Universität der Bundeswehr München, 

Werner-Heisenberg-Weg 39, 85579, Neubiberg, Germany, 

andreas.brieden@unibw.de 

1 - On clustering bodies and their applications 

Peter Gritzmann, Mathematics, TU München, Arcisstr. 21, 

D-80290, Munich, Germany, gritzman@ma.tum.de, Andreas 

Brieden 

We study convex sets that occur naturally in geometric clustering, our prime 

application being the consolidation of farmland. While the underlying convex 

optimization problems are NP-hard, polynomial-time approximation algorithms 

can be devised whenever appropriate polyhedral approximations of their 

related clustering bodies are available. We give various structural results that 

lead to tight approximations and report on their practical applications. 

2 - Data Classification by Cell Decompositions 

Steffen Borgwardt, Fakultät für Mathematik, Technische 

Universität München, Boltzmannstr. 3, 85748, Garching, 

Bayern, Germany, borgwardt@ma.tum.de 

The characterization of the vertices of a special polytope yields a combinatorial 

optimization approach to data classification: For a given geometric training 

set X and fixed cluster sizes, we can efficiently calculate a k-clustering of X 

and a full cell decomposition of the geometric space so that each cluster lies in 

its own cell. Having such a cell decomposition leads to intuitive and efficient 

operations for the classification of new data points and the prediction of data 

values. 

3 - Partition problems: Optimality and Clustering 

Uriel G. Rothblum, Industrial Engineering and Mgt., Technion, 

Technion City, 32000, Haifa, Israel, rothblum@ie.technion.ac.il 

Partition problems constitute a large class of combinatorial optimization problems. 

Of particular interest are problems where it is possible to restrict attention 

to solutions that exhibit clustering properties, facilitating the solution of the partition 

problem in polynomial time. The talk will introduce a classification of 

partition problem and survey of numerous approaches to solve such problems 

by focusing on partitions that exhibit clustering properties. The main technique 

concern the study of vertices of corresponding partition polytopes. 


4 - Minimizing risk and costs - two applications of geometric 

clustering 

Andreas Brieden, Universität der Bundeswehr München, 

Werner-Heisenberg-Weg 39, 85579, Neubiberg, Germany, 

andreas.brieden@unibw.de, Peter Gritzmann 

Finding risk adequate premiums in insurance and determing a best reallocation 

of farming lots seem to be completely different tasks. This talk demonstrates 

that both problems can be attacked by geometric clustering and also reports on 

a software especially designed for applying cluster optimization in practice. 

� TB-46 


8.2.14 

Agent-Based Modeling of Electricity 

Markets 

Stream: Agent-Based Modeling 


Chair: Massimo Genoese, Institute for Industrial Production, 

University of Karlsruhe, Hertzstraße 16, 76187, Karlsruhe, Germany, 

massimo.genoese@kit.edu 

1 - Are agent-based simulations robust? The wholesale 

electricity trading case 

Augusto Rupérez Micola, Business and Economics, Universitat 

Pompeu Fabra, Ciutadella Campus, Ramón Trías Fargas 25, 

08005, Barcelona, Spain, augusto.ruperezmicola@gmail.com, 

Albert Banal-Estanol 

Agent-based computational economics is becoming widely used in practice and 

this paper explores the consistency of some of its standard techniques. As a 

particular case, we focus on prevailing wholesale electricity trading simulation 

methods. We include different supply and demand representations and 

propose Experience Weighted Attractions to include several behavioral algorithms. 

We compare the results across assumptions and to economic theory 

predictions. The match is good under best-response and reinforcement learning 

but not under fictitious play. The simulations perform well under flat and 

upward slopping supply bidding, and also for plausible demand elasticity assumptions. 

Learning is influenced by the number of bids per plant and the 

initial conditions. The overall conclusion is that agent-based simulation assumptions 

are far from innocuous. We link their performance to underlying 

features, and identify those that are better suited to wholesale electricity markets. 

2 - An Agent-Based Model of the Relationship between 

Forward and Spot Electricity Markets 

Fernando Oliveira, Operations MAnagement, ESSEC Business 

School, Cergy, France, oliveira@essec.fr 

An important research topic in electricity markets is the relationship between 

forward and spot markets. We present an evolutionary model that we use to test 

the stability of the different strategies available for the agents regarding generation 

quantity and forward trading, taking into account uncertainty. We test 

under which conditions the trading models are robust under learning and under 

which conditions they fail to converge to equilibrium. In this model the agents 

learn by exploring the environment, by trial and error, in order to improve their 

profit. 

3 - Agent-based participatory simulation of the German 

electricity market 

Massimo Genoese, Institute for Industrial Production, University 

of Karlsruhe, Hertzstraße 16, 76187, Karlsruhe, Germany, 

massimo.genoese@kit.edu, Dominik Möst, Wolf Fichtner 

We modify an existing agent-based simulation model for the German electricity 

market to integrate real participants into the decision process of the model. 

This is proposed as agent-based participatory simulation (Guyot and Honiden, 

2006). In the simulation, a spot market for electricity is modelled, where real 

participants trade representing electricity generation companies. The bids for 

the electricity demand is inelastic and generated by the software agents. The 

participants have information about demand, outages, a price forecast for electricity, 

and fuel and certificate prices. 

149


� TB-47 


8.2.16 

Stochastic Models for Service Operations I 

Stream: Stochastic Models for Service Operations 


Chair: Giorgio Romanin-Jacur, Management and Engineering, 

University of Padova, Stradella San Nicola, 3, 36100, Vicenza, Italy, 

romjac@dei.unipd.it 

1 - A location model for planning hospital networks under 

uncertain demand 

Ana Mestre, Departamento de Engenharia e Gestão, Instituto 

Superior Técnico, Av. Rovisco Pais, 1, 1049-001, Lisboa, 

Portugal, anamestre@ist.utl.pt, Mónica Oliveira, Ana Paula 


Location models have been widely used to support hospital network planning 

where decisions like opening/closing facilities are usually driven by demand. 

Future demand for hospital services is hard to predict since it depends on several 

uncertain parameters, such as on demographic projections, utilization rates 

and technological changes. Therefore uncertainty should be considered. This 

study proposes a stochastic hierarchical, multiservice and multiperiod location 

model that aims to improve equity within a National Health Service structure 

and explicitly includes future demand scenarios. 

2 - Pediatric palliative care network organization 

Giorgio Romanin-Jacur, Management and Engineering, 

University of Padova, Stradella San Nicola, 3, 36100, Vicenza, 

Italy, romjac@dei.unipd.it, Giada Aspergh, Paola Facchin, Anna 

Ferrante, Laura Visona_Dalla_Pozza 

Pediatric palliative patients are children suffering for incurable pathologies. 

They present a wide pathologies spectrum, a large lifetime distribution and 

frequent changes in their conditions. The assistance aims at obtaining a clinical 

equilibrium permitting an acceptable life, utilizing hospital and territorial 

(medical and social) structures. We build up a simulation model describing 

patients’ movements among the structures, also in case of evolving assistance 

organization; it evidences lacks in satisfying requests and competitions for hospital 

admissions with ordinary patients. 

3 - Quality and Efficiency Tradeoff in Systems with Flexible 

Workforce 

Eser Kirkizlar, School of Management, State University of New 

York - Binghamton, PO Box 6000, School of Management, 

13902, Vestal, New York, United States, eser@binghamton.edu, 

Sigrun Andradottir, Hayriye Ayhan 

We consider a tandem system with flexible servers and exponential service 

times. We study the effective dynamic assignment of servers to stations with 

the objective of maximizing the long-run average profit. We assume that a revenue 

is earned each time a job is completed and that the quality of service is 

inversely proportional to the time a job spends in the system (we capture this 

deterioration in quality with a positive holding cost). We determine the optimal 

server assignment policy for small systems and provide effective server 

assignment heuristics for larger systems. 

4 - Investigating whether it is optimal to make replenishments 

simultaneously in a dual source model 

Soheil Abginehchi, Business Studies, CORAL, Aarhus School of 

Business, Aarhus University, Fuglesangs Allé 4, DK-8210, 

Aarhus V, Denmark, soha@asb.dk, Christian Larsen 

In multiple sourcing, when supplier lead times are stochastic it makes sense to 

split any replenishment order into several smaller orders to pool lead time risks. 

In literature it is always assumed these orders are issued simultaneously to the 

suppliers. Here we let this simultaneousness assumption be relaxed. We study 

a dual source system with non-identical suppliers and model the problem as a 

semi-Markov decision model, allowing the decision maker the choice whether 

he will simultaneously issue two orders to both suppliers or he will issue the 

orders to suppliers at two different times. 

150 

� TB-48 


8.2.04 

Heuristics 2 



Chair: Ana Maria A.C. Rocha, Production and Systems, University of 

Minho, Campus de Gualtar, 4710-057, Braga, Portugal, 

arocha@dps.uminho.pt 

1 - Modified differential evolution for nonlinear global optimization 

Md. Abul Kalam Azad, Algoritmi R&D Center, University of 

Minho, Portugal, School of Engineering, Campus de Gualtar, 

4710-057, Braga, Braga, Portugal, akazad@dps.uminho.pt, Edite 

M.G.P. Fernandes 

Many real world optimization problems can be described only by nonlinear relationships, 

which introduce the possibility of multiple local minima. The task 

of global optimization is to find a point where the objective function obtains its 

smallest value. Differential evolution is a population based heuristic approach 

that has been shown to be very efficient for solving bound constrained nonlinear 

problems. To handle general constraints, we propose a modified differential 

evolution based on a constraint fitness priority-based ranking method. We test 

our method with a benchmark set of problems. 

2 - Heuristic and Exact Approaches to the Pipeline Network 

Design Problem 

Gustavo Dias, PEP, COPPE/UFRJ, Rua General Severiano, 

209/302, Botafogo, 2290-040, Rio de Janeiro, Rio de Janeiro, 

Brazil, dias.silva@gmail.com, Laura Bahiense, Virgílio José 

Martins Ferreira Filho 

This paper presents computational tools to support technical and economic 

evaluation of building pipe networks to transport fuel. The combinatorial nature 

of the network design problem coupled with the non-linear equations ruling 

flow energy balance in pipes lead to convex and non-convex MINLP problems. 

These programs are solved trough reformulated linear relaxations generated 

mainly by the RLT methodology, heuristic ad hoc algorithms and an exact OA 

algorithm. Experiments were conducted with five pseudo-real instances to assess 

robustness and efficiency of the proposed solution approaches. 

3 - EXACT AND HEURISTIC APPROACHES FOR MINLPs IN 

PRODUCTION PLANNING OF OIL REFINING 

Ormeu Coelho, COPPE-Production Engineering, Federal 

University of Rio de Janeiro, Centro de Tecnologia, Bloco F, 

Sala 103, Cidade Universitária, Ilha do Fundão, 21945-970, Rio 

de Janeiro, Rio de Janeiro, Brazil, ormeucoelho@gmail.com, 

Laura Bahiense, Virgílio José Martins Ferreira Filho 

In the present work we present exact and heuristic approaches to solve a problem 

of production planning of oil refining arising in two Brazilian refineries. 

The problem is formulated as non-convex MINLP in which discrete decisions 

are related to the selection of operational modes and the use of additional oil 

quantities. Computational experiments are conducted with real instances taken 

from these refineries. They are solved exactly by Spatial Branch-and-Bound. 

The heuristic solutions are obtained with Outer Approximation and Feasibility 

Pump, and improvements were possible via Local Branching. 

4 - A Genetic Algorithm for Stabilization of a Quadruped 

Robot Locomotion 

Lino Costa, Dept. Production and Systems, University of Minho, 

School of Engineering, Campus de Gualtar, 4710-057, Braga, 

Portugal, lac@dps.uminho.pt, Ana Maria A.C. Rocha, Cristina 

Peixoto Santos 

An optimization approach for the locomotion gaits stabilization of quadruped 

walking robots is presented. To model the locomotion of the robot dog, a motion 

architecture based on CPGs oscillators is used. The optimization problem 

addressed has seven decision variables concerning the amplitude, offsets and 

knee angles. The main goal is to minimize the body vibration, maximize the 

velocity and the wide stability margin through genetic algorithms. Several experimental 

results show the effectiveness of this proposed approach.

Tuesday, 12:20-13:40 

� TC-01 


Aula Magna 




Chair: Denis Bouyssou, Université Paris Dauphine, 

CNRS-LAMSADE, Place du maréchal de lattre de tassigny, 75775, 

Paris Cedex 16, France, bouyssou@lamsade.dauphine.fr 

1 - Bilevel programming and price optimization problems 

Martine Labbé, computer Science, Université Libre de Bruxelles, 

CP210/01, Boulevard du Triomphe, 1050, Bruxelles, Belgium, 

mlabbe@ulb.ac.be 

Consider a general pricing model involving two levels of decision-making. The 

upper level (leader) imposes prices on a specified set of goods or services while 

the lower level (follower) optimizes its own objective function, taking into account 

the pricing scheme of the leader. This model belongs to the class of 

bilevel optimization problems where both objective functions are bilinear. In 

this talk, we review this class of hierarchical problems from both theoretical 

and algorithmic points of view and then focus on two special cases. In the 

first one, tolls must be determined on a specified subset of arcs of a multicommodity 

transportation network. In this context the leader corresponds to the 

profit-maximizing owner of the network, and the follower to users travelling 

between nodes of the network. The users are assigned to shortest paths with 

respect to a generalized cost equal to the sum of the actual cost of travel plus 

a money equivalent of travel time. The second case consists in determining 

optimal prices for bundles of products given that each customer will buy the 

bundle that maximizes her/his own utility function. Among others, we present 

complexity results, identify some polynomial cases and propose mixed integer 

linear formulations for those pricing problem. 

� TC-02 


3.2.14 

Cooperative Games and Combinatorial 

Optimization 



Chair: Nelson Uhan, School of Industrial Engineering, Purdue 

University, 315 N. Grant Street, Grissom Hall 262, 47907, West 

Lafayette, Indiana, United States, nuhan@purdue.edu 

1 - cooperative games and fractional programming 

Walter Kern, Applied Math, Twente University, P.O. Box 217, 

7500 AE, Enschede, Netherlands, kern@math.utwente.nl, Xian 

Qiu 

Straightforward analysis of the so-called epsilon-core of cooperative games 

leads to certain fractional optimization problems which are usually difficult and 

largely unexplored in the mathematical programming community. We focus on 

two interesting special cases (TSP and bin packing) and investigate complexity/approximation 

issues. 

2 - Optimal and fair partitions in additively separable hedonic 

games 

Haris Aziz, Computer Science, LMU Munich, Theoretical 

Computer Science, Oettingenstr. 67, 80538, Munich, Bavaria, 

Germany, haris.aziz@gmail.com, Felix Brandt, Hans Georg 

Seedig 

We study a natural succinct representation of hedonic coalition formation 

games known as additively separable hedonic games. Among other results, 

we show that computing a partition with the maximum egalitarian or utilitarian 

social welfare is NP-hard. We also show that verifying whether a given 

partition is Pareto optimal is NP-hard even for symmetric preferences. In contrast, 

we present a simple algorithm to compute a Pareto optimal partition when 

preferences satisfy certain mild conditions. 

EURO 24 - Lisbon 2010 TC-03 

3 - Cost Sharing for Economic Lot-Sizing Problems with 

Remanufacturing Options 

Nelson Uhan, School of Industrial Engineering, Purdue 

University, 315 N. Grant Street, Grissom Hall 262, 47907, West 

Lafayette, Indiana, United States, nuhan@purdue.edu, Mohan 

Gopaladesikan 

We consider a class of cooperative games that model the cost sharing issues 

that arise from the economic lot-sizing problem with remanufacturing options. 

By investigating the properties of various mathematical programming formulations 

and relaxations for the underlying lot-sizing problem, we obtain some 

insights into the existence of cost allocations in the core and the approximate 

core of these games, as well as the algorithmic aspects of computing such cost 

allocations. 

� TC-03 


3.2.15 

Inventory and routing problems 



Chair: Yury Kochetov, Information Technology Department, 

Novosibirsk State University, Pirogova str., 2, 630090, Novosibirsk, 

Russian Federation, jkochet@math.nsc.ru 

Chair: Jonata Araujo, Master Program in Applied Informatics, 

University of Fortaleza, Av. Washington Soares, 1321, 60811-905, 

Fortaleza, CE, Brazil, ljonata@gmail.com 

1 - A deterministic annealing algorithm for the simultaneous 

routing of loaded and empty container movements 

Kris Braekers, Transportation Research Institute (IMOB), 

Hasselt University, Universiteit Hasselt, campus Diepenbeek, 

Wetenschapspark — gebouw 5, 3590, Diepenbeek, Belgium, 

kris.braekers@uhasselt.be, Gerrit Janssens, An Caris 

Our problem is to create efficient vehicle routes fulfilling both loaded and 

empty container transport requests. Based on demand and supply, optimal 

empty container allocations are determined by an allocation model. The resulting 

problem is a full truckload pickup and delivery problem with time windows. 

An initial solution is obtained by a parallel insertion heuristic. After finding a 

local optimum, several local search operators are embedded in a deterministic 

annealing algorithm to improve the solution. Results show that we are able to 

find good solutions in a small amount of time. 

2 - An Iterated Local Search to solve a Multiobjective Integrated 

Distribution Problem 

Helena Ramalhinho Lourenço, Departamento de Economía y 

Empresa, Universitat Pompeu Fabra, R. Trias Fargas 25-27, 

08005, Barcelona, Spain, helena.ramalhinho@upf.edu, Angel A. 

Juan, Rita Ribeiro 

The problems arising in the logistics of commercial distribution are complex 

and involve several players and decision levels. One important decision is the 

design of the routes to distribute the products, in an efficient and inexpensive 

way. This article studies a multi-objective multi-period routing model with two 

objectives. The first one is the minimization of the total distance of the routes 

and second one is related with strategies that tight relationships with customers, 

by having the same driver delivered to the same customers along several periods. 

We propose a metaheuristic based on the Iterated Local Search to solve 

this problem. Results of the computational experiment are reported. 

151

TC-04 EURO 24 - Lisbon 2010 

� TC-04 


3.2.13 

Industrial and city problems 



Chair: Luc Muyldermans, Business School, Nottingham University, 

Jubilee Campus, Wollaton Road, NG8 1BB, Nottingham, 

luc.muyldermans@nottingham.ac.uk 

Chair: Christophe Duhamel, LIMOS, Université Blaise Pascal, 

campus des Cézeaux, 63173, Aubière, France, 

christophe.duhamel@isima.fr 

1 - Co-collection and postponement strategies: a kerbside 

household waste collection case study 

Gu Pang, Newcastle Business School, Northumbria University, 

City Campus East, NE8 1ST, Newcastle Upon Tyne, United 

Kingdom, g.pang@northumbria.ac.uk, Luc Muyldermans 

We study co-collection and postponement strategies and their impact on the 

total routing distance in Beeston, UK. Our aim is to provide the Waste Collection 

Authorities with a broader understanding of the collection alternatives that 

may improve routing efficiency. The collection options are modelled as Single 

and Multi-Compartment Capacitated Arc Routing Problems. A Guided Local 

Search metaheuristic was developed to solve the problems. The results suggest 

that significant reductions in routing distance are possible when co-collection 

and/or postponement strategies are applied. 

2 - Ants solve the Integrated Vehicle and Crew Scheduling 

Problem in the public urban bus system of Ljubljana, 

Slovenia 

David Pas, Faculty of Mathematics, Natural Sciences and 

Information Technologies, University of Primorska, Glagoljaska 

8, 6000, Koper, Slovenia, david.pas@student.upr.si, Balazs 

David, Jozsef Bekesi, Miklos Kresz, Andrej Brodnik 

We will present a new and unconvential approach to the important integrated 

vehicle and driver scheduling problem in the public urban bus transit that is 

based on ACO. Although the predominant technique for that problem is the 

Langragian relaxation with column generation it is known to be inflexible in 

modeling real world constraints and to incur long computational times. The 

ACO meta-heuristic has been applied to real-world data provided by the public 

bus companies in Ljubljana (Slovenia) and Szeged (Hungary). Moreover, we 

will evaluate the potential gain from parallelization. 

3 - Heuristics for the mobile oil recovery problem 

Christophe Duhamel, LIMOS, Université Blaise Pascal, campus 

des Cézeaux, 63173, Aubière, France, 

christophe.duhamel@isima.fr, Joan-Manuel Ortega-Ardila, 

Andréa C. Santos 

The Mobile Oil Recovery Problem consists in organizing a fleet of vehicles to 

collect the maximal amount of oil on a field in a working day. This problem 

generalizes the VRP and thus it is NP-hard. The total travelled distance has to 

be minimized as a secondary criteria. We propose several heuristics to build 

feasible solutions. A VND-like local search is also defined. It uses add/drop 

moves and some k-opt moves. The VND is embedded into a VNS metaheuristic 

and a GRASP. The efficiency of both approaches is compared on real-size 

instances. 

4 - Solving the bus network design problem for the suburban 

area in Hong Kong by a Hybrid Genetic Algorithm 

Wai Yuen Szeto, Department of Civil Engineering, University of 

Hong Kong, Department of Civil Engineering, University of 

Hong Kong, 000, HK, HK, Hong Kong, ceszeto@hku.hk, 

Yongzhong Wu, S.c. Wong 

This paper proposes a hybrid genetic algorithm to solve the bus network design 

problem for the suburban area in Hong Kong with the objective of improving 

the number of transfers and total travel time of passengers. In the proposed 

algorithm, a genetic algorithm is used to tackle the route design problem and a 

neighborhood search heuristic is used to tackle the frequency setting problem. 

A new representation scheme and specific genetic operators are also developed. 

The proposed method reduces the number of transfers and total travel time by 

20.6% and 7.0%, respectively 

152 

� TC-05 


3.2.16 

Location under uncertainty 



Chair: Laureano Fernando Escudero, Dept. de Estadística e 

Investigación Operativa, Universidad Rey Juan Carlos, c/Tulipan, S/n, 

28933, Mostoles (Madreid), Spain, laureano.escudero@urjc.es 

1 - On the solving the multi-period location problem under 

uncertainty 

Antonio Alonso-Ayuso, Statistics & Operations Research 

Department, Rey Juan Carlos University, c/Tulipan s/n, 28933, 

Mostoles, Madrid, Spain, antonio.alonso@urjc.es, Maria 

Albareda Sambola, Laureano Fernando Escudero, Elena 

Fernandez, Celeste Pizarro Romero 

We present a framework for solving the strategic problem of timing the location 

of facilities and customers assignment in a multi-period environment under uncertainty 

in the setup and maintenance costs, customers assignment cost, and 

the periods at which each customer has demand, among others. By considering 

a compact representation of the Deterministic Equivalent Model, we specialize 

the so-called Branch-and-Fix Coordination algorithm. Computational experience 

is reported. 

2 - A stochastic location problem of k-centers on graph 

Biljana Panic, Faculty of Organizational Sciences, Belgrade, 

Serbia, bilja@fon.rs, Mirko Vujosevic, Ognjen Pantelic, Tamara 

Valok 

While most of location models have deterministic inputs, there are a lot of 

problems with stochastic nature of parameters. An original stochastic k-center 

location problem is presented in the paper. At the beginning, node reliability, 

i.e. the reliability of covering a node from k centers is defined. A new algorithm 

is proposed to calculate this reliability. The optimization problem is to 

maximize the defined criterion function as minimal node reliability. 

3 - Stochastric Set Packing Problem 

Laureano Fernando Escudero, Dept. de Estadística e 

Investigación Operativa, Universidad Rey Juan Carlos, c/Tulipan, 

S/n, 28933, Mostoles (Madreid), Spain, 

laureano.escudero@urjc.es, Mercedes Landete, Antonio Manuel 

Rodríguez-Chía 

The Set Packing Problem (SPP) undder uncertainty studied via scenario analysis. 

A maximization of a composite function of the expected value minus the 

weighted excess probability is perfomred. The splitting variable representation 

is decomposed by dualizing the nonanticipativity constraints that link the 

deterministic SPP with a 0-1 knapsack problem for each scenario under consideration. 

The Lagrange multipliers updating is performed by using the Volume 

Algorithm. 

4 - On the location of Lottery Terminals in Spain 

Daniel Serra, Economics and Business, Universitat Pompeu 

Fabra, Trias Fargas 25-27, 08005, Barcelona, Spain, 

daniel.serra@upf.edu 

In this paper we study the location of new lottery terminals in Spain. Actually, 

more than 14.000 terminals are already located, but the model presented 

hear aims at finding new market "niches" to open new outlets, considering to 

minimise the impact on the income of existing ones, and taking into acount the 

buying power of the catchment area of each outlet. An integer formulation is 

presented, together with some results. 

� TC-06 


8.2.30 

DEA Application II - Education 



Chair: Maria Portela, Rua Diogo Botelho, 1327, 4169- 005 Porto, 

Porto, Portugal, csilva@porto.ucp.pt

1 - Efficiency evaluation of distance education in Rio de 

Janeiro 

João Carlos Soares de Mello, Engenharia de Produção, 

Universidade Federal Fluminense, Rua Passo da Pátria 156, São 

Domingos, 24240310, Rio de Janeiro, RJ, Brazil, 

jcsmello@pesquisador.cnpq.br, Fernando do Valle Silva 

Andrade, Luana Carneiro Brandão, Lidia Angulo-Meza 

This work quantitatively evaluates the Undergraduate Distance Educational 

Centre of Rio de Janeiro (CEDERJ). Preliminarily, we use basic Data Envelopment 

Analyses (DEA) in order to have its results as a reference. Then, we cluster 

the different poles into homogeneous groups and use advanced evaluation 

models for each one of them and therefore obtain more suitable benchmarks 

than when using basic DEA. Finally, we consider all of the centres together 

using Non Homogeneous Decision Making Units techniques. 

2 - Efficiency Evaluation of Private teaching Institutions by 

Data Envelopment Analysis 

Abdullah Korkut Üstün, Industrial Engineering, Eskisehir 

Osmangazi University, Eskisehir Osmangazi Universitesi 

Endüstri Mühendisligi Bölümü, 26030, Eskisehir, Turkey, 

korkut.ustun@gmail.com, Ezgi Aktar Demirtas 

Data envelopment analysis (DEA) is a technique that measures the relative efficiencies 

of decision making units. In this study, we try to determine the efficiency 

scores of private teaching institutions in a city in Turkey. We also 

determine the sources of inefficiency in these institutions. 

3 - Comparative Performance of Public Universities in 

Malaysia Using Data Envelopment Analysis 

Wan Rohaida Wan Husain, Warwick Business School, The 

University of Warwick, Warwick Business School, The 

University of Warwick, CV4 7AL, Coventry, United Kingdom, 

W.R.Wan-Husain@warwick.ac.uk 

This paper presents a study into the performance of public universities in 

Malaysia. The current research involves a small sample of twenty public universities 

in Malaysia, yet six variables have been identified as relevant inputs 

and outputs. This raises a concern of the discriminatory power of DEA assessment. 

We overcome the problem of the small sample size by combining two 

approaches. First, we use the hybrid returns-to-scale (HRS) model that allows 

us to combine the features of both CRS and VRS models in one formulation and 

achieve a significant improvement in discrimination compared to the standard 

VRS model. Second, we enrich the HRS model by the use of weight restrictions. 

The latter are constructed using the trade-off approach which preserves 

the technological meaning of our DEA model. 

4 - Performance assessment of Portuguese secondary 

schools 

Maria Portela, Rua Diogo Botelho, 1327, 4169- 005 Porto, Porto, 

Portugal, csilva@porto.ucp.pt, Ana Camanho, Diogo Borges 

This paper describes a performance assessment of Portuguese secondary 

schools using data envelopment analysis. Schools are assessed through a value 

added perspective, where average grades on exit of secondary education are 

contextualised by average grades on entry on secondary education. An assessment 

of all Portuguese secondary schools was undertaken showing best 

performers and worst performers. Assurance regions were used to prevent unreasonable 

weighting schemes of exam average grades on exit (output weights 

were linked to the number of exams of the school in each course). 

� TC-07 


8.2.47 

DEA - General topics II 



Chair: Dimitris Despotis, Department of Informatics, University of 

Piraeus, 80, Karaoli & Dimitriou Street, 18534, Piraeus, Greece, 


1 - Using FDH technology to determine anchor points in 

DEA 

Abolfazl Keshvari, Math, IUST, Iran, Islamic Republic Of, 

abkeshvari@gmail.com, Mohammad Reza Alirezaee 


We present a new algorithm to determine anchor points in VRS DEA by using 

FDH model. In this algorithm, we use FDH instead of regular VRS model; and 

because FDH models can be solved by the enumeration method, the entire process 

will be much faster and simpler. Also, we discuss problems without inputs 

or outputs. In these problems, all extreme efficient DMUs are anchor points and 

we can determine anchor points by use of FDH model. Also, we present a comparison 

of algorithm in Bougnol and Dula (2009) and the proposed algorithm 

and show that the proposed algorithm requires fewer computations. 

2 - A Performance Study on Business Centers of 

Chunghwa Post Company in Taiwan 

Chin-Piao Lin, Department of Management, Fo Guang 

University, 26247, Yilan, natureosho@gmail.com, Shinn Sun 

The purpose of this paper is to assess m The purpose of this paper is to evaluate 

the performance of 23 practiced business responsibility centers of Chunghwa 

Post Company in Taiwan over the years of 2003 to 2008. Overall performance, 

service performance and productivity growth of the 13 centers are examined. 

The important remarks regarding center pergormance are concluded. 

3 - Congestion in stochastic data envelopment analysis: 

An input relaxation approach 

Luka Neralic, Faculty of Economics and Business, University of 

Zagreb, J. F. Kennedy 6, Stefaniceva 7, 10000, Zagreb, Croatia, 

lneralic@efzg.hr 

The input relaxation model and its stochastic version, recently introduced in 

Data Envelopment Analysis (DEA) literature, uses more flexibility in changes 

of the used input combination to find maximum possible output and can be 

useful to resource management. We study congestion issues in this setting. 

Deterministic equivalent to the stochastic congestion model is obtained. The 

deterministic equivalent is typically non-linear. It is, however, shown that uder 

fairly general conditions this non-linear model can be replaced by an ordinary 

deterministic DEA model. When allowable limits of data variations for evaluating 

decision making unit is permitted, sensitivity analysis is studied. 

4 - Banknote Printing at Modern Central Banking: Trends, 

Costs and Efficiency 

Miguel Sarmiento, International and Monetary Affairs Division, 

Central Bank of Colombia, Carrera 11B # 119-36 Apto 302, 

4152, Bogota, Colombia, nsarmipa@banrep.gov.co 

This paper examines trends in banknote printing during the 2000-2005 period 

for a cross-section of 56 central banks. A cost function using a panel data model 

with random effects was estimated. Based on these results, a DEA model was 

used to measure technical cost efficiency and changes in productivity. It was 

found that the technical efficiency of most central banks increased during the 

period, especially when the private sector became involved. The Malmquist 

Index showed a moderate increase in productivity, mainly due to increases in 

scale efficiency rather than technical change. 

� TC-08 


6.1.36 

Production Scheduling 



Chair: Dirk Briskorn, Department for Supply Chain Management and 

Operations Management, University of Cologne, 

Albertus-Magnus-Platz, 50923, Köln, Germany, 

briskorn@wiso.uni-koeln.de 

1 - Yield vs. Flow-Time Tradeoff in a Production System 

Israel Tirkel, Industrial Engineering & Management, Ben-Gurion 

University, 8/3 Ha’Zamir St., 90805, Mevasset Zion, 

tirkel@bgu.ac.il, Gad Rabinowitz 

Modern industry struggles with Yield and Flow-Time (FT) tradeoff due to their 

impact on profit, but considers them separately. Our study originates in semiconductors 

but applies to other industries. Analytical and simulation models 

are developed to investigate the impact of inspections on Yield and FT in a 

deteriorating production system. Results indicate Yield and FT increase with 

growing inspection rate until Yield reaches a maximum and starts to decline, 

while FT continues to increase. Decision support tool is provided for applying 

inspections while simultaneously considering Yield and FT. 

153


2 - Rolling stock maintenance planning 

Francis Sourd, Innovation & Research, SNCF, Paris, France, 

francis.sourd@sncf.fr, Mathilde Carlier, Raphaël Even 

Rolling stock maintenance consists of a series of periodic operations that are to 

be scheduled with respect to maximal time-lags. In order to optimize maintenance 

costs, each operation must be scheduled as late as possible. In order to 

model real-life problems, we use a resource consumption model derived from 

the resource-constrained scheduling problem (RCPSP) and a generic representation 

of the maintenance cycles. The model is solved with a heuristic algorithm 

which is based on a large neighborhood search. 

3 - Scheduling flexible maintenance actitivies on a single 

machine 

Dirk Briskorn, Department for Supply Chain Management and 



briskorn@wiso.uni-koeln.de, Stefan Bock, Andrei Horbach 

We focus on single machine scheduling subject to machine deterioration. The 

maintenance level specifies the machine’s current maintenance state. While 

jobs are processed the maintenance level drops by a certain — possibly jobdependent 

— amount. A maintenance level of less than zero is associated with 

the machine failure. Consequently, scheduling maintenance activities that raise 

the maintenance level may become necessary in order to prevent maintenance 

level being becoming negative. We present complexity results and approaches 

considering two types of maintenance activities. 

� TC-09 


6.2.53 

Mathematical Programming, Machine and 

Statistical Learning 



Chair: Tamas Terlaky, Industrial and Systems Engineering, Lehigh 

University, H.G. Mohler Lab., 200 W. Packer Avenue, 18015, 

Bethlehem, Pennsylvania, United States, terlaky@lehigh.edu 

1 - On the Rescaling Algorithms for Linear Optimization 

Tamas Terlaky, Industrial and Systems Engineering, Lehigh 

University, H.G. Mohler Lab., 200 W. Packer Avenue, 18015, 

Bethlehem, Pennsylvania, United States, terlaky@lehigh.edu, 

Dan Li 

The perceptron algorithm is not a polynomial time algorithm to solve linear inequality 

systems, but Dunagan/Vempala showed that with high probability the 

largest inscribed ball can be inflated sufficiently to reach polynomial solvability. 

We extend the rescaling idea used to improve the complexity of the von 

Neumann algorithm, herewith opening the road to improve the complexity of 

the colourful feasibility problem. Various properties and parallelizablity of the 

re-scaling phase are analyzed, too. 

2 - Comparison of Performance of All-together Support 

Vector Machines for Multiclass Classification 

Keiji Tatsumi, Division of Electrical, Electronic and Information 

Engineering, Osaka University, Yamada-Oka 2-1, 565-0871, 

Suita, Osaka, Japan, tatsumi@eei.eng.osaka-u.ac.jp, Tetsuzo 

Tanino 

In this paper, we focus on a new all-together method of support vector machines 

(SVMs) for multiclass classification called multiclass multiobjective 

SVM (MMSVM), which maximizes geometric margins in the sense of multiobjective 

optimization. However, all of its many Pareto solutions are not guaranteed 

to have high generalization ability, and it is not known what kind of 

classifier each Pareto solution corresponds to. Thus, through numerical experiments, 

we compare solutions obtained by MMSVM and other methods with 

respect to geometric margins and generalization. 

3 - On the Incomplete Oblique Projections Method for Solving 

Box Constrained Least Squares Problems 

154 

Hugo Scolnik, Computer Science, University of Buenos Aires, 

Ciudad Universitaria, Pabellon I, 1428, Buenos Aires, Argentina, 

hscolnik@gmail.com 

This paper aims to extend the applicability of the IOP algorithm (Scolnik, 

Echebest, Guardarucci, Vacchino) for solving inconsistent linear systems to the 

box constrained case which arise in many important applications and optimization 

problems. The new algorithm employs incomplete projections onto the set 

of solutions of the augmented system Ax-r=b, together with box constraints. 

Convergence is analyzed and numerical experiments are presented comparing 

its performance with some well-known methods. 

4 - Gaussian kernel models for irregular demand processes 

José Luís Carmo, CIO and University of Algarve, Portugal, 

jlcarmo@ualg.pt, Antonio Rodrigues 

This work focuses on the study of forecasting problems and optimal decision 

making, in the context of inventory management, given irregularly observed 

demand processes. We propose adaptive models based on gaussian kernels, either 

for point forecasting, or for density forecasting, of the time of occurrence 

and the magnitude of future demand. Optimal decisions, under asymmetric 

cost functions, are derived from quantiles computed from the density forecasts. 

We produce encouraging empirical results of this approach, tested with both 

simulated and real irregularly spaced time series. 

� TC-10 


6.2.56 

Graphs and Networks VIII 



Chair: Bernard Ries, Warwick Business School, University of 

Warwick, CV4 7AL, Coventry, Switzerland, 

Bernard.Ries@wbs.ac.uk 

1 - 11 Methods for proving infeasibility of IP’s. 

Anastasia Kouvela, Operational Research, London School of 

Economics and Political Science, 29 Abercorn Pl, Flat 171, 

NW89DU, London, United Kingdom, a.kouvela@lse.ac.uk, 

Dimitris Magos, Gautam Appa, Yiannis Mourtos 

We present a summary of different methods to tackle infeasibility and evaluate 

their viability for large scale 0-1 Integer Programming Problems. A challenging 

example is that of proving the non-existence of three 10x10 Mutually Orthogonal 

Latin Squares; which has been formulated as a five index assignment 

problem. We discuss methods extensively analysed in the literature, such as 

Hermite Normal Form, Aggregation, Lattice Basis Reduction and the Nullstellensatz 

Certificate of infeasibility; and some resent ideas that derive from the 

study of the intersection and row-column Graph. 

2 - The metric dimension of generalized Petersen graphs 

Mirjana Cangalovic, Department for Mathematics, Statistics and 

Operational Research, Faculty of Organizational Sciences, 

University of Belgrade, Jove Ilica 154, Belgrade, Serbia, Serbia, 

canga@fon.rs, Jozef Kratica, Vera Kovacevic Vujcic 

We consider the problem of determining the metric dimension of the generalized 

Petersen graph G(n,k). It is proved that the metric dimension of G(n,k), 

for parameter k not smaller than 2, is greater or equal to 3. For k equal to 2 we 

find explicitly a metric basis of the cardinality 3 and the corresponding metric 

coordinates of all vertices of the graph. It follows that the metric dimension of 

G(n,2)is equal to 3. 

3 - A Branch-and-Price Mechanism for Bimodal Multicommodity 

flow problems 

Ashwin Arulselvan, Warwick Business School, University of 

Warwick, Room E1.02 (Social Studies Building), Warwick 

Business School university of warwick, CV4 7AL, Coventry, 

Warwickshire, United Kingdom, ashwin.arulselvan@wbs.ac.uk, 

Steffen Rebennack, Panos Pardalos 

We provide a path based formulation for a bimodal multicommodity flow problem 

and solve it using a branch-and-price approach. We also study some stabilization 

techniques to accelerate the convergence.

� TC-11 


8.2.38 


Technology II 









1 - Reducing product development complexity through information 

management 

Ahm Shamsuzzoha, Department of Production, University of 

Vaasa, Lindroosintie 2, C19, FI-65101, Vaasa, Finland, 

ahsh@uwasa.fi, Petri Helo 

Product Development process is a collaborative network to drive increased productivity. 

The focus of this research is to integrate the essential features of information 

flow with product development process. An empirical study is conducted 

through a case company. Existing dependency pattern and required information 

flow were collected and review meetings conducted. We considered 

to mitigate the complexity of information flow or dependency pattern between 

component-to-component and component to customer requirement. Several 

recommendations were proposed. 

2 - Marketing Information Systems (MIS) as a Tool for Performance 

Measurement in Nigeria Aluminium Industry: 

Operations Research Perspective 

Joshua Magbagbeola, ACTUARIAL SCIENCE AND 

INSURANCE, JOSEPH AYO BABALOLA UNIVERSITY, 

P.M.B. 5006, ILESHA, IKEJI-ARAKEJI, OSUN STATE, 234, 

IKEJI-ARAKEJI, ILEHSA, OSUN STATE, Nigeria, 

kunle_magbagbeola@yahoo.com 

More than 90% of Nigerian’s income is from exporting of crude oil. To expand 

and search for new avenue for revenue, other industrial areas like the Auminium 

industry have to be investigated. The study examined among other things the 

performance of Aluminium industry in Nigeria with particular reference to the 

secondary sector of the industry and special cognisance to the appreciation of 

Operations Research as a veritable tool in Information Systems. Questionnaires 

were carefully prepared, administered and analyzed accordingly. 

3 - Multicriteria Decision Analysis and Geographic Information 

Systems (GIS) 

Ceren Gundogdu, administration, Yildiz Technical University, 

Barboros Bulvarı Yildiz Kampusu, H-Blok, 34349, ˙Istanbul, 

Turkey, ceren_erdin@yahoo.com 

Geographic Information Systems (GIS) is an information-technology based 

knowledge system providing all kinds of positional and spatial descriptive information 

in a related format. GIS is required in almost all types of enterprises 

today and they play an important role in the solution of corporate problems. By 

using GIS we show some example related to supply chain management, electronic 

business platform and mobile GIS in corporate decision support system 

and for solving some problems. 

4 - Evaluate Collaboration Design Systems Performance 

based on CEGRA 

Cheng-Ru Wu, Yuanpei University, Taiwan, 

alexru00@ms41.hinet.net, Chiu-Chin Chen 

This study focused on using the cause-effect grey relational analysis (CE- 

GRA) decision support tools for the Taiwan bureau’s control information systems 

(CIS) analysis the collaboration design information systems performance 

case studies. CEGRA in the combination-based approach, this article provides 

decision-makers are doing collaboration design, to more practical and accurate 

in line with the standard structure of domestic industries to enhance the control 

of collaboration design the overall effectiveness of information systems. 


� TC-12 


8.2.39 

ANP 03 


Processes 


Chair: Yasemin C. Erensal, Industrial Engineering, Marmara 

University, Göztepe, 34000, Istanbul, Turkey, yerensal@gmail.com 

1 - Sequencing Attributes of Energy Portfolio Optimization 

in Industrial Regions 

Alp Muzaffer Arslan, Management Faculty, ITU, Kaptan Arif 

Cad. Tonozlu Sk. No: 32/23, 34710, Istanbul, Turkey, 

muzalparslan@gmail.com, Gulgun Kayakutlu 

Different technologies of renewable energy are critical for industrial regions 

but to be used in a portfolio. This study aims to identify, classify and prioritize 

the influential factors. Classes of factors are defined as technological, political, 

social and environmental. Dependency among attributes and among classes 

caused the application of ANP in order to determine the priorities. Initial survey 

is run among the academics of energy field; further work is continued with 

the small and medium companies. Case in Turkey shows that the most influential 

attribute is the European energy policies. 

2 - EThe process of a Taiwan pharmaceutical company selects 

drugs for international contract manufacturing 

Su-Chuan Shih, Business Administration, Providence University, 

200 Chung Chi Road, Business Administration, 43301, Taichung 

City, Taiwan, Taiwan, scshih@pu.edu.tw 

This research evaluated the group decision-making strategies of contract manufacturing 

in different drug types by applying the Analytic Network Process 

(ANP) method. The results had revealed that while accepting big pharmaceutical 

companies’ orders of contract manufacturing, the efficiency and effectiveness 

of management is the most important facet that Taiwan’s pharmaceutical 

company considered, the other ones are the innovative partner relationship and 

the industry factor. The results of index weighing ranking are Return on Assets 

(ROA), establishing partnerships with big pharmaceutical companies, and 

reduce risks in contract manufacturing. 

3 - An ANP Based BOCR Approach for Marketing Strategy 

Selection 

Tuncay Gürbüz, Industrial Engineering, Galatasaray University, 

Ciragan cad. No.36, Ortakoy, 34357, Istanbul, Turkey, 

tuncaygurbuz09@gmail.com, Y. Esra Albayrak, Yasemin C. 

Erensal 

In today’s world, producing is important and yet its importance is not even 

comparable to that of selling. As the competition is fierce, the companies has 

to attach greater importance to the sales. And for a good performance in sales, 

the company has to have an effective marketing strategy. Because the chosen 

strategy can work out pretty good and rise the sales as well as it can be a quite 

wrong strategy and hence drop the sales. In order to choose a suitable marketing 

strategy, in this study, a BOCR model will be built and analyzed with ANP. 

For each dimension, a suitable strategy will be chosen but at the end the final 

aggregation will give us the overall best strategy to follow. 

4 - The Relationship of Salary and Performance for Food 

and Beverage Industry Based on DEMATEL with a 

MCDM Model 

Yi-Chun Chen, Department of Leisure Management, Taiwan 

Hospitality & Tourism College, No 268 Chung-Hsing ST., 

Feng-Shan Village, Shou-Feng County, 974, Hualien, Taiwan, 

chen.vivien@gmail.com, Gwo-Hshiung Tzeng 

The Purpose of This Paper Is to Analyze and evaluate the Food and Beverage 

Industry Employees Salary Verses Employees performance. In This Research 

DEMATEL Technique is Used to Build the Network Relationship Map Between 

Salary and Performance; Then ANP Relative Importance Weights Can 

Be Obtained Based the Results of DEMATEL Technique; Finally a MCDM 

Model is Conducted to Evaluate and Improve Satisfaction of Employee Needs. 

However, the Main Contribution of This Study is to help the employer clarifying 

the needs of employees and the related factors. 

155


� TC-13 


2.2.21 

Reliability in Location 



Chair: Dmitry Krass, Rotman School of Mgmt, University of 

Toronto, 105 St. George st., M5S 3E6, Toronto, Ontario, Canada, 

krass@rotman.utoronto.ca 

1 - Location Problems with Two Unreliable Facilities on a 

Line 

Mozart Menezes, MIT-Zaragoza International Logistics 

Program, Zaragoza Logistics Center, 50197, Zaragoza, Zaragoza, 

Spain, mmenezes@zlc.edu.es, Oded Berman, Dmitry Krass 

We study the 2-Median and 2-Center problems for facilities that may be subject 

to (possibly correlated) disruptions; customer demand is uniformly distributed 

over a unit segment. To variants of the problem are analyzed: (1) the customers 

have full information on the status of facilities, and (2) no such information is 

available. In addition analyzing location patterns, we provide the decomposition 

of the optimal cost into the components corresponding to the cost of travel, 

unreliability and incomplete information. The sensitivity of the optimal location 

and optimal cost to the probability of disruption and to the correlation of 

failures is presented and managerial insights are discussed. 

2 - Hedging against territorial disruptions in capacitated 

distribution systems 

Federico Liberatore, Kent Business School, University of Kent, 

Annexe, CT2 7PE, Canterbury, United Kingdom, 

fl51@kent.ac.uk, Maria Paola Scaparra 

We present a multi-level model to improve the reliability of supply systems in 

the face of disastrous events that affect wide geographical areas. Examples of 

areal disruptions are: floods, earthquakes, hurricanes, and the spreading of biological 

and chemical agents. Two optimal solution approaches are proposed 

and compared. 

3 - An Exact Linear Reformulation of the Expected Maximum 

Covering Problem 

Jesse O’Hanley, Kent Business School, University of Kent, CT2 

7PE, Canterbury, United Kingdom, j.ohanley@kent.ac.uk, Sergio 

García Quiles, Maria Paola Scaparra 

In this talk we consider the expected maximum covering problem, which can 

be used to design coverage networks that are robust to random facility failures. 

Although naturally represented in nonlinear form, we show how the model can 

be reformulated as an exact mixed integer linear program using simple linear 

constraints to iteratively evaluate high-degree polynomial terms involving 

probability products. Results and analyses are presented for various test problems 

showing the effectiveness of the new formulation in comparison to an 

approximation model and heuristic methods. 

4 - Optimal Solution to n-Median With Unreliable Facilities 

on a Line 

Dmitry Krass, Operations Management, University of Toronto - 

Rotman School of Management, 105 St. George Street, 

M5S-3E6, Toronto, Ontario, Canada, 

Krass@Rotman.Utoronto.Ca, Oded Berman 

We present an approach for solving an n-facility median problem with unreliable 

facilities. The direct approach leads to loss of tractability for n>3. However, 

by recasting the problem as a weighted combination of median problems 

an general case can be solved. 

� TC-14 


2.2.15 

Vendor Managed Inventory 



Chair: Michael Vidalis, Business Administration, University of 

Aegean, Sachtouri 19 Agios Dimitrios, 17341, Athens, Greece, 

mvid@ath.forthnet.gr 

156 

1 - Implementing a Supplier Managed Inventory Policy in 

an Internal Logistics Setting 

Tim Govaert, Industrial Management, University of Ghent, 

Technologiepark 903, 9052, Zwijnaarde, tim.govaert@ugent.be, 

El-Houssaine Aghezzaf, Sofie Van Volsem 

This paper discusses the issue of effective inventory control and internal logistics 

management within a complex production and assembly system. The 

main objective of this investigation is to examine the performance of inventory 

control policies based on the concept of supplier managed inventory (SMI) and 

in the same time to optimize the relevant internal logistic Costs. We first start 

by defining the structure of inventory system and its critical parameters and 

then discuss the related internal logistics. We assume that there is a large offsite 

warehouse that supplies raw materials and components to the facility. This 

warehouse supplies a small area within the facility which in turn supplies the 

production and assembly operations. The supply is fully controlled using an 

SMI based policy. We investigate the effect of the critical parameters related to 

inventory as well as to the internal logistics (the number of incoming/outgoing 

trucks from suppliers/to retailers, the number of forklifts and lift trucks, etc.). 

The ultimate goal is of course an effective allocation of internal resources and 

increase of the service level. 

2 - Coordination of Production and Dispatching Decisions 

in Vendor Managed Inventory Systems 

Onur Kaya, INDR, Koc University, Koc University, Eng 206, 

Sariyer, 34450, Istanbul, Turkey, okaya@ku.edu.tr, Deniz 

Kubali, E. Lerzan Ormeci 

We focus on the coordination of transportation and production policies between 

a manufacturer and a retailer in a vendor-managed inventory system. The manufacturer 

needs to determine the length of the production cycle as well as the 

shipment timings and the quantities to the retailer. We consider both the deterministic 

and the stochastic cases in a general setting and we also consider 

simpler dispatch policies, like time-based and quantity-based shipments. We 

also do a computational study to compare the performances of these policies 

and to analyze the effects of the parameters to the system. 

3 - Supplier managed inventory in case of stochastic demand 

and finite supply 

Sofie Van Volsem, Industrial Management, Ghent University, 

Technologiepark 903, BE 9052, Gent, Belgium, 

sofie.vanvolsem@ugent.be, El-Houssaine Aghezzaf, Tim 

Govaert 

The typical IRP deals with distribution of a single product to a set of customers 

with constant demand rates, from a single distribution center (DC) with an unlimited 

supply, but these assumptions do not always reflect reality. We consider 

finite DC capacity and stochastic customer demands. Under these assumptions, 

DC stockouts can occur, causing backorders and possibly customer stockouts. 

We introduce an SMI—model to analyze the trade-off between all costs involved. 

Furthermore, the influences of customer clusters and timetables on DC 

inventory are analyzed and improvements are proposed. 

4 - Modeling a Supply Chain with two members, (S, s) inventory 

policy Poisson external demand and Coxian-2 

lead time 

Michael Vidalis, Business Administration, University of Aegean, 

Sachtouri 19 Agios Dimitrios, 17341, Athens, Greece, 

mvid@ath.forthnet.gr, Vassilios Vrisagotis 

We model a dynamic supply network with two members, stochastic demand 

and replenishment. External demand is pure Poisson, retailer follows (S,s) policy 

and the lead time follows a Coxian-2 distribution. The supply network is 

modeled as a continuous time Markov process with discrete states. The structure 

of the transition matrix is explored and a computational algorithm is developed 

to generate this for different values of system characteristics. Performance 

measures such as fill rate, WIP and flow time are obtained. Also expressions 

for the holding costs and shortage costs are derived.

� TC-15 


2.2.12 

Dynamic and Stochastic Vehicle Routing 



Chair: Lars Magnus Hvattum, Industrial Economics and Technology 


Alfred Getz veg 3, Sentralbygg 1, N-7491 Trondheim, Norway, 7491, 

Trondheim, Norway, lars.m.hvattum@iot.ntnu.no 

1 - Real-time delivery of perishable goods using past request 

data to increase customer service quality 

Francesco Ferrucci, WINFOR (Business Computing and 

Operations Research), University of Wuppertal, Gaussstrasse 20, 

42119, Wuppertal, NRW, Germany, fferrucci@winfor.de, Stefan 

Bock, Michel Gendreau 

In this talk, we focus on a real-world variant of the well-known Dynamic Vehicle 

Routing Problem (DVRP). In the considered DVRP, goods have to be 

delivered under high time pressure because of their high perishability. In order 

to reduce delivery times and to increase quality of customer service experience, 

future demands are anticipated. For this purpose, sophisticated methods 

for analyzing historical data and request forecasting are applied. Recent computational 

results on designed and real-world data are presented. 

2 - Robust optimization of internal transports 

Brigitte Werners, Wirtschaftswissenschaft, Ruhr-University 

Bochum, Universitätsstr, 44780, Bochum, Germany, or@rub.de 

Significant reductions in internal transports at one of the Deutsche Post DHL’s 

main parcel sorting centers can be achieved by applying the robust solution of 

a modified three-dimensional linear assignment model. The suggested MILP 

model minimizes necessary manual transportation effort by layout modifications. 

Specific requirements are taken into account and uncertain data are handled 

by applying robustness criteria which can be generally applied to develop 

robust solutions in a dynamic decision situation. 

3 - Metaheuristics for the dynamic stochastic dial-a-ride 

problem with expected return transports 

Michael Schilde, Department of Business Administration, 

University of Vienna, Brünner Straße 72, Bauteil 2, Raum 179, 

1210, Vienna, Austria, michael.schilde@univie.ac.at, Karl 

Doerner, Richard Hartl 

The Austrian Red Cross faces the problem of designing vehicle routes to serve 

dynamically arising transportation requests from patients’ home locations to 

hospitals using a fixed vehicle fleet. Each request may cause a return transport 

in the opposite direction on the same day. Some stochastic information 

about these return transports is available. We show that using this kind of information 

while designing the vehicle routes can be beneficial to solution quality. 

Different modifications of the variable neighborhood search metaheuristic were 

tested using real world inspired test instances. 

4 - Using Decision Trees for a Stochastic Maritime Routing 

Problem 

Lars Magnus Hvattum, Industrial Economics and Technology 


Alfred Getz veg 3, Sentralbygg 1, N-7491 Trondheim, Norway, 

7491, Trondheim, Norway, lars.m.hvattum@iot.ntnu.no, Eystein 

Fredrik Esbensen, Kjetil Fagerholt, Bjørn Nygreen 

A real-world shipping company services two types of requests. The first has a 

fixed starting time and late arrivals result in severe penalties. The second has no 

time windows, but the total amount of goods transported in any time interval is 

restricted by contracts. The main stochastic element is port congestions, which 

may result in a substantial risk of arriving late for a fixed time contract. The 

size and structure of the case allow a solution method based on decision trees. 

Simulation results are provided illustrating the usefulness of the method. 


� TC-16 


2.2.14 

OR Applications in Railways 



Chair: Dennis Huisman, Econometric Institute, Erasmus University, 

Rotterdam, Netherlands, huisman@ese.eur.nl 

1 - Integrated Planning of Railway Transportation Resources 

David De Almeida, Innovation & Research Department, SNCF, 

45 rue de Londres, 75379, Paris CEDEX 08, France, 

david.de_almeida@sncf.fr, Faten Benhizia, Stéphane 

Dauzere-peres 

In this presentation, we deal with the integration of rolling-stock and driver 

planning in railway production. This prospective research is partly motivated 

by previous works in the airline industry and in public transportation. We first 

propose a mixed integer linear programming model where coupling constraints 

are introduced to ensure the consistency in the use of the two resource types. 

Some preliminary numerical experiments obtained with a standard solver are 

presented. Alternatives for handling problems of industrial size in reasonable 

calculation times will be further discussed. 

2 - Solving Large Scale Crew Scheduling Problems Efficiently 

Ricardo Saldanha, Innovation, SISCOG - Sistemas Cognitivos, 

SA, Campo Grande 378-3, 1700-097, Lisboa, Portugal, 

rsaldanha@siscog.pt, Erwin Abbink, Luis Albino, Twan 

Dollevoet, Dennis Huisman, Jorge Roussado 

We address the problem of assigning anonymous drivers and conductors to 

more than 60,000 trips (complete trains or parts of trains) running on a standard 

week period. Typically huge problems like this are solved in several steps 

where in each step only part of the decision variables is considered. We show 

that solving the problem considering a priori all decision variables together results 

in solutions that are significantly more efficient. This is only possible 

with the use of advanced techniques and algorithms capable of tackling the additional 

complexity of the problems to be solved. We test our approach with 

real-life data supplied by the Dutch railway operator, Netherlands Railways. 

3 - Tram traffic management at signalised junctions with 

tram stops 

Jeremi Rychlewski, Civil and Environmental Engineering, 

Poznań University of Technology, Piotrowo 5, 61-138, Poznań, 

Poland, jeremi.rychlewski@put.poznan.pl 

The aim of this paper is to discuss strategies for managing priority tram traffic 

at junctions with traffic lights and tram stops at entry or at exit of the junction. 

Problems with tram stops at junction entry include variable service time 

at the tram stop, but also turnout and traffic light management when two trams 

going in different directions are to be served. Tram stop platforms at junction 

exit should not be too long, but that puts a risk of a second tram blocking the 

junction. The problems are based on experience from Poznan and other Polish 

cities. 

4 - Developments in crew re-scheduling at Netherlands 

Railways 

Dennis Huisman, Econometric Institute, Erasmus University, 

Rotterdam, Netherlands, huisman@ese.eur.nl, Daniel Potthoff 

After a major disruption, the re-scheduling of train drivers and guards is the 

bottleneck in the operation of Netherlands Railways (NS). For example, there 

were hardly any trains running on time during a week with snowfall in December 

2009. OR based algorithms can help to solve these problems. In this talk, 

we present the latest developments in these algorithms and we discuss how NS 

want to apply them during the day of operations. 

157


� TC-17 


1.3.14 

Models for Vehicle Routing 



Chair: Nurul Mohamed, Kent Business School, University of Kent, 1 

Kingsmead Road, CT1 1BN, Canterbury, Kent, United Kingdom, 

nurul.h@gmail.com 

1 - Multi-depot vehicle routing problem (MDVRP): comparative 

study of alternative formulations 

Tânia Ramos, Department of Management and Centro de Estudo 

de Gestão do IST, ISCTE-IUL and IST-UTL, Avenida das Forças 

Armadas, Edificio ISCTE, 1649-026, Lisboa, Portugal, 

tania.ramos@iscte.pt, Maria Isabel Gomes, Ana Paula 


The MDVRP appears as a generalization of the classical VRP. There are different 

mathematical formulations of VRP used in the literature. This paper aims to 

compare exact formulations to solve MDVRP. We worked with the three index 

formulation adapted to the MDVRP and proposed the two-commodity flow formulation 

for the MDVRP. In the three index formulation, we compared three 

alternative formulations based on Miller-Tucker-Zemlin constraints to eliminate 

subtours. The mixed-integer linear programming models developed are 

applied to some small and medium size scale instances. 

2 - Two objective functions for a special Split Delivery Vehicle 


Marc Uldry, Département d’Informatique, Université de 

Fribourg, Boulevard de Pérolles 90, 1700, Fribourg, Switzerland, 

marc.uldry@unifr.ch, Marino Widmer 

Different products have to be delivered by trucks. Due to the trucks capacity, 

each order is split into one or more deliveries supplied from a main depot 

or from local depots. Simple and complex constraints have to be satisfied as 

drivers and trucks availability or the fact that only specific trucks can be fulfilled 

in local depots. A MIP considers two different objective functions. The 

first one defines a vehicle routing where the total travel time is minimized while 

the second one defines a vehicle routing where the number of different trucks 

supplying each individual customer is minimized. 

3 - Branch-and-price for the multi-depot pickup and delivery 

problem with heterogeneous fleet and soft time windows 

Andrea Bettinelli, Dipartimento di Matematica, Università degli 

Studi di Milano, via Saldini 50, 20133, Milano, Italy, 

andrea.bettinelli@unimi.it, Alberto Ceselli, Giovanni Righini 

The multi-depot pickup and delivery problem with heterogeneous fleet and soft 

time windows calls for finding a minimum cost routing for a fleet of vehicles 

with different capacities and based at different depots, satisfying a given set of 

customers. For each customer a demand must be picked up at a source and 

delivered at a destination in the same route. Each location has a time window 

for the service that can be violated at the cost of a linear penalty. We propose a 

branch-and-price algorithm. The pricing problem is solved through a bidirectional 

dynamic programming method. 

4 - Split Delivery Vehicle Routing Problem 

Nurul Mohamed, Kent Business School, University of Kent, 1 

Kingsmead Road, CT1 1BN, Canterbury, Kent, United Kingdom, 

nurul.h@gmail.com, Said Salhi, Gábor Nagy 

Split Delivery Vehicle Routing Problem (SDVRP) is one of Vehicle Routing 

Problem (VRP) variants where each customer can be served more than once. 

We will explore and analyse some new approaches based on the saving and 

sweep algorithms together with some refinement schemes to solve the SDVRP. 

A set partitioning based approach is put forward and used in ILOG Cplex to get 

a better solution. Preliminary results will be highlighted along future research. 

158 

� TC-18 


1.3.15 

Complex Systems under Uncertainty: 

Networks and Data Mining 











1 - On a Class of Branching Problems in Broadcasting and 

Distribution 

Sohail Chaudhry, Management and Operations, Villanova 

University, Villanova School of Business, 800 Lancaster Avenue, 

19085, Villanova, PA, United States, 

sohail.chaudhry@villanova.edu, Edward Rosenthal, In-Chan 

Choi, Jinbong Jang 

The class of branching problems in broadcasting and distribution (BBD) that 

we introduce in this paper represents a large class of combinatorial and network 

optimization models. We introduce the following network optimization problem: 

given a finite directed graph with a cost function on the arcs, demands at 

the nodes, and a single source s, find the minimum cost connected subgraph 

from s such that its total demand is no less than lower bound D. 

Theoretically, BBD generalizes several known discrete optimization problems, 

including the knapsack problem, Edmonds’ minimum branching problem, the 

partially ordered knapsack problem, and certain scheduling problems. BBD 

has broad practical application as well. We show how BBD models disaster 

relief efforts such as food or medicine distribution. In addition, we describe a 

"broadcasting’ application which shares the same properties as the distribution 

problems, which, loosely speaking, are to maximize effect at minimum cost. 

We prove that our problem is strongly NP-complete, give an integer programming 

formulation, outline two heuristic approaches, and illustrate them with a 

numerical example. In additional, we report on our computational experience 

on randomly generated problems. 

2 - Data mining approach for network intrusion detection 

using mobile agent 

Djamal Dris, Lamos laboratory, université de Bejaia, route de 

Targua ouzemour Bejaia, 06000, Bejaia, Bejaia, Algeria, 

drisdjamal@hotmail.com, Lynda Sellami, Khaled Sellami, 

Mohamed Ahmed-nacer 

A data mining techniques that incorporates the mobile agent system based Intrusion 

Detection System (IDS) has been defined to guaranty an efficient computer 

network security architecture. We provide an overview on data mining 

algorithms that we have implemented: association rules algorithm. Whose is 

used to compute the intra- and inter- audit record patterns, which are essential 

in describing program or user behavior. We propose an agent-based architecture 

for intrusion detection systems where the learning agents continuously 

compute and provide the updated (detection) models to the detection agents. 

3 - Probabilistic Analysis of Multidimensional Assignment 

Problems 

Pavlo Krokhmal, Mechanical and Industrial Engineering, 

University of Iowa, 3131 Seamans Center, 52242, Iowa city, IA, 

United States, krokhmal@engineering.uiowa.edu 

We consider a class of multidimensional assignment problems (MAPs) that 

generalize the well-known Linear and Quadratic Assignment Problems. Properties 

of large-scale randomized instances of MAPs are investigated under assumption 

that their assignment costs are iid random variables. For a broad 

class of probability distributions, we demonstrate that optimal values of random 

MAPs converge in L1 and almost surely as problem size increases. Computational 

properties of large-scale randomized MAPs are discussed, including 

polynomial algorithms that deliver epsilon-approximate solutions almost surely

4 - Operating Overall Production under Chance Constraints 

Baruch Keren, Industrial Engineering and Management 

Department, SCE - Shamoon College of Engineering, 

Bialik/Basel Sts., P.O.B. 950, 84100, Beer Sheva, Israel, 

baruchke@sce.ac.il, Zohar Laslo, Gregory Gurevich 

A single product can be manufactured in n plants with heterogeneous characters. 

Each plant has its specific stochastic production capability. The expected 

capability and the standard deviation of each plant can be increased by allocation 

of additional budgets. The periodic demands for the product are forecasted 

and given by a probability function for each period. The problem is to determine 

the total budget needed and its distribution among the n plants in order 

to ensure a complete fulfillment of the demands according to the due dates and 

the pre-given confidence levels. 

� TC-19 



Universality in complex systems 



Chair: Rui Gonçalves, Engenharia Civil, Faculdade de Engenharia da 

U. Porto, R. Dr. Roberto Frias, 4200-465, Porto, rjasg@fe.up.pt 

1 - Euclidean Jordan algebras, generalized Krein conditions 

and strongly regular graphs with one eigenvalue 

in modulus less than one 

Luís Vieira, Engenharia Civil, Faculdade de Engenharia do 

Porto, R. Dr. Roberto Frias, 4200-465, Porto, lvieira@fe.up.pt 

Let G be a strongly regular graph. We first present the generalized Krein parameters 

of G in the environment of Euclidean Jordan algebras and we analyze 

the spectra of G when one of the eigenvalues distinct from the regularity is in 

modulus less than one. Next, we analyze the generalized Krein parameters of 

G when his order is big and establish some theorems on the spectra and on the 

parameters of G. Finally, we present some conditions on the Krein parameters 

of a strongly regular graph H recurring to the generalized Krein parameters of 

H. 

2 - The Relationship between Crude Oil Prices and Exchange 

Rates 

André Salles, Industrial Engeneering, Federal University of Rio 

de Janeiro - UFRJ, Av. Ataulfo de Paiva, 348 ap. 501 - Leblon, 

22440033, Rio de Janeiro, Rio de Janeiro, Brazil, as@ufrj.br 

This work uses a Bayesian approach: to study the risk and returns of the oil 

prices; and examines the relationship between crude oil prices and exchange 

rate. The methodology used here takes in consideration the violation of homoscedasticity 

and Gaussian distribution, in the construction of the models. 

The data used in this study consists of the daily closing exchange rate, of US 

dollar to Euro, and oil prices, from June 2006 to March 2009. 

3 - Universality in the stock market 

Rui Gonçalves, Engenharia Civil, Faculdade de Engenharia da U. 

Porto, R. Dr. Roberto Frias, 4200-465, Porto, rjasg@fe.up.pt, 

Helena Ferreira, Alberto Pinto 

We compute the analytic expression of the probability distributions FIp,+ and 

FIp,- of the normalized positive and negative Ip index returns r(t), with periodicity 

p. We study North American, European and World wide indices, energy 

sources and exchange rates. We define the alpha re-scaled Ip index returns 

that we call, after normalization, the alpha positive and alpha negative fluctuations. 

We use the Kolmogorov-Smirnov statistical test, as a method, to find 

the values of alpha that optimize the data collapse of the histogram of the alpha 

fluctuations with the Bramwell-Holdsworth-Pinton (BHP) probability density 

function. Since the BHP probability density function appears in several other 

dissimilar phenomena, our results reveal a universal feature of the stock exchange 

markets. 


� TC-21 


6.2.47 

Software for OR/MS I - Optimization 



Chair: Bjarni Kristjansson, Maximal Software, Ltd., Boundary 

House, Boston Road, W7 2QE, London, United Kingdom, 


1 - Why is Maximal Software Now Giving Away Free Development 

Copies of the MPL Modeling System? 

Bjarni Kristjansson, Maximal Software, Ltd., Boundary House, 

Boston Road, W7 2QE, London, United Kingdom, 

bjarni@maximalsoftware.com, Sandip Pindoria 

In today’s challenging economy, many are now looking for ways to save on 

their IT budget, including when purchasing optimization software. We at Maximal 

have now decided to fundamentally change how we sell our software, by 

greatly reducing how much you pay for development copies of MPL, in many 

cases bringing the actual cost down to zero. We will explain several new programs: 

"Pay Maintenance Only, "Subscription-Based Pricing," "Free Development 

Copies of MPL" and "Free Software for Academics." 

2 - Performance of Optimization Software - An Update 

Hans Mittelmann, School of Math&Stats, Arizona State 

University, Box 871804, 85287-1804, Tempe, AZ, United States, 

mittelmann@asu.edu 

We report on the current status of our benchmarking effort for both discrete and 

continuous optimization software. 

3 - Xpress-Kalis: Automatic Linear Relaxations for Combined 

CP-MIP Problem Solving with Application to Nuclear 

Power Plant Maintenance Outage Scheduling 

Susanne Heipcke, Xpress team, FICO, 54 rue Balthazar de 

Montron, 13004, Marseille, susanneheipcke@fico.com, Fabrice 

Buscaylet 

After a brief overview of the capabilities of the Xpress-Mosel language for handling 

multiple problems and multiple solvers, we explain the concepts behind 

automatic linear relaxations for combined CP-MIP solving with Xpress-Kalis 

and Xpress-Optimizer. As an application example demonstrating the suitability 

of the implementation for large-scale optimization problems, we show how 

the problem of planning nuclear power plant maintenance outage for 50+ reactors 

over several years has been modeled and solved using the automatic linear 

relaxations functionality. 

4 - Adaptable Robust Optimization and Chance Constraints 

approximation using automatic model generation 

in AIMMS 

Ovidiu Listes, AIMMS, Paragon Decision Technology, 


o.listes@aimms.com 

We illustrate how AIMMS is able to generate automatically the Robust Counterpart 

of a model in case the uncertain data belongs to uncertainty sets like 

Box, Ellipsoid, Polyhedron or Scenarios. The automatic generation extends 

to adjustable variables following Linear Decision Rules, as well as to robust 

approximations of Chance Constraints. We show how the intuitive, effective 

modeling concepts in AIMMS allow for fast, flexible experiments and comparison 

of results based on various uncertainty sets. 

� TC-22 



Health Care Policy Making I 

Stream: Health Care Management [c] 


Chair: Dave Worthington, The Management School, Lancaster 

University, Dept. Of Management Science, LA1 4YX, Lancaster, 

Lancashire, United Kingdom, d.worthington@lancaster.ac.uk 

159


1 - New Integer-Programming Formulation for Kidney Exchange 

Problem 

Dorien de Regt, TU Delft, Mekelweg 4, 2628 CD, Delft, 

Netherlands, D.N.deregt@student.tudelft.nl, Abdur Rais, Miguel 

Constantino, Ana Viana 

A kidney transplant patient having an associated kidney donor may not receive 

the donor kidney because of incompatibility. However, by exchanging 

donor kidneys between two or more incompatible patient-donor pairs, successful 

transplants can be performed. For a set of incompatible patient-donor 

pairs, Kidney Exchange Problem (KEP) finds optimal matching of patients and 

donors across the incompatible pairs for successful kidney transplants. In this 

talk, a new IP formulation for the KEP will be presented. The new description 

and computational results will be compared to other known formulations. 

2 - Setting Staffing Requirements for Time Dependent 

Queueing Networks: The Case of Accident and Emergency 

Departments 

Navid Izady, Lancaster University, United Kingdom, 

n.izady@lancaster.ac.uk, Dave Worthington 

An incentive scheme aimed at reducing patients waiting times in A&Es was 

introduced by the UK government in 2000. It requires 98 percent of patients to 

be served within 4 hours of arrival. Setting the minimal hour by hour staffing 

levels for achieving the target, in the presence of complexities like time-varying 

demand, is the subject of this talk. We propose an iterative scheme which uses 

queueing models and simulation to produce a good solution. The implementation 

of this method in a typical A&E suggests that significant improvements 

can be achieved without increasing total staff-hours. 

3 - Optimal Multileaf Collimator Leaf Sequencing in IMRT 

Treatment Planning 

Z. Caner Ta¸skın, Department of Industrial Engineering, Bo˘gaziçi 

University, Bebek, 34342, ˙Istanbul, Turkey, 

caner.taskin@boun.edu.tr, Cole Smith, Edwin Romeijn, James 

Dempsey 

Intensity modulated radiotherapy (IMRT) is a powerful technique for radiotherapy 

treatment delivery. We discuss the problem of efficiently delivering a given 

beam intensity profile. In particular, we need to decompose the intensity profile 

into a set of beam shapes satisfying the consecutive-ones property. We propose 

an exact algorithm to solve this problem under different efficiency criteria. Our 

results indicate that several clinical instances can be solved to optimality within 

a few minutes. 

4 - Measuring the Effect of Workers’ Health on Productivity 

Fredrik Odegaard, Richard Ivey School of Business, University 

of Western Ontario, 1151 Richmond Street North, N6A 3K7, 

London, Ontario, Canada, fodegaard@ivey.uwo.ca, Pontus Roos 

In this paper we present a methodology to study the interaction between health 

and productivity. Health refers to the physical and mental condition of an individual, 

while productivity refers to the physical output and quality of a firm. 

Health is modeled as a latent variable, and considered in the framework of the 

capability approach. We illustrate the methodology with results from a workplace 

intervention study of four large Swedish manufacturing firms over the 

years 2000-2003. 

� TC-23 


6.2.49 

MOO: Facility Location Problems 



Chair: Carlos Ferreira, Dep. of Economics, Management and 

Industrial Engineering, University of Aveiro, Campus Universitario 

de Santiago, 3810-143, Aveiro, Portugal, carlosf@ua.pt 

1 - The embedment of a GIS in a decision support system 

for bicriteria location problems 

160 

Sérgio Fernandes, Instituto Politécnico de Setúbal, Escola 

Superior de Tecnologia, Campus do Instituto Politécnico de 

Setúbal, Estefanilha, 2910-761, Setúbal, Portugal, 

sergiof@est.ips.pt, Maria Eugénia Captivo, João Clímaco 

SABILOC is a Decision Support System aimed at supporting decision making 

concerning bicriteria location models. When the facilities to be located have 

environmental effects, they usually depend on different factors like the altitude 

and the morphology of the potential locations, the winds, the temperature, the 

humidity, etc, most of which can be more easily evaluated with a Geographical 

Information System (GIS). In this work we describe and exemplify the qualitative 

added value of the embedment of a GIS platform in SABILOC. 

2 - An interactive method for multi-objective integer and 

mixed-integer programming applied to a facility location 

problem 

Carlos Ferreira, Dep. of Economics, Management and Industrial 

Engineering, University of Aveiro, Campus Universitario de 

Santiago, 3810-143, Aveiro, Portugal, carlosf@ua.pt, Rui Borges 

Lopes, Beatriz Sousa Santos 

We propose an interactive method following an open communication protocol 

for multi-objective integer and mixed-integer programming. In each step of 

the human/computer dialogue, the decision maker (DM) provides indications 

about the sub regions he/she desires to continue the search for non-dominated 

solutions (nds). The method enables to progressively eliminate criteria regions, 

either by dominance or unfeasibility, and ends when the DM considers to have 

sufficient knowledge about the set of nds. The proposed method is applied to a 

facility location problem (the set covering problem). 

3 - A multi-objective evolutionary algorithm for the capacitated 

location-routing problem 

Rui Borges Lopes, Dep. of Economics, Management and 

Industrial Engineering, CIO / University of Aveiro, Campus 


rui.borges@ua.pt, Carlos Ferreira, Beatriz Sousa Santos 

We consider a discrete multi-objective Capacitated Location-Routing Problem 

(CLRP) with two levels (depots and clients) and a capacitated and homogeneous 

vehicle fleet. For this model, an evolutionary (genetic) algorithm is presented. 

The algorithm uses a new chromosome encoding and Pareto ranking 

for crossover selection. The method aims at obtaining a good approximation 

of the whole non-dominated set. Results for some instances adapted from the 

CLRP literature are presented and discussed. 

4 - Routing Location Multi Criteria Optimization Metaheuristics 

Claudia Margarita Villagran de Leon, Faculty of Medicine., 

University San Carlos of Guatemala, 9 Avenida 9-45 Zona 

11,CentroUniversitario, CUM,Colonia Roosevelt., 01011, 

Guatemala, Guatemala, Guatemala, 

margavilla2008@hotmail.com 

Location Routing problems appear in many applications with different constraints, 

depending on the application, which are studied by several algorithms 

in (OR). We Consider a Multi Criteria routing location problem analyzing economical 

as well as social criteria, deriving a metaheuristic solution for a Travel 

Touristic Management Plan. 

� TC-24 


6.2.50 

Data Mining in Bioinformatics 


Medicine 


Chair: Giovanni Felici, IASI, CNR, 00185, Rome, Italy, 

giovanni.felici@iasi.cnr.it 

1 - Genetic Programming for feature extraction in supervised 

learning 

Mauro Castelli, DISCO- Department Informatics, Systems and 

communication, University of Milano Bicocca, viale sarca 336,

20126, milan, castelli@disco.unimib.it, Ilaria Giordani, 

Francesco Archetti, Enza Messina, Leonardo Vanneschi 

In this talk we address the problem of reducing the dimensionality of the space 

of features for machine learning systems in case of large datasets. We propose 

a GP based approach for a combined feature extraction/classification method, 

which creates a fixed and limited number of new features, to be considered during 

the classification process. The approach uses multi-objective optimization, 

with the following criteria:classification performances;size and complexity of 

the classifier and the extracted features expressions;Entropy of the extracted 

features.Experimental results will be discussed. 

2 - Metaheuristics for sequences consensus problems 

Paola Festa, Dept. of Mathematics and Applications, University 

of Napoli Federico II, Compl. MSA - Via Cintia, 80126, Napoli, 

Italy, paola.festa@unina.it 

It has been shown that a large number of molecular biology problems can be 

formulated as combinatorial optimization problems, including sequence alignment 

problems, genome rearrangement problems, string selection and comparison 

problems, and protein structure prediction and recognition. This paper 

defines some string selection and string comparison problems, also known as 

consensus problems, and describes several metaheuristic approaches for finding 

good-quality solutions. 

3 - Feature Selection with Hyperspheres 

Paola Bertolazzi, CNR, IASI, 61980, Rome, Italy, 

paola.bertolazzi@iasi.cnr.it, Giovanni Felici, Paola Festa 

In this paper we consider a class of integer and mixed integer programming 

models used to represent Feature Selection (FS) problems in very high dimensional 

spaces. FS arises in Data Analysis and Data Mining to properly reduce 

the dimension of the space where the data are represented. Such dimensional 

reduction is performed to make the analysis tractable while retaining the largest 

amount of information, given the specific objective of the data analysis task. 

We consider a new variant where the target space can be considered an affine 

transformation of the original one where the data lie in hyperspheres of homogeneous 

classes. Such problem is modeled with linear and mixed integer 

programs and solved with commercial softwares when tractable, else with adhoc 

heuristics. The proposed space transformation presents several interesting 

features for the separation of data as it is shown in the experiments reported. 

4 - Feature Discretization and Clustering in the DMB logic 

mining environment 

Paola Bertolazzi, CNR, IASI, 61980, Rome, Italy, 

paola.bertolazzi@iasi.cnr.it, Giovanni Felici, Emanuel 

Weitschek, Guido Drovandi 

Feature Discretization is applied to the numerical attributes of a dataset and 

consists of the identification of a possibly small set of intervals of values for 

each attribute, than can then be mapped in a qualitative or binary scale. In Feature 

Clustering, the discrete features can be aggregated in clusters with common 

properties. We present two algorithms for discretization and clustering 

integrated in the DMB (Data Mining Big) software system, a logic data mining 

system for classification in large data sets. In particular, we consider biological 

data, often represented by large arrays of integer or real numbers, that correspond 

to measures on the objects (e.g. gene expressions by microarray). 

� TC-25 


6.2.48 

Volatility Spillover and Liquidity Risk 



Chair: Emrah Sener, Computational Finance, Ozyegin University, 

Kusbakisi Cad. No: 2, Altunizade, 34662, Istanbul, Turkey, 

emrah.sener@ozyegin.edu.tr 

Chair: Feyzullah Egriboyun, Finance, Sabanci University, Istanbul, 

Turkey, feyz@alumni.cmu.edu 

1 - Dynamic spillover effects of deviation in covered interest 

rate parity through term structure 

Feyzullah Egriboyun, Finance, Sabanci University, Istanbul, 

Turkey, feyz@alumni.cmu.edu, Sait Satiroglu, Emrah Sener 


We investigate if there is a dynamic relationship between the collapse of shortterm 

covered interest rate parity and the long-term cross currency basis swap 

market during the 2008 crisis. Our investigation spans an unusual period of unrest 

in the money markets of six emerging and four developed currencies. Our 

analysis suggests that the deviation in FX swap market has significant spillover 

effects on the basis swaps as a result of higher liquidity concerns. Volatility 

and mean spillovers are detected in developed countries while in emerging 

countries, volatility spillover is not observed. 

2 - Discrete and continuous-time liquidity models 

Selim Gokay, Mathematik, ETH Zurich, Ramistrasse 101, 8092, 

Zurich, Switzerland, selim.gokay@math.ethz.ch, Mete Soner 

Illiquidity risk incorporates the effect of the size and the time of the trade into 

the price process of the asset. This price impact might be permanent or temporary. 

In this talk we will address some temporary and permanent illiquidity 

models for option pricing. In particular, we will focus on the supply curve 

model introduced by Cetin, Jarrow and Protter. This is an example of a temporary 

impact model and we will explore the recent works of Cetin, Soner and 

Touzi in continuous-time and Gokay and Soner in discrete-time concerning this 

model. 

3 - Volatility Spillover Dynamics between the Emerging 

CDS, Bond and Equity Markets 

Emrah Sener, Computational Finance, Ozyegin University, 

Kusbakisi Cad. No: 2, Altunizade, 34662, Istanbul, Turkey, 

emrah.sener@ozyegin.edu.tr, Sevan Ulutas, Osman Yilmaz 

We analyze the impulse response relationship, variance decomposition and 

volatility spillover dynamics between the CDS, equity, and bond markets. In 

order to investigate the volatility transmission, we use a multivariate GARCH 

model. Our research includes three emerging markets (Turkey, Russia and 

Brazil) during the pre-crisis, crisis and post-crisis periods. The findings show 

that all three markets transmit volatility to each other and hence innovations 

from one market increase trading activity in the other two markets. 

� TC-26 


3.1.11 

Cooperative Games and Applications 



Chair: M a Luisa Carpente, Matemáticas, University of La Coruña, 

Campus de Elviña, 15071, A Coruña, luisacar@udc.es 

1 - The proportional coalitional Shapley value 

Francesc Carreras, Applied Mathematics II, Technical University 

of Catalonia, ETSEIAT, Colom 11, 08222, Terrassa, Catalonia, 

Spain, francesc.carreras@upc.edu, José María Alonso_meijide 

We propose a modification of the Shapley value for monotonic games with a 

coalition structure. The resulting coalitional value is a twofold extension of 

the Shapley value since: (1) the amount obtained by any union coincides with 

the Shapley value of the union in the quotient game; and (2) the players of the 

union share this amount proportionally to their Shapley value in the original 

game. We provide axiomatic characterizations of this value that are close to 

those existing in the literature for the Owen value and include applications to 

coalition formation in bankruptcy and voting problems. 

2 - Bilateral assignment markets with the same core 

Javier Martinez de Albeniz, Matematica Economica, Financera i 

Actuarial, Universitat de Barcelona, Av. Diagonal, 690, 08034, 

Barcelona, Spain, javier.martinezdealbeniz@ub.edu, Marina 

Nunez, Carles Rafels 

In the framework of the bilateral assignment games, we study the set of matrices 

leading to assignment markets with the same core. We state conditions, in 

terms of the matrix entries, that ensure that the related assignment games have 

the same core and we prove that the set of matrices leading to the same core 

form a join-semilattice with a finite number of minimal elements and a unique 

maximum. Conditions under which the join-semilattice reduces to a singleton 

or is in fact a lattice are also identified. 

161


3 - The Shapley value and some stability conditions 

M a Luisa Carpente, Matemáticas, Universidade da Coruña, 

Facultade de Informática. Campus de Elviña, 15071, A Coruña, 

Spain, luisacar@udc.es, Balbina-Virginia Casas-Méndez, 

Ignacio García-Jurado 

We study the Shapley value solution for coalitional games in some situations in 

which there are upper bounds on the possible payoffs for some coalitions. We 

are concerned with the notion of the core and some stability conditions derived 

from these upper bounds. For instance, minimum cost spanning tree problems 

and some sequencing situations that have been studied in the literature under a 

pessimistic and an optimistic point of view fit perfectly in our setting. 

� TC-27 


8.2.06 

FREIGHT TRANSPORTATION 

Stream: Transportation and Logistics 


Chair: Walter Ukovich, ORTS - DEEI, University of Trieste, via 

Valerio, 10, 34127, Trieste, Italy, ukovich@units.it 

1 - A simulation-optimization approach for drawing 

stowage plans for containerships 

Rina Mary Mazza, DEIS - Dipartimento di Elettronica, 

Informatica e Sistemistica, Università della Calabria, Via P. 

Bucci 41C, 87036, Rende, (CS), Italy, rmazza@deis.unical.it, M. 

Flavia Monaco, Marcello Sammarra, Gregorio Sorrentino 

We consider the Ship Stowage Plan Problem according to the terminal planner’s 

point of view. Within the global aim of reducing the time to perform loading 

operations, the problem focuses on minimizing the number of reshuffles (unproductive 

moves) required to fetch a container from a stack on the terminal 

yard. The simulation-optimization solution methodology proposed uses a tabu 

search algorithm to explore the solution space combined with a discrete-event 

simulation model used to estimate the cost of each solution, i.e. container handling 

and transfer time from the yard to the ship. 

2 - Freight transportation planning in railway networks 

with rapid transhipment terminals 

Davide Anghinolfi, Università di Genova, 16100, Genova, Italy, 

davide.anghinolfi@unige.it, Massimo Paolucci, Simona Sacone, 

Silvia Siri 

We propose a planning procedure for serving freight transportation demand in 

a railway network whose terminals are equipped with fast train to train transhipment 

devices. The demand consists of sets of boxes with different origins, 

destinations and delivery times. A plan gives for each box the routing, the 

sequence of trains and the assignment to train wagons. The planning procedure 

first finds the alternative sequences of trains for serving orders then solves 

a generalized assignment problem by a combined integer programming and 

heuristic approach. Some experimental results are shown. 

3 - A Mathematical Programming Approach to the Multi- 

Port Master Bay Plan Problem 

Massimo Paolucci, Dipartimento di Informatica, Sistemistica e 

Telematica, Universita‘ di Genova, Via Opera Pia 13, 16145, 

Genova, Italy, Italy, paolucci@dist.unige.it, Daniela Ambrosino, 

Davide Anghinolfi, Anna Sciomachen 

We face the line planner problem of determining aggregate stowage plans for 

containerships to satisfy a transportation demand. This problem, denoted as 

Multi-Port Master Bay Plan Problem (MP-MBPP), consists in determining the 

stowage plan for each port on the ship route so that stability constraints are satisfied 

and port operations are optimized, i.e., the workload for the quay cranes 

is balanced and non-productive movements are minimized. We propose a mathematical 

programming approach combined with a heuristic approximation and 

we reporting experimental results showing its effectiveness. 

4 - A Metamodeling Approach to the Management of Intermodal 

Transportation Networks 

162 

Walter Ukovich, ORTS - DEEI, University of Trieste, via 

Valerio, 10, 34127, Trieste, Italy, ukovich@units.it, Valentina 

Boschian, Mariagrazia Dotoli, Maria Pia Fanti, Giorgio 

Iacobellis, Gabriella Stecco 

The contribution specifies an Integrated System (IS) devoted to the efficient 

management and control of Intermodal Transportation Networks (ITN). The 

proposed IS is designed to take both tactical decisions, in an off-line mode, and 

operational decisions, in real time. In either case, the core of the presented IS 

is a reference model that uses information from the real system. The reference 

model is based on a metamodeling approach: a top-down procedure based on 

the UML formalism, a graphic and textual language able to describe systems 

from structural and behavioral viewpoints. To show the IS application at the 

tactical decision level, an ITN real case study is considered and simulated. 

� TC-29 


8.2.11 

Theory of Boolean Functions 



Chair: Gyorgy Turan, Math., Stat. and Comp. Sci., Univ. of Illinois 

at Chicago, 851 S.Morgan, M/C 249, 60608-7045, Chicago, IL, 

United States, gyt@uic.edu 

1 - The arity gap of functions of Boolean variables and lattice 

polynomial functions 

Miguel Couceiro, Mathematics, University of Luxembourg, 

University of Luxembourg, FSTC, 6, rue Richard 

Coudenhove-Kalergi, L-1359 , Luxembourg, Luxembourg, 

miguel.couceiro@uni.lu 

In this talk we consider the problem of determining the minimum decrease in 

the number of essential variables of a given function f when variables of f are 

identified (the so-called arity gap of f). We will present complete classifications 

of Boolean and pseudo-Boolean functions according to their arity gap. 

Using these results we shall derive an analogous classification of polynomial 

functions over arbitrary distributive lattices, showing that the only polynomial 

functions having arity gap 2 are the truncated median functions, and all other 

have arity gap 1. 

2 - Construction and learnability of canonical Horn formulas 

Marta Arias, Llenguatges i Sistemes Informàtics, Universitat 

Politècnica de Catalunya, C Jordi Girona 1-3, 08034, Barcelona, 

Barcelona, Spain, marias@lsi.upc.edu 

In the talk I will describe an alternative construction of an existing canonical 

representation for definite Horn theories, the Guigues-Duquenne basis (or GD 

basis), which minimizes a natural notion of implicational size. I will show how 

this canonical representation can be extended to general Horn, by providing a 

reduction from definite to general Horn. If time permits, I will also show a 

striking connection between a well-known learning algorithm for Horn formulae 

by Angluin, Frazier and Pitt and this canonical representation. This is joint 

work with José L. Balcázar. 

3 - Essential sets and Horn minimization 

Endre Boros, RUTCOR, Rutgers University, 08854, Piscataway, 

New Jersey, United States, Endre.Boros@rutcor.rutgers.edu, 

Ondrej Cepek, Alex Kogan, Petr Kucera 

We introduce a biology motivated classification of the implicants of a Horn 

function, and derive consequences for the minimum representations. 

4 - On the approximate minimization of Horn formulas 

Gyorgy Turan, Math., Stat. and Comp. Sci., Univ. of Illinois at 

Chicago, 851 S.Morgan, M/C 249, 60608-7045, Chicago, IL, 

United States, gyt@uic.edu 

We consider the complexity of approximately minimizing propositional Horn 

formulas. An inapproximability result is proved under a complexity theoretic 

assumption. An efficient approximation algorithm is given with a sublinear performance 

guarantee in the Steiner-minimization version, where new variables 

can be added in a restricted manner. The algorithm is based on a procedure for 

partitioning bipartite graphs. We also consider the case when only the original 

clauses are allowed to be used. 

Joint work with Amitava Bhattacharya, Bhaskar DasGupta and Dhruv Mubayi.

� TC-30 


8.2.13 

PROMETHEE: Axiomatic basis and other 

issues 


other Issues 


Chair: Gabriela Fernández Barberis, Quantitative Methods, San Pablo 

CEU University, 23, Julián Romea St., 28003, Madrid, Spain, Spain, 

ferbar@ceu.es 

1 - Semi Orders, Interval Orders and Pseudo Orders Preference 

Structures in Multiple Criteria Decision Aid 

Method 

Gabriela Fernández Barberis, Quantitative Methods, San Pablo 

CEU University, 23, Julián Romea St., 28003, Madrid, Spain, 

Spain, ferbar@ceu.es 

During the last decades, an important number of Multicriteria Decision Aid 

Methods have been proposed to help the decision maker to select the best 

compromise alternative. Meanwhile, the PROMETHEE family of outranking 

method has attracted much attention from academics and practitioners. In this 

paper, an extension of these methods is presented, consisting of analyze its 

functioning under New Preference Structures (NPS). The preference structures 

tale into account are, namely: semi-orders, intervals orders and pseudo-orders. 

2 - D-sight: a new decision aid software 

Quantin Hayez, SMG - Faculty of Engineering, ULB - Brussels 

Free University, Boulevard du TRiomphe CP210/01, 1050, 

Brussels, Belgium, qhayez@ulb.ac.be, Yves De Smet, Bertrand 

Mareschal 

D-Sight is a new software that implements the PROMETHEE & GAIA methods, 

including their newer extensions. It introduces new visual modelling tools 

and graphical representations of the results of the multicriteria analysis. The 

interaction with the decision-maker has been emphasized through several sensitivity 

analysis tools. Numerical examples will be used to illustrate the features 

of the software. 

� TC-31 


8.2.15 

Network Planning in Postal Logistics 

Stream: OR Applications in Industry 


Chair: Hans-Jürgen Sebastian, Deutsche Post Endowed Chair of 

Optimization of Distribution Networks, RWTH Aachen University, 

Templergraben 64, 52062, Aachen, Germany, 

sebastian@or.rwth-aachen.de 

1 - Allocation of international long-haul transportation 

lanes to sorting centers at Deutsche Post 

Christoph Hempsch, Deutsche Post DHL, Sträßchensweg 10, 

53113, Bonn, Germany, Christoph.Hempsch@deutschepost.de, 

Andreas Marschner, Thomas Müller 

Deutsche Post delivers about 2.5 million parcels in Germany every day. A 

smaller volume is exported to European countries via an international long-haul 

road network. The preparation for export of parcels for a European country is 

concentrated at one of the 33 national parcel sorting centers. Thus, international 

long-haul lanes to gateways in this country are starting at the corresponding 

national sorting center. The talk presents a mixed-integer model formulated 

to determine the potential of alternative allocations of the international lanes 

under demanding service-level requirements. 

2 - A Strategic Model to Simultaneously Find Optimal Locations, 

Allocations and Flows in Complex Network Environments 

Thomas Müller, Deutsche Post Chair of Optimization of 

Distribution Networks, RWTH Aachen University, 



Mueller@or.rwth-aachen.de, Julia Hillebrandt 

We observe a very large-scale distribution network where distinguishable items 

have to be transported from sources to sinks. Since the total quantity is very big, 

but the quantity going from a certain source to a sink is quite small, it is useful 

to consolidate item flows in hubs. We introduce a strategic network design 

model that simultaneously finds optimal locations for consolidation facilities, 

optimizes network flows going through the facilities, and allocates sources and 

sinks to these flows. Even big instances are solved to optimality using a mixed 

integer linear programming solver. 

3 - Scheduling of Long-Distance Transports in a Service 

Network with Capacity and Time Constraints 

Tobias Winkelkotte, Deutsche Post Chair of Optimization of 

Distribution Networks, RWTH Aachen University, 


winkelkotte@or.rwth-aachen.de, Li Sun 

We talk about tactical transportation scheduling in logistics networks where locations 

and allocations are given. Due to vehicle capacities, there are usually 

several transports on each connection, which have to be achieved in such a short 

time period that some distances cannot be overcome within this time. Therefore, 

we introduce service level constraints, such that only a certain fraction of 

the transports must arrive on time. Big instances of the model are solved, and 

numerical results are presented. 

4 - Optimization of Postal Transportation Networks based 

on Opposite Travel Mode 

Li Sun, Institute of System Engineering, Southeast University, 

Nanjing, China, sunsuper@126.com 

China’s postal transportation network is a multi-layer network based on the 

Postal Center Office System. It is needed to design an appropriate structure for 

each sub-network according to its unique characteristics. In order to lower the 

unloaded ratio of vehicles during transportation, we apply the opposite travel 

mode (OTM) to structure the delivery network in local postal areas. Therefore 

we develop a non-linear 0-1 planning model and a corresponding heuristic algorithm 

to improve the operating costs of local postal networks. A simulation 

test is also presented. 

� TC-32 


8.2.17 

OR in Agriculture and Forest Management 

Stream: OR in Agriculture and Forest Management [c] 


Chair: Paulo Borges, University Of Lisbon, Lisbon, Portugal, 

pjaborges@gmail.com 

1 - Multiple criteria and multiple objectives decisionmaking 

for a Brazilian sugar and ethanol milling company 

Fernando Marins, Production, UNESP - São Paulo State 

University, Av. Ariberto Pereira da Cunha, 333, 12516-410, 

Guaratinguetá, SP, Brazil, fmarins@feg.unesp.br, Valério 

Salomon, José Arnaldo Montevechi, Aneirson Silva 

This paper presents a Lexicographic Goal Programming model combined with 

Analytic Hierarchy Process, for the aggregate production planning, commercialization, 

and distribution in a Brazilian sugar and ethanol milling company. 

The model depicts the comprehensive production process of sugar, ethanol, 

molasses and derivatives, including decisions on agricultural steps such as cutting 

phases, sugarcane load and transport — and on industrials steps, mainly 

those related to crushing, type of production process, weekly production and 

storage, as well as the distribution and commercialization stages. 

2 - Modelling farmers’ decision-making using multiattribute 

utility theory: a comparative analysis 

Laura Riesgo, Dept. of Economics, Pablo de Olavide University, 

Ctra. Utrera km.1, 41013, Seville, Spain, laurariesgo@upo.es, 

José A. Gómez-Limón, Jordi Gallego 

163


The aim of this paper is: a) Can farmers’ MAUFs be considered as structural 

elements (constant in the short and medium term)?, and b) among the different 

MAUF specifications already implemented in empirical studies, which one better 

fit farmers’ behavior? Only answering the first question affirmatively, the 

MAUT can be considered a useful approach to simulate ex-ante agricultural 

systems’ response to hypothetic future scenarios (policy reforms, etc). If this is 

the case, the response of the second question could help analysts to build more 

accurate models choosing the most realistic MAUF. 

3 - A Model for Participative Forest Management in the 

Mediterranean Area 

Concepcion Maroto, Estadistica e Investigacion Operativa, 

Universidad Politecnica de Valencia, Camino de Vera S/N, 

46022, Valencia, Valencia, Spain, cmaroto@eio.upv.es, 

Concepción Ginestar, Juan Uriol, Baldomero Segura, Juan 

Fernando Usó 

Decision making in the strategic management of public forest is a complex 

problem due to two factors. Firstly, public administrations need to balance 

several economic, environmental and social criteria and secondly, they should 

carry out a participation process with the different social groups or stakeholders 

involved. In this work we propose a model for strategic management of 

public forests in a Mediterranean area. This model integrates multiple criteria 

and group decision making techniques. 

4 - Decision Analysis to address risk when assessing impacts 

of Agro Forestry Policies and Economic Scenarios 

- a Case Study in Southern Portugal. 

Paulo Borges, University Of Lisbon, Lisbon, Portugal, 

pjaborges@gmail.com, Brigite Botequim, Jose Borges, Jordi 

Garcia_Gonzalo, Rui Fragoso 

This paper demonstrates the use of a Decision Analyses approach to address 

risk when assessing impacts of Common Agricultural Policy (CAP) and/or of 

prices on land use patterns in rural areas. For testing and demonstration purposes, 

this research considered an application to the Alentejo, region in Southern 

Portugal, encompassing 31 farm types extending over 2106 ha. The approach 

was integrated in a Decision Support System (DSS) to generate scenarios 

and corresponding decisions and to analyze solutions. Results show the 

usefulness and relevance of the proposed approach. 

� TC-33 


8.2.19 

Energy Planning Models 

Stream: Energy, Environment and Climate 


Chair: Carlos Henggeler Antunes, DEEC, University of Coimbra and 

INESC Coimbra, Rua Antero de Quental 199, 3000-033, Coimbra, 

Portugal, ch@deec.uc.pt 

1 - Risk Diffusion in Natural Gas Networks 

Parviz Darvish, Operations Management, ESSEC Business 

school, Avenue Bernard Hirsch, BP 50105, Cergy, 95021, Cergy 

Pontoise cedex, Val d’Oise, France, parviz.darvish@essec.fr, 

Fernando Oliveira 

Natural gas is a strategic source of energy to Europe as in many countries, it 

is essential for producing electricity, residential and industrial consumption. 

The infrastructure required to store, transfer, distribute and deliver natural gas 

to consumers is complex and exposed to many risks including demand uncertainty 

and supply disruptions due to natural disasters, terrorist attacks, or political 

conflicts. In this paper we study alternative topologies of the network, 

analyzing how risk averse, profit-maximizing firms, are affected by different 

risk factors in different areas of the network. 

2 - A multi-sectoral multiobjective model with interval coefficients 

to study E3 interactions 

164 

Carla Henriques, INESC Coimbra, Rua Antero de Quental, 199, 

3000-030, Coimbra, Portugal, coliv@inescc.pt, Carlos 

Henggeler Antunes 

Multi-sectoral economy-energy-environment (ME3) models enable a prospective 

analysis of the economic structure and the energy system, including the 

environmental impacts, for decision support in policy making. Multiobjective 

(MO) approaches provide a framework for dealing with the conflicting axes of 

evaluation in sustainability problems. MO linear programming models based 

on input-output analysis are used to study the interactions between those concerns. 

A MOLP ME3 model with interval coefficients to capture the uncertainty 

of the coefficients is proposed for assessing E3 impacts of distinct policies. 

3 - Flexible design of sustainable energy systems under 

uncertainty 

Gonçalo Pereira, DEM - Departamento Engenharia Mecânica, 

Instituto Superior Técnico, Lisbon, Portugal, 

goncalo.duarte.pereira@ist.utl.pt, Alexandra Moutinho, Carlos 

Silva 

The design of sustainable energy systems requires that decision makers have 

access to a planning tool to evaluate different investment scenarios, taking into 

account the uncertainty of different parameters in order to decide upon what, 

how and when investments should be made. This paper presents a long-term 

energy planning tool that proposes flexible investment strategies, allowing the 

evaluation of which is the best option under different scenarios that describe 

different uncertain parameters. The tool uses a short-term energy model to take 

into account the dynamics of renewable resources. 

4 - Analyzing the Impact of Wind Power in the Electricity 

Market Prices of Portugal 

Ruben Ramalho, IN+ Center for Innovation, Universidade 

Técnica de Lisboa, Taguspark Campus of IST - Av. Prof. Dr. 

Aníbal Cavaco Silva, 2744-016, Porto Salvo, Portugal, 

ruben.s.ramalho@ist.utl.pt, Filipa Amorim 

The main goal of this paper relies in determining this balance for the wind 

power technology in the Portuguese wholesale electricity market. A secondary 

goal is to determine the economic benefits of wind power technology related to 

GHG emissions and associated costs for the Portuguese power system. Finally, 

a third goal is to make an overall balance considering the impact of wind technology 

in the total costs of the Portuguese power system that include private 

investment in generation capacity, RES support schemes costs, operation costs 

and the wind effect in the wholesale market prices. 

� TC-34 


8.2.23 

Generalized Convexity I 

Stream: Convex Optimization 


Chair: Gyula Maksa, Department of Analysis, University of 

Debrecen, Institute of Mathematics, University of Debrecen, 4010, 

Debrecen, Hungary, maksa@math.klte.hu 

1 - Pseudoconvexity of the ratio between a quadratic function 

and an affine function on the non-negative orthant 

Laura Martein, department of statistics and applied mathematics, 

University of Pisa, via Ridolfi, 10, 56124, Pisa, Italy, 

lmartein@ec.unipi.it 

The pseudoconvexity of the ratio f of a quadratic function and an affine one has 

been recently characterized on the domain D corresponding to the positivity 

of the denominator. Motivated by the fact that in optimization problems the 

decision variables are often required to be non-negative, in this paper we will 

characterize, firstly, the maximal domains of pseudoconvexity of f and, successively, 

we will specialize the obtained results in order to obtain conditions 

which guarantee that the maximal domains of pseudoconvexity of f contain 

the non-negative orthant. At last, some methods are suggested for constructing 

pseudoconvex functions which are the ratio of a quadratic function and an 

affine one. 

2 - A class of generalized convex functions: optimality and 

duality results 

Laura Carosi, Statistics and Applied Mathematics, University of 

Pisa, Via Ridolfi, 10, 56124, Pisa, Italy, lcarosi@ec.unipi.it, 

Riccardo Cambini

We introduce a class of generalized convex functions and we analyze its relationships 

with the generalized convex properties commonly studied in the 

literature. In particular, the proposed class is compared with the ones of invex 

and generalized invex functions. The new property is used to derive necessary 

and sufficient optimality conditions and to obtain duality results for vector 

optimization problems. 

3 - Bernstein-Doetsch type results for h-convex functions 

Attila Házy, Department of Applied Mathemaics, University of 

Miskolc, Miskolci Egyetem, 3515, Miskolc-Egyetemváros, 

Hungary, matha@uni-miskolc.hu 

In our talks we introduce the more general concept of the h-convexity, and the 

concept of the so called (H,h)-convexity. This type of h-convexity is a common 

generalization of the usual convexity, the Godunova-Levin functions, the 

Breckner s-convex functions and the so called P-functions. 

The main goal of the talk is to prove some regularity and Bernstein-Doetsch 

type result for h-convex and (H,h)-convex functions. We also collect some 

facts on such functions and collect some interesting, easily-proved properties 

of h-convex functions. 

4 - Regularity and convexity results on approximately hconvex 

functions 

Pál Burai, Applied Mathemathics and Probability Theory, 

University of Debrecen, Faculty of Informatics, Pf.: 12., 4010 , 

Debrecen, Hungary, burai.pal@inf.unideb.hu, Attila Házy 

In 1915 Bernstein and Doetsch proved the following: if a Jensen-convex function 

is bounded from above at a point of its domain, then it is continuous on 

the whole domain and convex. The main goal of this talk to prove a Bernstein- 

Doetsch type result on approximately h-convex functions. 

� TC-35 


6.2.46 

Researching Facilitated Modelling 



Chair: Etiënne Rouwette, Thomas van Aquinostraat 1.2.33, PO Box 

9108, 6500 HK, Nijmegen, E.Rouwette@fm.ru.nl 

1 - Shifting Perspectives from the Individual- to the Team- 

Level: The Role of Mental Model Convergence 

Sara McComb, Texas A&M University, United States, 

mccomb@tamu.edu 

How do individuals become team members? Why do teams that appear to 

be similar have very different outcomes? Understanding mental model convergence 

may provide insight into these types of questions. Individual team 

members create cognitive frameworks that facilitate team collaboration and 

converge across members over time. This paper describes the convergence 

process, as it may explain how team members shift their perspectives from the 

individual- to the team-level, and presents empirical results demonstrating the 

relationship between the convergence process and effective performance. 

2 - Exploring the role of Cognitive Style on the Structuring 

and Definition of Ill-structured Management Problems. 

L. Alberto Franco, Warwick Business School, University of 

Warwick, ORIS Group, Gibbet Hill Road, CV4 7AL, Coventry, 

United Kingdom, alberto.franco@warwick.ac.uk, Luiz Felipe 

Nasser-Carvalho 

We report preliminary findings from a quasi-experimental field study on the 

role of managers’ Need for Cognition and Need for Closure on characteristics 

of their: a) causal problem representations, b) problem definitions and, c) 

attitudes towards structuring ill-structured problems. 

The findings suggest that while Need for Cognition influences positively managers’ 

confidence in, and enjoyment of their structuring efforts, Need for Closure 

hinders managers’ understanding of ill-structured problems. Implications 

for research and practice of Problem Structuring Methods are discussed. 

3 - Strategy Maps as a Tool to Support Strategic Management 

— How to Optimally Map Your Strategy 

Melanie Windolph, Chair of Management Accounting and 

Control, University of Goettingen, Platz der Goettinger Sieben 3, 


37073, Goettingen, Germany, 

melanie.windolph@wiwi.uni-goettingen.de, Judith Huelle, York 

Hagmayer, Klaus Moeller 

The study’s purpose is to analyze main problems when creating strategy maps 

for supporting strategy’s implementation. To examine key problems in the creation 

process, we conducted two experiments with students from the university 

of Goettingen and managers from the automotive industry. The results indicate 

that while participants did choose strategic goals covering global and specific 

aspects of the firm’s strategy, they repeatedly failed in defining the main relations 

between the goals. Furthermore, participants’ self-assessment was significantly 

better than the maps’ objective evaluation. 

4 - Decision Development in Facilitated Modelling Wokshops 

Etienne Rouwette, Nijmegen School of Management, Radboud 

University Nijmegen, 6500 HK, Nijmegen, Netherlands, 

e.rouwette@fm.ru.nl, L. Alberto Franco 

Facilitated modelling has its most direct impact on the way a convened group of 

participants produce their decisions. By facilitating the decision development 

process, and capturing the content of group discussions in the form of models, 

we attempt to provide effective group decision making support. Despite its 

central role in assisting the group decision development process, limited attention 

has been paid to this dimension in the literature of facilitated modelling. 

Indeed, although a large number of published cases studies of facilitated modelling 

interventions are available, the majority of these studies do not provide 

any detail below the level of the modelling workshop; that is, modelling procedures 

and general outputs and outcomes may be described but there is no 

portrayal of the decision development process within the workshop. To address 

this gap, we map the different conceptualisations of decision development that 

seem embedded within the facilitated modelling tradition, and contrast them 

with well-established theoretical models from the group communication field. 

Our analysis identifies a number of research possibilities for the study of facilitated 

modelling workshops from a decision development perspective, and 

suggests a research strategy that can help to further develop facilitated modelling 

theory and practice. Central to this strategy are interaction coding and 

analysis methods for the examination of facilitated modelling workshops. By 

adopting such a research strategy we show how a decision development focus 

can increase our understanding of the rich and complex nature of facilitated 

modelling ’as it happens’. 

� TC-36 


3.1.05 

Fuzzy Decision Making and Applications 




Chair: Burcu Caglar, Industrial Engineering, Uludag University, 

Uludag University Dept. of, Industrial Engineering Gorukle Campus, 

16059, Bursa, Turkey, burcucaglar@gmail.com 

1 - Fuzzy Reactive Project Scheduling 

Dorota Kuchta, Informatics and Management, Wroclaw 

University of Technology, ul. Smoluchowskiego 25, 50-372, 

Wroclaw, Poland, dorota.kuchta@pwr.wroc.pl 

In the literature no methods of project schedule control and modification (i.e 

for reactive project scheduling) are known for the case of fuzzy activities duration 

times. We will propose a method of performing project control at regular 

time intervals and of updating the project schedule using updated information 

about the actual crisp duration of finished activities and about the(less and less) 

fuzzy duration of unfinished activities, as well as about the (also less and less) 

fuzzy number of available resources. 

2 - Determining Strategic Priorities With Fuzzy TOPSIS 

Method 

Ihsan Yüksel, Busines Administration, Kirikkale University, 

Kırıkkale University, IIBF, 71450 , Kırıkkale, Turkey, 

yuksel@kku.edu.tr, Metin Dagdeviren, Erdem Aksakal 

165


Companies, under the series affect of inhibitive and incentive factors, try to 

achieve their goals with limited financial (stock, equity capital, business capital) 

and non-financial (human resources, knowledge, competence) resources. 

This can be possible after determining the strategic priorities of the company. 

The aim of this study is to determine the company strategies with fuzzy TOPSIS 

technique which depends on the strengths and weaknesses caused by internal 

environment, opportunities and threats caused by external environment. 

3 - Multifactor Bullwhip Effect Analysis Using Fuzzy Model 

Ronay Ak, Industrial Engineering, Istanbul Technical University, 

Istanbul Teknik Universitesi, Isletme Fakultesi, Endustri 

Muhendisligi Bolumu, 34367, Istanbul, Turkey, 

ronay_ak@yahoo.com, Gulgun Kayakutlu, Cafer Erhan Bozdag 

Bullwhip Effect (BWE) studies are focused on either the causes or the solutions 

of demand fluctuations. This study aims to measure the effects of demand and 

lead time fluctuations on a typical Supply Chain (SC) of three echelons representing 

supplier, manufacturer and retailer. In such a Supply Chain, demand 

is analyzed upstream despite the lead time flows downstream. Contribution of 

this study is the analysis of integrated fuzzy demand-lead time model. The 

proposed model will give an opportunity for better Supply Chain plans. 

4 - Statistical Procedures for Robotic Assembly Line Balancing 

Problems 

Burcu Caglar, Industrial Engineering, Uludag University, Uludag 

University Dept. of, Industrial Engineering Gorukle Campus, 

16059, Bursa, Turkey, burcucaglar@gmail.com, H. Cenk 

Özmutlu, Ali Yurdun Orbak, Seda Ozmutlu 

Robotic systems have been an essential part of assembly lines, because of their 

advantages such as flexibility and automation.This study presents an application 

of type II robotic assembly line balancing (rALB-II) problem, in which the 

assembly tasks have to be assigned to robots.In order to maintain a balanced 

workload while achieving a desired production cycle time, a fuzzy clustering 

based algorithm is employed for the job assignment problem. The proposed 

algorithm is applied to a real robotic assembly line system and its advantages 

over the existing system is explained. 

� TC-37 


3.1.09 

Models for Decision Making & Decision 

Analysis 

Stream: Decision Support Systems 


Chair: Fatima Dargam, SimTech Simulation Technology, Ries Strasse 

120, 8010, Graz, Austria, F.Dargam@SimTechnology.com 

Chair: Pascale Zaraté, Institut de Recherche en Informatique de 

Toulouse, Toulouse University, 118 route de NarBonne, 31062, 

Toulouse, France, zarate@irit.fr 

1 - Influence of the criteria in the bayesian ahp 

Pilar Gargallo, Facultad de Económicas, Universidad de 

Zaragoza, Gran Vía 2, 50005, Zaragoza, Spain, 

pigarga@unizar.es, José María Moreno-jimenez, Alfredo 

Altuzarra 

This work proposes different methodologies for measuring the influence of a set 

of criteria on the final priorities of the Analytic Hierarchy Process (AHP) in a 

global context (a hierarchy). The priorities have been obtained by means of the 

Bayesian prioritization procedure of Altuzarra et al. (2007). Cross-validation 

methods have been used when measuring the influence. The methodology is 

illustrated by means of an empirical example. 

2 - Structuring and assessing large and complex decision 

problems using MCDA 

Michael Bruhn Barfod, Department of Transport, Technical 

University of Denmark, Bygningstorvet, Building 115, DK-2800, 

Kgs. Lyngby, Denmark, mbb@transport.dtu.dk, Steen Leleur 

This paper presents an approach for the structuring and assessing of large and 

complex decision problems using multi-criteria decision analysis (MCDA). 

The MCDA problem is structured in a decision tree and assessed using the 

REMBRANDT technique featuring a procedure for limiting the number of pair 

wise comparisons. A case study dealing with the structuring and prioritisation 

of projects from a large pool with limited funds is used for illustrating the approach. 

Finally, strengths and weaknesses in the MCDA approach are discussed 

and conclusions are made. 

166 

3 - Transitioning the Multistatic Tactical Planning Aid 

(MSTPA) towards Decision Support 

Christopher Strode, Systems and Technology, NATO Undersea 

Research Centre, Viale S. Bartolomeo 400, 19126, La Spezia, 

Italy, strode@nurc.nato.int 

The MSTPA tool is a multistatic sensor model able to determine the probability 

of detection, track holding, and classification of a target. This report surveys 

the approaches to be taken in order to transition the model towards decision 

support, that is, one that not only determines a performance measure for a proposed 

network geometry, but one that must determine an optimum geometry 

for a given scenario. This transition will require the addition of data mining, 

optimization and game theory modules to assist the operator in making the most 

informed decision. 

� TC-38 


6.2.44 

Stochastic Valuation of Derivatives and 

Commodities II 






Chair: Henrik Andersson, Accounting, Stockholm School of 

Economics, P.O. Box 6501, 113 83, Stockholm, Sweden, 

Henrik.Andersson@hhs.se 

1 - Pricing and Positioning of Remanufactured Products 

Using a Nested Logit Model 

Necati Aras, Industrial Engineering, Bogazici University, Bebek, 

34342, Istanbul, Turkey, arasn@boun.edu.tr, Esra Mutlu, I. 

Kuban Altinel 

We focus on the selection of remanufactured products to be offered by a firm 

in addition to the existing brand-new versions. We develop a mixed-integer 

nonlinear programming formulation using nested logit model to determine the 

product’s best prices so that the firm’s profit is maximized. The model is solved 

by decomposing it into two sub-problems. The pricing sub-problem is solved 

by a modified simplex search procedure whereas the product selection problem 

is solved via complete enumeration. Using an extended model, we also find the 

optimal quality level of the remanufactured products. 

2 - Comparative Evaluation of the Unique Elements in the 

Evolution of e- and m- Auctions 

Charis Marentakis, Dept. of Industrial Management and 

Technology, University of Pireaus, Karaoli and Dimitriou Street, 

18534, Pireaus, Greece, chmarent@unipi.gr, Dimitrios Emiris 

The evolution of auctions conducted over the Internet and mobile networks 

(e and m auctions, resp.) created an interdisciplinary research area, combining 

Economics, OR, Information Technology and Communications (ICT). We 

bridge the gap between design mechanisms and ICT infrastructure. It employs 

an evolutionary, 3-level auction classification model to progressively determine 

the mechanism and adapts it to an e and/or m context. The unique design elements 

for each auction class are compared. Findings are tested and evaluated 

in a pragmatic freight transport services setting. 

3 - When operating cash flows are mean-reverting 

Henrik Andersson, Accounting, Stockholm School of 

Economics, P.O. Box 6501, 113 83, Stockholm, Sweden, 

Henrik.Andersson@hhs.se 

This paper deals with investment analysis of mean-reverting cash flows and 

changing levels of utilization in the forest industry. Generally, mean-reversion 

enhances the value of a basic project and reduces the value of managerial flexibility 

due to less uncertainty. However, the outcome is crucially dependent 

upon whether the equilibrium price is above or below today’s price. A generalized 

version of the Feynman-Kac formula is applied but as option modeling is 

no success in the business community, a rough cut approximation is suggested 

as a comparison to standard DCF-valuation.

4 - Progressive Design of Auctions for Freight Transportation 

Services 

Ilias Petridis, Economic&Business Strategy, University of 

Piraeus, 94, Thoukididou str, 17455, Alimos - Athens, Greece, 

ilias_lat@yahoo.gr, Dimitrios Emiris, Charis Marentakis 

Pricing and service composition through auctions is of great importance for the 

major stakeholders in the freight business, carriers and shippers. We stress the 

need for systematic design of complex auction mechanisms for freight transportation 

and describe a progressive auction mechanism development model. 

The model may be used as an evolutionary design template to standardize the 

auction design process, as a tool for the logistician to initiate an auction, to 

procure transportation services or to participate in auction-based marketplaces 

to trade for logistics services. 

� TC-39 


6.2.45 

Optimal Control: Recent Advances I 






1 - Optimal Pricing of a Conspicuous Product During a Recession 

Gustav Feichtinger, Institute for Mathematical Methods in 

Economics, Vienna University of Technology, Argentinierstr. 

8/105-4, 1040, Vienna, Austria, gustav@eos.tuwien.ac.at, 

Jonathan Caulkins, Dieter Grass, Peter M. Kort, Richard Hartl, 

Andrea Seidl 

The paper considers the problem of how to price a conspicuous product, when 

the economy is in recession, which is modeled as having two effects: it reduces 

demand and it freezes capital markets so borrowing is not possible. The firm 

faces the following trade-off. On the one hand it feels pressure to reduce the 

price to maintain sales volume in the face of reduced demand. On the other 

hand, reducing the price damages brand image and thus long term demand. It 

will be shown that a decision maker has to adapt the pricing strategy according 

to the severity of the recession. 

2 - A Maximum Principle in Mixed Constrained Optimal 

Control Problems under Weakened Assumptions of 

Regularity 

Dmitry Karamzin, FEUP, University of Porto, Porto, 4200-537, 

Porto, Portugal, dmitry_karamzin@mail.ru, Aram Arutyunov, 

Fernando Pereira 

In the present work optimal control problems with mixed constraints are investigated. 

A novel weakening of the conventional regularity assumptions on 

mixed constraints is introduced. A maximum principle is derived in which the 

maximum condition is of nonstandard type: the maximum is taken over the 

closure of the set of regular points, but not over the whole feasible set. 

3 - Time Reparameterization in Optimal Control of Hybrid 

Dynamic Systems 

Maxim Staritsyn, Laboratory of Systems Analysis and 

Computational Methods, Institute for System Dynamics and 

Control Theory of Siberian Branch of Russian Academy of 

Sciences, 664033, Lermontov St., 134, Irkutsk, Russia, 664033, 

Irkutsk, Russian Federation, starmax@bk.ru 

We address an optimal control problem for a hybrid dynamic system where the 

jumps of a trajectory may occur only if it hits a given surface. Such a system 

can be regarded as an impulsive one. A discontinuous time reparameterization 

techniques is suggested to reduce this problem to a problem with bounded controls 

under phase and functional constraints. By extending a solution concept 

we show that the reduced problem is equivalent to the optimization in the class 

of generalized solutions to the hybrid dynamic system. 


4 - Sustainability, Optimality, and Viability in the Ramsey 

model 

Noel Bonneuil, Ined-Ehess, 133, bld Davout, 75980, Paris, 

France, bonneuil@ined.fr 

Viability in the Ramsey model is presented with a constraint of minimal consumption, 

then with an additional criterion of economic sustainability. The 

comparison of viability kernels with or without sustainability shows how much 

consumption should be reduced and when. The viable-optimal solution in the 

sense of inter-temporal consumption is obtained on the viability boundary of 

an auxiliary system. Technological progress works against population growth 

to favor the possibility for a given state of being viable or viable-sustainable. 

� TC-41 


3.1.06 


II 



Chair: Markus Schwaninger, Institut für Betriebswirtschaft, 

Universität St.Gallen, St.Gallen, Switzerland, 

markus.schwaninger@unisg.ch 

1 - Exploring Anti Counterfeiting Strategies: Making the 

Case for Quantitative Strategy Evaluation and System 

Dynamics 

Oliver Kleine, Industry and Service Innovations, Fraunhofer 

Institut for Systems and Innovation Research, Breslauer Straße 

48, 76139, Karlsruhe, Germany, oliver.kleine@isi.fraunhofer.de, 

Marcus Schröter 

Today, product counterfeiting and piracy are fully recognized as essential business 

risks to nearly any industry. However, as this scourge is prevailing, the 

strategies proposed to counter this threat often fail to yield the expected success 

when deployed in a company’s specific business context. Besides still 

open issues in understanding the dynamic complexity of this phenomenon, it 

is in particular decision support tools that stand for an immediate management 

need. This paper elaborates on a previous contribution and extents the presented 

system dynamics framework for strategy evaluation. 

2 - Urban water consumption management Using System 

dynamics 

M Reza Abdi, School of Management, Bradford University, 

Emm Lane, BD9 4JL, Bradford, Wet Yorkshire, United 

Kingdom, r.abdi@bradford.ac.uk 

The purpose of this paper is to develop a methodology for monitoring water 

consumption that highlights possible water saving strategies. The paper 

presents the process of building a system dynamics model of water consumption 

management using the VENSIM software. The model provides a formal 

causal-descriptive framework along with computer simulation for the analysis 

of dynamic, complex and socio-economic water consumption problems, which 

includes feedback loops and dynamic relationships over the time. In particular, 

the computer simulation methodology provides an experimental platform 

for the water consumption problems through a case study, and illustrates engagement 

of managerial roles with the society to establish the water efficiency 

strategies. 

3 - Modelling of the Economic Crisis: Have We Learn the 

Lession? 

Markus Schwaninger, Institut für Betriebswirtschaft, Universität 

St.Gallen, St.Gallen, Switzerland, 

markus.schwaninger@unisg.ch, Stefan Groesser 

The current economic crisis could deliver valuable lessons for economic agents. 

However, it seems that those have not learned essential lessons, continuing their 

"business as usual". A dynamic simulation model presented in this chapter 

highlights that this is likely to lead to the next crunch in the offing. Even before 

we have mastered this crisis the next one is already looming. Does prevention 

have a chance? How can it be achieved? 

167


� TC-42 


3.1.07 

Data Mining and Forecasting 

Stream: Data Mining and Applications [c] 








1 - Improving Forecasting Performance: A Meta-Model Approach 

Coskun Hamzaçebi, Industrial Engineering, Karadeniz Technical 

University, Kanuni Campus, 61000, Trabzon, Turkey, 

hamzacebi@ktu.edu.tr, Alper Aytekin, M.Fatih Bayramoglu 

Although there are some techniques used for time series forecasting, none of 

them is not the best in all situations. In this study, in order to improve forecasting 

performance an easy meta-model is proposed. Proposed model is tested 

with real-world and simulation time series. 

2 - Forecasting consumption of bulk customers with frequent 

tank level readings 

Natanel Sadres, Artelys, 75002, Paris, France, 

natanel.sadres@artelys.com, Emmanuelle Patay, Louis-Philippe 

Kronek, Marie-Eléonore Marmion 

Air Liquide distribution of bulk products, as Oxygen, is structured by areas 

grouping hundreds of customers for which required deliveries have to be determined 

every day. It is critical to maintain a minimum quantity of product 

within vessels. Hence, it is crucial to have a continuous estimation of the 

consumption of each customer. Air Liquide and Artelys co-developed a forecasting 

method for this real-life challenge. An exponential smoothing model 

completed with a linear auto-regressive model has been coupled with physicalbased 

pre-processing rules to provide an efficient adaptive method. 

3 - Model generation for equity-futures spread forecasting 

Roman Sologub, Innovations and High Technology, Moscow 

Institute of Physics and Technology, 508, 86, Altufievskoe sh., 

Moscow, 127349, Moscow, alucardische@gmail.com 

The investigated problem is to make a short-term forecast of the equity futures 

price relative difference. This forecast is a part of the optimal in-day statistical 

arbitrage trade strategy. To make this forecast we use the model-generation 

approach. A model maps historical prices to future spread values. The model 

is defined by a set of superpositions of smooth functions. The model quality is 

calculated by the back-test. The model of the optimal structure was deployed 

on the real market. The results are compared with government bonds. 

� TC-43 


8.2.02 

Global Optimization 1 

Stream: Global Optimization 


Chair: Leocadio G. Casado, Computer Architecture and Electronics, 

Universidad de Almeria, Ctra Sacramento s/n, La Canada de San 

Urbano, 04120, Almeria, Spain, leo@ual.es 

1 - On interval Branch-and-Bound for separable functions 

with common variables 

168 

Jose Luis Berenguel, Computer Architecture and Electronics, 

Universidad de Almería, Ctra. Sacramento s/n, La Cañada de 

San Urbano, 04120, Almeria, Spain, jlberenguel@gmail.com, 

Leocadio G. Casado, Eligius M.T. Hendrix, I. Garcia, Frederic 

Messine 

Interval Branch-and-Bound methods are powerful methods which aim for guaranteed 

solutions of Global Optimization problems. The computational effort to 

reach this aim increases exponentially with the problem dimension in the worst 

case. For separable functions this effort is less as lower dimensional subproblems 

can be solved individually. We investigate possibilities to design specific 

methods for cases where the objective function can be considered separable, 

but common variables occur in the subproblems. 

2 - Multidimensional Scaling Based on Universal Evolutionary 

Global Optimizer and Quadratic Programming 

Juana López-Redondo, Computer Architecture and Electronics, 

University of Almeria, Carretera Sacramento, S/N, 04120, 

Almeria, Spain, jlredondo@ual.es, Pilar M. Ortigosa, Julius 

Zilinskas 

Multidimensional scaling is a technique for exploratory analysis of multidimensional 

data defined by pairwise dissimilarities between objects. An essential 

part of the technique is optimization of a continuous function possessing 

many optimization-adverse properties. In this work a two-level optimization 

algorithm has been applied with universal evolutionary global optimizer in the 

upper level and quadratic programming at the lower level. The developed algorithm 

has been experimentally investigated and the results show that it performs 

better than other algorithms tested. 

3 - Copositivity and constrained fractional quadratic problems 

Paula Amaral, Faculdade de Ciências e Tecnologia, Universidade 

Nova de Lisboa, Departamento de Matemática, Campo da 

Caparica, 2829-516, Caparica, Lisbon, Portugal, 

paca@fct.unl.pt, Immanuel Bomze, Joaquim Judice 

Completely Positive (CpPP) and Copositive Programming (CoP) formulations 

for the Constrained Fractional Quadratic Problem (CFQP) and Standard Fractional 

Quadratic Problem (StFQP) are introduced. Dual and Primal attainability 

are discussed. Semidefinite Programming (SDP) formulations are proposed for 

finding good lower bounds to these fractional programs. A global optimization 

branch-and-bound approach is proposed for the StFQP. Applications of 

the CFQP and StFQP, related with the correction of infeasible linear systems 

and eigenvalue complementarity problems are also discussed. 

4 - Comparison of global and local meta modelling approximations 

for global optimization 

Sergei Kucherenko, Centre for Process Systems Engineering, 

Imperial College London, SW7 2AZ, London, United Kingdom, 

s.kucherenko@ic.ac.uk, Balazs Feil, Nilay Shah 

We consider global and local meta modelling approximations for global optimization. 

The global meta modelling approach is based on the Quasi Random 

Sampling - High Dimensional Model Representation method. It is capable of 

dealing with high dimensional, multimodal problems of low or moderate complexity. 

It is also can be used for global sensitivity analysis of the objective 

functions. Global sensitivity analysis is an efficient technique for model analysis 

and complexity reduction and as such can be used for designing efficient 

global optimization strategies. The radial basis function (RBF) method utilizing 

the local modelling strategy shows very good performance on complex test 

problems. 

� TC-44 


8.2.03 

Interregional Security Work 



Chair: Viveca Asproth, Information Technology and Media, Mid 

Sweden University, 83125, Östersund, Sweden, 

viveca.asproth@miun.se 

Chair: Stig C Holmberg, Information Technology and Media, Mid 


Östersund, Sweden, shbg@ieee.org 

1 - C4I2 for Interreg Security 

Erik Borglund, Department of Information Technology and 

Media, Mid Sweden University, 87188, Härnösand, Sweden, 

erik.borglund@miun.se, Lena-Maria Öberg

Collaboration between interregional actors in crisis management is costly and 

difficult to practice on regular basis. We argue that crisis management needs to 

take into account the unknown. The uncertainty of future crisis makes training, 

practice and design of C4I2 systems a real challenge. In this paper we present 

one possible method, which could help us to design C4I2 systems that minimize 

the unanticipated risk. We propose a combination of scenario planning 

and the interaction design technology, personas, as the silver bullet that prepare 

us for the unknown and reduce uncertainty. 

2 - Spatial Information for Interregional Security Work 

Andreas Ring, DSV, ITM, Mittuniversitetet, Akademigatan 1, 

Hus Q, SE-83125, Östersund, Sweden, andreas.ring@miun.se, 

Viveca Asproth 

Disaster management demands rapid and timely coordination, not only between 

members within a team and between different teams. A system for exchange 

of experiences and models of response to emergence situations within 

and between nations may impede the emergence response. In this paper spatial 

information for a computer system for disaster management is discussed. 

Questions to be answered are: Which spatial information is needed? What 

is already present and how accessible is it? How should the information be 

aggregated and presented? 

3 - Soft Early Warning for Regional Security 

Viveca Asproth, Information Technology and Media, Mid 

Sweden University, 83125, Östersund, Sweden, 

viveca.asproth@miun.se, Stig C Holmberg, Christina Nyström 

Awareness and preparedness are identified as being of paramount importance 

for the security level in geographical regions. A regional security systems according 

to those insights are here seen as a soft early warning system (SEWS). 

In such a security SEWS each individual living in a region acts as a networked 

human sensor. Out conception of a regional security SEWS will be given as an 

idealised design. 

4 - Business Simulator as a Tool to Improve Learning Process 

- Experimental Results 

Miroljub Kljajic, Faculty for organizational sciences, University 

of Maribor, Kidriceva cesta 55, 4000, Kranj, Slovenia, 

miroljub.kljajic@fov.uni-mb.si, Mirjana Kljajic-Borstnar, Andrej 

Skraba, Davorin Kofjac 

The methodology in the decision assessment of complex systems using simulation 

model is described. Students took part in the experiment with the task of 

solving a management problem. Experimental results were analyzed and discussed 

in the students’ projects. After the experiment, students had to complete 

an opinion questionnaire. The results of Criteria Function and student’s opinion 

support the hypothesis that simulation model application contributes to greater 

understanding of the problem and greater convergence in the decision-making 

process. 

� TC-45 


8.2.12 

OR in Military I 

Stream: OR in Military 


Chair: Ana Isabel Barros, Information & Operations, TNO - Defense, 

Security and Safety, POBox 96864, 2509 JG , The Hague, 

Netherlands, ana.barros@tno.nl 

1 - Procurement Decision Support for Portuguese MoD: 

The MACBETH Approach and the Acquisition of 8x8 

AWV 

Carlos Bana e Costa, Engineeing & Management, IST, Institute 

of Science and Technology, Technical University of Lisbon, 

Avenida Prof. Dr. Cavaco Silva, 2780-990, Porto Salvo, 

Portugal, carlosbana@ist.utl.pt, Mario Simoes-Marques 

Since 2003 the Portuguese MoD is using MACBETH to support tender evaluation 

processes. The paper describes a joint acquisition program of AWV 

that involved harmonizing requirements; issuing Tender Program and Technical 

Specs; proposal analysis; tests; negotiations; BAFO evaluation and award 

contract to the globally most attractive proposal. The MACBETH methodology, 

previously adopted on large national and international tenders, proved to 

be a flexible, consistent and robust decision support tool 


2 - Equipment procurement systems methodology - the 

military equipment selection 

Irena Peharda, FOI, Croatia, peharda@hotmail.com 

This paper presents an overview of the equipment procurement methodology, 

with the goal of aligning the equipment characteristics and organizational goals. 

The possibility of combining value-focused thinking, risk evaluation, and negotiation 

procedure based on analytical modeling, is considered, to improve 

aligningment. The described alignment methodology is tested on a complex 

decision-making problem: the military equipment selection for the Croatian 

Armed Forces, the 8x8 Armored Wheeled Vehicles. 

3 - Optimization of U.S. Army Equipment Readiness 

Javier Salmeron, Operations Research, Naval Postgraduate 

School, 1411 Cunninham Rd, 93943, Monterey, CA, United 

States, jsalmero@nps.edu, David Alderson, Robert Dell, Lee 

Ewing 

The U.S. Army has sought analysis and guidance in support of their efforts to 

achieve the readiness goals set by Army Regulation 220-1, which is intended to 

prepare each Army unit to meet its Modified Table of Organization and Equipment 

requirements. We develop Optimal Readiness Allocation Model, a mixedinteger 

mathematical optimization model that maximizes readiness across the 

force. Computational results for a problem with 1,900 units and 2,500 items 

demonstrate readiness can be significantly improved, even when restrictions on 

unit-to-unit redistributions are present. 

4 - A cost effective spare parts package for military equipment 

deployment 

Ana Isabel Barros, Information & Operations, TNO - Defense, 


Netherlands, ana.barros@tno.nl, Nicole van Elst, Karin de 

Smidt-Destombes, Jan Hontelez, Harm Mulder 

Many of today’s technological systems, such as military equipment are characterised 

by a high level of complexity and sophistication. The military context 

requires a guarantee of high availability, since the consequences of equipment 

downtime can have very serious repercussions and may lead to mission failure. 

In order to achieve a high level of availability, the availability of spare 

parts is essential. Therefore, during the initial period of military deployment, 

when the re-supply lines cannot be guaranteed, a deployment spares package is 

required. We present a method to determine a cost effective spare parts deployment 

package when the total amount of available systems (aircraft) is larger 

than the required number of systems and where only the required number of 

systems is used (cold stand by redundancy on system level). 

� TC-46 


8.2.14 

Agent-based Modeling I 

Stream: Agent-Based Modeling 





Chair: Markus Günther, Department of Business Administration, 


markus.guenther@univie.ac.at 

1 - Spatial social networks in an agent-based model of new 

product diffusion 

Elmar Kiesling, Department of Business Administration, 

University of Vienna, Brünnerstraße 72, 1210, Vienna, Austria, 

elmar.kiesling@univie.ac.at, Markus Günther, Christian 

Stummer, Rudolf Vetschera, Lea M. Wakolbinger 

In recent years, agent-based modeling has become a popular tool for investigating 

diffusion processes and forecasting innovation adoption. In such models, 

information exchange between consumer agents plays a crucial role because it 

forms "information cascades" that considerably impact the emergent adoption 

patterns. In this talk, we present an approach for generating geographically 

dispersed social networks using real-world population density data and discuss 

experiences from an application in an agent-based simulation of the market 

introduction of a new product. 

169


2 - A Class of Random Asymmetric Multi-Agent Models for 

Innovation Diffusion 

Christos Emmanouilides, Department of Economics, Aristotle 

University of Thessaloniki, 541 24, Thessaloniki, Greece, 

cemman@econ.auth.gr 

The paper presents a class of statistical models for the diffusion of innovations 

in a social space consisting of individual agents that interact randomly with 

each other across spatial hierarchies over time. The dynamics of the models is 

studied through numerical simulations. Models have multiple equilibriums and 

in some case more complex attractors. Under suitable restrictive assumptions 

reduced models are derived. These models can be estimated using longitudinal 

data techniques that result to consistent and efficient estimates of the effects of 

complex multi-agent interactions on diffusion. 

3 - Organizational Innovativeness and Diffusion of Innovation 

Xiaohui Shi, York Management School, University of York, 

Heslington Road, YO10 4DR, York, United Kingdom, 

xs518@york.ac.uk, Kiran Fernandes 

This paper seeks to understand and simulate the diffusion process from an economic 

and agent based view.We consider innovativeness as a dynamic characteristic 

of a firm and therefore use geographical spatial data in our model.We 

present innovation as a dynamic ’element’ that can be adopted by a firm based 

on their level of innovativeness and use real world data to simulate states that 

change over a period of time.Output of this work should be of value to both 

academics and practitioners who are keen to understand the dynamic nature of 

organizational innovativeness and the diffusion process 

4 - Intra-firm knowledge diffusion and the emergence of innovations: 

An agent-based simulation approach 

Markus Günther, Department of Business Administration, 


markus.guenther@univie.ac.at, Christian Stummer 

In this talk, we introduce an agent-based simulation approach that addresses 

intra-firm knowledge diffusion and its impact on the firm’s innovativeness. The 

model extends prior research on the impact of network structure on knowledge 

diffusion in several aspects. First and foremost, it supplements (purely) 

knowledge-based measures by innovation-oriented ones. Furthermore, the simulation 

allows for various knowledge enhancing activities as well as information 

decay over time. 

� TC-48 


8.2.04 

Nonlinear Optimization and Applications 1 



Chair: Isabel Espírito Santo, Production and Systems, Minho 

University, Campus de Gualtar, 4710-057, Braga, Portugal, 

iapinho@dps.uminho.pt 

1 - Dengue in Cape Verde: modelling and optimal control 

Helena Sofia Rodrigues, Escola Superior de Ciências 

Empresariais, Instituto Politécnico de Viana do Castelo, Av. 

Miguel Dantas, 4930-678 , Valença, Portugal, 

sofiarodrigues@esce.ipvc.pt, M. Teresa Torres Monteiro, Delfim 

F. M. Torres 

The dengue is a vector-borne disease common in tropical areas. In 2009, for 

the first time, an outbreak of dengue was reported in Cape Verde and affected 

more than 20000 persons. This paper proposes a model for dengue disease, 

that includes human and mosquitoes components. The aim is to analyze the 

relationship between several controls and consequently the number of affected 

persons. The data used is from the episode that happened in Cape Verde (October 

to December 2009). 

2 - Offline Biplot Analysis of Nondominated Sets in Multiobjective 

Engineering Optimization 

170 

Lino Costa, Dept. Production and Systems, University of Minho, 

School of Engineering, Campus de Gualtar, 4710-057, Braga, 

Portugal, lac@dps.uminho.pt, Pedro Oliveira 

Real-world problems often involve a large number of objectives. Several multiobjective 

evolutionary algorithms have emerged to tackle problems with increasing 

number of objectives. However, the representation and visualization 

of the nondominated sets is not simple since a large amount of information is 

involved. Thus, there are enormous difficulties concerning the decision making 

process. A methodology based on Biplot graphical representations is proposed 

to retrieve information from nondominated sets obtained for several engineering 


3 - Comparison of Classic and Multi-objective Genetic Algorithms 

for Optimal Design and Control of Chemical 

Processes 

Silvana Revollar, Universidad Simón Bolívar, 89000, Sartenejas, 

Venezuela, srevolla@usb.ve, Rosalba Lamanna, Pastora Vega, 

Mario Francisco 

This work deals with the application of a methodology based on multiobjective 

genetic algorithms for the optimal design and control of chemical processes. 

The mathematical formulation results into a mixed-integer non linear dynamic 

optimization problem. In the proposed multi-objective GA the economic costs 

and dynamic performance indices are considered as separated optimization objectives. 

The activated sludge process is used to test the GA performance, 

comparing the multi-objective formulation to the classical one where only economic 

costs are optimized. 

4 - Collision Avoidance for the ATM problem: A mixed 0-1 

nonlinear approach 

Francisco Javier Martin-Campo, Statistics and Operations 

Research, University Rey Juan Carlos, C/ Tulipán s/n 

Departamental II Building, Office 045., 28933, Móstoles, 

Madrid, Spain, javier.martin.campo@urjc.es, Laureano Fernando 

Escudero, Antonio Alonso-Ayuso 

The main objective of this work is tackling the Collision Avoidance Problem 

applied to the air traffic. We propose a mixed 0-1 nonlinear model that will be 

solved by using Taylor approximations for the nonlinear constraints, integrated 

into a simple algorithm based in a few iterations. This model is able to provide 

the best strategy for an arbitrary aircraft configuration such that the conflicts in 

the airspace are avoided, minimizing the acceleration changes, and forcing to 

return to the initial flight plan when no aircrafts are in conflict.

Tuesday, 14:00-15:20 

� TD-01 


Aula Magna 




Chair: David Pisinger, DTU Management, Produktionstorvet 424, 

2800, Kgs. Lyngby, Denmark, pisinger@diku.dk 

1 - Can the computer make scientific discoveries? 

Pierre Hansen, MQG, GERAD and HEC Montreal, 3000 Chemin 

de la Cote-Sainte-Catherine, H3T 2A7, Montreal, Quebec, 

Pierre.Hansen@gerad.ca 

Is there a systematic method leading to scientific discovery? To this question, 

Francis Bacon answers yes, while both Albert Einstein and Karl Popper answer 

no. As in so many cases, enormous progress in computer power, and in its 

efficient use, renew the question. Taking examples from geometry, graph and 

number theory, physics and chemistry, we will illustrate some clear successes 

(and some failures). This will lead to some speculation on the balance between 

automation and inspiration in discovery. 

� TD-02 


3.2.14 

Topics in Combinatorial Optimization 

Stream: Combinatorial Optimization [c] 


Chair: Marek Libura, Systems Research Institute, Polish Academy of 

Sciences, Newelska 6, 01-447, Warszawa, Poland, 

libura@ibspan.waw.pl 

1 - Robustness analysis in combinatorial optimization. 

Marek Libura, Systems Research Institute, Polish Academy of 

Sciences, Newelska 6, 01-447, Warszawa, Poland, 

libura@ibspan.waw.pl 

We consider the generic combinatorial optimization problem in which the set 

of feasible solutions is fixed, but the weights of the ground set elements may 

vary. For such a problem we study subsets of weights for which an initially 

optimal solution remains robust. Our approach is therefore a natural extension 

of the standard stability analysis. We present results concerning the robustness 

region, robustness radius and robustness tolerances, which are defined as direct 

analogues of the stability region, stability radius and stability tolerances 

considered in the sensitivity analysis framework. 

2 - Cooperative Cuts: Graph Cuts with Submodular Edge 

Weights 

Stefanie Jegelka, Empirical inference, Max Planck Institute for 

Biological Cybernetics, Spemannstr. 38, 72076, Tuebingen, 

Germany, jegelka@tuebingen.mpg.de, Jeff Bilmes 

We introduce cooperative cut, a minimum cut problem whose cost is a submodular 

function on sets of edges: the cost of an edge that is added to a cut set 

depends on the edges in the set. Applications are e.g. in probabilistic graphical 

models and image processing. We prove NP hardness and a polynomial lower 

bound on the approximation factor, and upper bounds via four approximation 

algorithms based on different techniques. Our additional heuristics have attractive 

practical properties, e.g., to rely only on standard min-cut. Both our 

algorithms and heuristics appear to do well in practice. 

3 - New formulations for the k-club problem — a comparative 

study 

Maria Almeida, Instituto Superior de Economia e 

Gestão-UTL/Centro de Investigação Operacional, Rua do 

Quelhas, 6, 1200-781, Lisboa, talmeida@iseg.utl.pt, Filipa 

Carvalho 

EURO 24 - Lisbon 2010 TD-03 

A k-club is a clique relaxation which represents a dense structure in a graph if 

k is small (a 1-club is simply a clique). Finding a maximum cardinality k-club 

is NP-hard for any k. In this talk two types of formulations for the k-club problem 

are considered: formulations with node variables only and formulations 

that combine them with edge variables. We show how to strengthen these formulations 

with valid inequalities and how to embed the strengthened models 

in exact and heuristic algorithms to solve the problem. Comparative computational 

results are presented. 

4 - Solving the Quadratic Assignment Problem by Means 

of General Purpose Mixed Integer Linear Programming 

Solvers 

Huizhen Zhang, Statistics and Operations Research, Rey Juan 

Carlos University, C/ Tulipán s/n, 28933 , Móstoles, Madrid, 

Spain, zhzzywz@gmail.com, Cesar Beltran-Royo 

The Quadratic Assignment Problem (QAP) is one of the most difficult combinatorial 

optimization problems with a diversity of applications. Linearization 

is a well-known solution method for the QAP where one formulates the QAP 

as a (mixed) integer linear programming ((M)ILP) problem. Kauffmann and 

Broeckx’s linearization (1978) is the smallest QAP-MILP formulation but it is 

also one of the weakest ones. In this work we analyze how Kauffmann and 

Broeckx’s formulation can be tightened and used in the framework of the semi- 

Lagrangian relaxation in order to solve the QAP by means of general purpose 

mixed integer linear programming solvers. 

� TD-03 


3.2.15 

Transportation and logistics 



Chair: Christophe Duhamel, LIMOS, Université Blaise Pascal, 

campus des Cézeaux, 63173, Aubière, France, 

christophe.duhamel@isima.fr 

Chair: Jalel Euchi, Quantitatives Methods, Faculty of Economics and 

Management of Sfax, Route de l’aéroport km 4.5, 3018, Sfax, Sfax, 

Tunisia, jalel.euchi@fsegs.rnu.tn 

1 - A Hybrid Tabu Search to Solve the Heterogeneous 

Fixed Fleet Vehicle Routing Problem 

Jalel Euchi, Quantitatives Methods, Faculty of Economics and 

Management of Sfax, Route de l’aéroport km 4.5, 3018, Sfax, 

Sfax, Tunisia, jalel.euchi@fsegs.rnu.tn, Habib Chabchoub 

The Heterogeneous Fixed Fleet Vehicle Routing Problem (HFFVRP) is a variant 

of the Vehicle Routing Problem (VRP) that aims to provide service to a 

specific customer group with minimum cost using a limited number of vehicles. 

We assume that the number of vehicles is fixed. We must decide how to 

make the best use of the fixed fleet of vehicles. In this paper we describe a Tabu 

Search algorithm embedded in the Adaptive Memory (TSAM) procedure to 

solve the HFFVRP. Computational experiments indicating the performance of 

the algorithm concerning quality of solution and processing time are reported. 

2 - Scheduling of road construction projects by means of 

tabu search algorithm 

Jacek Hejducki, Electronics, University of Technology Wrocław 

Poland, Wybrze˙ze Wyspiańskiego 27, 50-370, Wrocław, 

Dolnoslaskie, Poland, 163646@student.pwr.wroc.pl, Zdzisław 

Hejducki 

This paper deals with some problems of synchronizing construction activities 

differing in their execution times. A matrix methodology of calculating the 

times of execution of the activities, ensuring that there will be no collisions 

between them, is presented. The methodology is illustrated with numerical examples 

showing the successive steps of the algorithm and is applied to road 

works modelled as the flow shop problem. Negative transport times are used 

to model the specific constraints of the road construction problem. The tabu 

search algorithm is adapted to solve the problem. 

3 - Parallel Cooperative Grasp for the HVRP 

Christophe Duhamel, LIMOS, Université Blaise Pascal, campus 

des Cézeaux, 63173, Aubière, France, 

171

TD-04 EURO 24 - Lisbon 2010 

christophe.duhamel@isima.fr, Philippe Lacomme, Caroline 

Prodhon, Christian Prins 

Parallel implementations of metaheuristics appear naturally as effective alternative 

to speed up search in approximate iterative framework including but not 

limited to tabu, memetic algorithm and grasp. This article focus on definition 

of a cooperative parallel grasp for the Heterogeneous Vehicle Routing Problem 

benchmarked on 94 instances with 80 to 300 nodes to service. The numerical 

experiments prove the method is strongly efficient and outperforms the sequential 

grasp in both term of solutions quality and computation time. 

� TD-04 


3.2.13 

Agriculture, forestry and environmental 

problems 



Chair: Bernard De Baets, Applied Mathematics, Biometrics and 

Process Control, Ghent University, Coupure links 653, Gent, 

Belgium, Bernard.DeBaets@rug.ac.be 

Chair: Quintin Martin, Statistic, University of Salamanca, Plaza los 

Caidos s/n, 37008, Salamanca, Spain, Spain, qmm@usal.es 

1 - Local search based multiobjective metaheuristic for forest 

planning. 

Monica Hernandez, Applied Economics (Mathematics), 

University of Malaga, Campus El Ejido s/n, 29071, Malaga, 

Spain, m_huelin@uma.es, Trinidad Gomez, Julian Molina, 

Rafael Caballero, Maria Amparo Leon 

In this work, we develop a non linear Multiobjective Combinatorial programming 

model for forest planning, which resolution is strongly conditioned by its 

complexity. Thus, we have implemented a Multiobjective Local Search procedure 

based on Scatter Search, and enhanced by a seeding procedure. The 

characteristics and effectiveness of this method are compared to other heuristic 

widely used in the literature, NSGA-II, through computational experiments on 

simulated plantations. 

2 - A genetic algorithm for the optimization of land subdivision 

in the process of agrarian reform 

Mayron César Oliveira Moreira, Applied Mathematics and 

Statistics, University of Sao Paulo, USP - Sao Carlos - ICMC - 

SME - Av. Trabalhador sao-carlense, 400 Cx. Postal 668 - São 

Carlos-SP - CEP 13560-970, Rua Cap. Alberto Mendes Jr., 346, 

apto. 22, Vila Costa do Sol - CEP 13566-010, São Carlos, São 

Paulo, Brazil, mayron@icmc.usp.br, José Ambrósio Ferreira 

Neto, Urbano Fra Paleo, José Norberto Muniz 

Land subdivision in the projects of agrarian reform presents a major challenge 

to planners and decision makers responsible for project implementation in developing 

countries such as Brazil. This is due to the complexity of delineating 

new compact land parcels with a similar productive capacity in a varied environment 

and large geographical area. Usually, land varies highly in soil classes, 

yielding a very different agricultural aptitude. In order to increase the efficiency 

of the procedure and facilitate the land subdivision process, an application using 

Genetic Algorithm (GA) was developed. 

3 - Subset Selection from Multi-Experiment Data Sets with 

Application to Milk Fatty Acid profiles 

172 

Karolien Scheerlinck, Department of Applied Mathematics, 

Biometrics and Process Control, UGhent, Coupure links 653, 

9000, Ghent, Flanders, Belgium, karolien.scheerlinck@ugent.be, 

Bernard De Baets, Ivan Stefanov, Veerle Fievez 

The development of routine analyses to allow for the handling of large amounts 

of samples and to avoid cost and time expensive analytical techniques is of high 

value. These routine analyses most often require calibration using the detailed 

analyses as reference values. A representative subset reflecting the complete 

range of the variables of interest is required for this purpose. In this paper this 

subset selection problem is tackled for multi-experiment data sets. Conventional 

techniques such as the Kennard and Stone algorithm and OptiSim are 

compared to a new approach based on Genetic Algorithms. The challenge here 

is to find an adequate objective function and to modify the standard crossover 

and mutation operators to keep the number of desired samples fixed. These 

techniques are applied on a data set containing the concentration of 45 fatty 

acids, determined by a simplified reference method, in 1033 milk samples, 

stemming from six different experiments. The objective is to select a subset of 

100 samples in which each of the six different experiments is sufficiently represented. 

While there is no obvious way to generalize the conventional methods 

for multi-experiment data sets, this can quite easily be accomplished for Genetic 

Algorithms by modifying the objective function. Our results indicate that 

Genetic Algorithms are very capable of handling the subset selection problem 

for multi-experiment data sets. 

4 - Heuristic Methods for Modelling the Behaviour of Climate 

Variables Using a Multilayer Perceptron 

Quintin Martin, Statistic, University of Salamanca, Plaza los 

Caidos s/n, 37008, Salamanca, Spain, Spain, qmm@usal.es 

An empirical methodology to fit a model for forecasting climate variables is 

developed. The method is based on an artificial neural network (ANN) of 

the multilayer perceptron type (MLP). To configure the model, the series of 

mean monthly minimum temperature (TminMean) data observed at a weather 

station in Avila (Central Spanish Plateau), belonging to the synoptic and climatological 

network of the Spanish National Institute of Meteorology (NIM), 

were used. Experiments were undertaken to determine the number of training 

and test patterns, the number of data per pattern, layer activation functions, 

the training algorithm, and the end-of-training condition that would allow later 

application of the model. We also experimentally determined the type of data 

pre-processing for which the model provided the best yield. The following 

were considered: differentiation, deseasonalisation, anomalies, normalisation 

and standardisation. The resulting model was applied to the TminMean series 

at the stations of the synoptic and climatological network of the NIM on the 

Spanish Central Plateau. A high degree of fitting was observed between the 

real and simulated series, as shown by the values of the determination coefficients 

(R2), and the mean square error (MSE) and the dispersion and sequence 

plots of the real and simulated series. 

� TD-05 


3.2.16 

Parameter tuning and interactive 




Chair: Martin Josef Geiger, Logistics Management Department, 

Helmut-Schmidt-University, Holstenhofweg 85, 22041, Hamburg, 

Germany, m.j.geiger@hsu-hh.de 

Chair: Jana Ries, Department of Mathematics, University of 

Portsmouth, United Kingdom, jana.ries@port.ac.uk 

1 - On the importance of the initial setting when tuning online 

the parameters of metaheuristics 

Paola Pellegrini, Applied Mathematics, Università Ca’ Foscari, 

Dorsoduro, Venice, Italy, paolap@unive.it, Thomas Stützle, 

Mauro Birattari 

Metaheuristics have a number of parameters that need to be instantiated. Several 

on-line tuning approaches have been proposed for adapting these parameters 

during a run, and hence for using the most suitable setting for each problem 

instance. In this study, we analyze the impact of the initial value of the parameters 

on the performance of two adaptation approaches: we apply two variants of 

an adaptive ant colony optimization algorithm to the traveling salesman problem 

and we study the effects of different initializations. 

2 - Fuzzy Instance-specific Parameter Tuning of Instance 

Batches 

Jana Ries, Department of Mathematics, University of 

Portsmouth, United Kingdom, jana.ries@port.ac.uk, Patrick 

Beullens

Parameter tuning and parameter control are established strategies for solving 

the Parameter Setting Problem for meta-heuristics. Instance-specific parameter 

tuning (IPTS) is designed such that a set of parameter values is found according 

to instance-characteristics a priori to the meta-heuristic. An approach is 

presented that models decision maker preference in an IPTS using fuzzy logic 

by comparing the size and the occurrence of specific tour length dimensions for 

a batch of instances. The results indicate that this batch model is a promising 

approach leading to a parameter-free heuristic. 

3 - Interactive Heuristic Search and Decision Making for 

Multi-Criteria Timetabling Problems 

Martin Josef Geiger, Logistics Management Department, 

Helmut-Schmidt-University, Holstenhofweg 85, 22041, 

Hamburg, Germany, m.j.geiger@hsu-hh.de 

The talk presents a interactive search and decision making approach for a 

curriculum-based timetabling problem, based local search metaheuristics. Two 

different aggregation techniques are used and studied. First, a weighted sum 

aggregation, and second, a reference point based approach. Experimental investigations 

are carried out for benchmark instances taken from track 3 of the 

International Timetabling Competition ITC 2007. After ranking among the best 

five approaches world-wide in the ITC 2007, we now extend our work towards 

interactive search and decision making. 

4 - Tabu Search parameters selection for Quadratic Assignment 

Problem based on General Factorial Design 

Mahdi Bashiri, Shahed University, Iran, Islamic Republic Of, 

bashiri.m@gmail.com, Hossein Karimi 

Tabu Search has been used in the literature for solving the Quadratic Assignment 

Problem and it has some parameters which must be determined to have a 

better algorithm performance. In this paper General Factorial Design has been 

used to select the parameters for QAP.We analyzed the proposed TS algorithm 

with two parameters of maximum iteration and short term memory size and 

the results show that the first parameter has no effect on objective value for the 

selected problem data.Finally the proposed method has been illustrated and the 

parameters for the example of QAP has been shown and analyzed 

� TD-06 


8.2.30 

DEA Application III - Transportation 



Chair: Sebastián Lozano, Dept. of Industrial Management, University 

of Seville, Escuela Superior de Ingenieros, Camino de los 

Descubrimientos, s/n, 41092, Seville, Spain, slozano@us.es 

1 - Operational Environment and its Influence on International 

Airports Performance 

Rui Marques, Portugal, rcmar@civil.ist.utl.pt, Pedro Simões, 

Pedro Carvalho 

This paper computes the efficiency of the 141 international major airports using 

robust benchmarking non-parametric techniques. It employs data envelopment 

analysis (DEA) and the recent methods of bootstrap and order-m to provide robustness 

to the scores obtained. Moreover, some explanatory factors included, 

such as GDP, regulation, percentage of non-aeronautical revenues, etc., in order 

to evaluate their influence on airports’ performance. The results of the research 

are here discussed and some policy suggestions and recommendations are provided. 

2 - Efficiency comparison between 14 regions of RAI (Iranian 

Railway) using DEA 

Mohammad Mehdi Movahedi, Management Department, Islamic 

Azad University, Firoozkouh Branch„ Islamic Azad University, 

Firoozkouh Branch, Management Department„ None, 0098, 

Firoozkouh, Tehran, Iran, Islamic Republic Of, 

mmmovahedi@gmail.com, Seyed Mohieddin Hoseini 


Safe, fast, efficient and effective railway is a main factor of economic development 

level of every country. This paper uses the DEA method to evaluate 

and compare the efficiency of 14 Regions of RAI. In addition, we introduce the 

reference unit(s) for every inefficient region, and determine the amount of input 

decrease and/or output increase needed for inefficient units to become efficient. 

Investigations on returns to scale indicate that 8 regions are operating under 

IRS and only one region under DRS. Our suggestion is that RAI further invests 

into the regions operating under IRS. We also submit suitable suggestion on 

how to improve the efficiency of the inefficient regions. 

3 - Technical and scale efficiency of interstate bus companies 

in Brazil 

Carlos Ernani Fries, Department of Production and Systems 

Engineering, Federal University of Santa Catarina, Campus 

Trindade - CTC/EPS, C.P. 476, 88010-970, Florianópolis, Santa 

Catarina, Brazil, ernani@deps.ufsc.br, Antonio G.N. Novaes, 

Mônica M. M. Luna, Helena C. Medeiros 

Interstate bus transport in Brazil is regulated by the National Agency of Land 

Transport (ANTT). The system is formed by more than 200 bus companies. 

Since the production in passenger-kilometers is heavily concentrated, one important 

question is to investigate the existence of returns to scale in conjunction 

with technical efficiency and company size. The DEA approach was used in 

the analysis. The results show three groups of bus firms, ordered by size, with 

clear differences in efficiency. Additionally, 73% of the total production is represented 

by firms showing decreasing returns to scale. 

4 - A Network DEA model of airlines 

Sebastián Lozano, Dept. of Industrial Management, University 

of Seville, Escuela Superior de Ingenieros, Camino de los 

Descubrimientos, s/n, 41092, Seville, Spain, slozano@us.es, 

Ester Gutiérrez, Placido Moreno, Jose L. Salmeron 

Conventional Data Envelopment Analysis (DEA) models consider the system 

as a single-process black box. Network DEA considers the system as composed 

by distinct processes or stages, each one with its own inputs and outputs 

and with intermediate flows among the stages. In this research, a network DEA 

approach to airlines efficiency assessment is presented. The network DEA approach 

has more discriminant power than single-process DEA and the computed 

targets, efficiency scores and rankings are more valid. However, more 

detailed data (i.e. at the process level) and more complex models are needed. 

� TD-07 


8.2.47 

Scheduling problems in production and 

service 



Chair: Dirk Briskorn, Department for Supply Chain Management and 



briskorn@wiso.uni-koeln.de 

1 - Scheduling painting shops with due date restrictions 

Stefan Bock, WINFOR (Business Computing and Operations 

Research) Schumpeter School of Business and Economics, 

University of Wuppertal, Gaußstraße 20, D-42097 Wuppertal, 

42097, Wuppertal, NRW, Germany, sbock@winfor.de, Kathrin 

Klamroth 

In this paper a painting shop with sequence-dependent setup costs is considered. 

Analogously to the approach of Gilmore and Gomory, setup costs are 

assumed to be proportional to absolute state differences. Since subsequent production 

stages need to be supplied according to a predetermined time table, 

due date restrictions incur. We show that the integration of due dates leads to 

a strongly polynomial problem if the number of setup states is a constant. By 

making use of new dominance rules, problem instances of larger size can be 

solved optimally by a Branch&Bound procedure. 

2 - Detecting, Measuring and Repairing Instabilities in 

Schedules of Tasks with Skill Requirements 

Murat Firat, Mathematics and Computer Science, Eindhoven 

University of Technology, Postbus 513, 5600 MB, Eindhoven, 

Netherlands, m.firat@tue.nl, Cor Hurkens, Alexandre Laugier 

173


In this work we extend the notion of blocking pair in the classical model of 

Gale Shapley and define it in a framework with multiple many-to-one assignments 

in the context of multiple skills. We aim to find good quality schedules 

that are as stable as possible. Throughout our solution methodology we use IP 

models to measure instability and to assign the technicians to jobs. 

3 - SAT-based local search for project scheduling 

Andrei Horbach, Institute of Business Administration, University 

Kiel, Olshausenstrasse 40, D-24098, Kiel, Germany, 

horbach@bwl.uni-kiel.de 

Encouraged by recent results of applying SAT to various optimization problems 

such as project scheduling (RCPSP) and sports league scheduling we consider 

further possibilities to employ the techniques of propagation and clause learning. 

Contrary to our previously reported results we make use of SAT techniques 

to do some important work in a heuristic search framework. We present our 

preliminary results for project scheduling with minimum total completion time 

objective. 

4 - Scheduling of aircraft landings using aircraft classes 

Raik Stolletz, Department of Management Engineering, 

Technical University of Denmark, Nils Koppels Alle, Building 

426, 2800, Kgs. Lyngby, Denmark, raist@man.dtu.dk, Dirk 

Briskorn 

This presentation focuses on the aircraft landing problem. We consider a set 

of aircraft classes such that two aircraft of the same class differ by their target 

landing time. We develop polynomial time algorithms to minimize total landing 

cost, where landing cost is specified by a piece-wise linear convex function 

of the landing time. Moreover, we present integer programming models and 

show how the developed optimality properties can be used to increase efficiency 

of standard solvers. 

� TD-08 


6.1.36 

Scheduling Applications 



Chair: Natalia Shakhlevich, School of Computing, University of 

Leeds, Woodhouse Lane, LS2 9JT, Leeds, United Kingdom, 

N.Shakhlevich@leeds.ac.uk 

1 - An effective algorithm to solve Energygrass Supply 

Scheduling problem 

Laszlo Torjai, BDE Research Nonprofit Public Benefit Ltd., 

Xavér str. 1., 7624, Pécs, Hungary, torjai@ktk.pte.hu 

Energygrass Supply Scheduling (ESS) is a bi-objective scheduling model 

emerged from a real-life logistics problem. There are non-preemtive jobs with 

integer processing times and without precedence conditions. The length of the 

discrete-time horizon is equal to the number of jobs and every job has to be 

finished in different time period. Each job is processed on one of the identical 

machines. The first objective is to find a feasible schedule such that the number 

of machines is minimal. The second one, subject to minimal number of 

machines, is minimizing the idle times of them. 

2 - Resource Allocation in Preclinical Pharmaceutical Research 

John Gittins, Statistics, Oxford University, 1 South Parks Road, 

Oxford, OX1 3TG, Oxford, United Kingdom, 

gittins@stats.ox.ac.uk, Shuo Qu, Anne-marie Oreskovich 

Models and software for the allocation of resources in preclinical pharmaceutical 

research will be described. The models are based on stochastic optimisation. 

The relationship between resource allocation and rate of progress, and the 

influence of competition, are modelled. Examples will be discussed, showing 

that large improvements in profitability are often possible as a result of applying 

more rational resource allocation policies 

3 - ZigBee based sensor network design for indoor space 

gazing system 

174 

Jun Matsuo, Hokkaido University, Japan, 

matsuo@complex.eng.hokudai.ac.jp, Hidenori Kawamura, Keiji 

Suzuki, Takeshi Ikeda, Akio Sashima, Koichi Kurumatani 

We realize space gazing system for achievement of safety and convenience that 

watches living indoor space such as home, shopping center. The service images 

include smooth navigation and avoidance of danger situation. At first, we 

adopt to build wireless sensor networks with ZigBee. In this network design, 

the problems are how the ZigBee which is a low rate and low power network 

technology processes much information and detects the indoor environment. 

These problems are depended not only the density of the information but also 

the indoor space properties including the group of human behaviors. 

4 - Scheduling Patterns of Repetitive Tasks: A Case Study 

in Health Care 

Natalia Shakhlevich, School of Computing, University of Leeds, 

Woodhouse Lane, LS2 9JT, Leeds, United Kingdom, 

N.Shakhlevich@leeds.ac.uk, Alessandro Condotta 

Integrated pathway management has become an important concept in achieving 

waiting time targets and improving health care procedures. Our study is applicable 

to those health-care procedures which consist of a series of treatments 

delivered in accordance with strictly regulated standard protocols. Such protocols 

are typical, e.g., for chemotherapy treatments which are characterized 

by strictly pre-specified multi-day and intra-day patterns. We present the main 

features of the associated models, introduce our solution approach and discuss 

preliminary results obtained for real-world data. 

� TD-09 


6.2.53 

Optimization of Transport Problems on 

Networks II 



Chair: Simone Göttlich, Mathematics, TU Kaiserslautern, Post Box 



1 - Solving Equilibrium Problems with Equilibrium Constraints 

via Relaxation Approaches 

Sonja Veelken, Mathematics, RWTH Aachen University, 

Mathematik C, Templergraben 55, 52062, Aachen, Germany, 

veelken@mathc.rwth-aachen.de 

Equilibrium Problems with Equilibrium Constraints (EPECs) arise for example 

as bilevel game-theoretic models used to describe electricity markets with locational 

prices. They form an extension of Mathematical Programs with Equilibrium 

Constraints (MPECs) and so far, little is known about what will be 

the numerical method of choice for this relatively new class of nonlinear programming. 

We study the generalization of a relaxation scheme to EPECs, that 

proved very successful for MPECs. Moreover, we test and compare a related 

numerical solution approach on a randomly generated test set. 

2 - Column Generation Based Heuristic for Capacitated 

Arc Routing Problem with Soft Time Windows 

H. Murat Afsar, Industrial Systems, University of Technology of 

Troyes, 12 rue Marie Curie, BP 2060, 10010, Troyes, France, 

murat.afsar@utt.fr 

Capacitated arc routing problems are inspired by winter gritting and urban 

waste collection where a subset of edges on undirected graphs to be serviced. 

Each edge has a time window where the service cost is minimal. Outside the 

time windows, service is authorized while the cost increases linearly by time. 

We propose an heuristic approach based on Dantzig-Wolfe decomposition and 

column generation. The master problem is set partitioning model and the subproblem 

is a non elementary shortest path with resource constraints. The subproblem 

is solved heuristically and exactly by dynamic programming. 

3 - Paradox in Multi-commodity Min Cost Network Flow 

Problem 

Ankit Khandelwal, Consumer Banking Risk Analytics, Standard 

Chartered Bank, 51 Bras Basah Road #08-04, Plaza by the Park, 

189554, Singapore, Singapore, ankit.khandelwal@gmail.com, 

Sonia 

Minimum cost multi-commodity flow problem is one of the classical optimization 

problems that arise when several commodities share arcs in a capacitated 

network. This type of network has found many practical applications in various 

diversified fields. In this paper we attempt to extend the concept of More (same) 

for less paradox in Multi-commodity Min Cost Flow Problem and obtain the 

conditions for its existence. Some supporting illustrations are also included.

4 - Improving public disabled transportation using a mathematical 

programming approach 

Gizem Ozel, Industrial Engineering, Dokuz Eylül University, 

Tinaztepe Campus, 35160, ˙Izmir, Turkey, 

ozelgizem@gmail.com, Alper Unal, Seyda Topaloglu 

Public transportation of disabled people in Izmir is not at a desired service 

quality since specialized buses for disabled people have been allocated to the 

bus routes without considering disabled population and degree of disability on 

each route. The aim of this study is to improve disabled transportation. First, 

weighted scoring method and fuzzy TOPSIS is used to find the need of disabled 

buses on each route.Then, a mathematical programming model is developed for 

optimal assignment of these buses. This way, the number of disabled passengers 

is greatly increased compared to current situation. 

� TD-10 


6.2.56 

OR in Sports 1 

Stream: OR in Sports 


Chair: Dennis Van den Broeck, MOSI, VUB, P.P. Rubensstraat 9, 

1880, Kapelle-op-den-Bos, Belgium, devdbroe@vub.ac.be 

1 - Scheduling the play-offs in Belgian football 

Dries Goossens, Operations Research and Business Statistics 

(ORSTAT), Katholieke Universiteit Leuven, Naamsestraat 69, 

3000, Leuven, Belgium, Dries.Goossens@econ.kuleuven.be 

Since this season, the competition format in the highest division of Belgian 

football includes play-offs, consisting of 3 double round robin tournaments, 

with 6 or 4 teams. Apart from the usual scheduling constraints in each of these 

tournament, a number of constraints linking the 3 tournaments should be taken 

into account. For instance, two teams playing in different play-off competitions 

share the same stadium and thus should not play at home in the same 

round. This presentation shows how a play-off schedule was developed which 

was actually used to determine the 2009-2010 champion. 

2 - Combining the AHP and VIKOR Methodologies for 

Ranking Basketball Players 

Seyhan Sipahi, Quantitative Methods, Istanbul University School 

of Business, IU Isletme Fakultesi Avcilar Kampusu 34320, 

Avcilar, 34320, Istanbul, sipahi@istanbul.edu.tr, Bilge Donuk 

The purpose of this study is to provide a hybrid multi-criteria model for ranking 

basketball players who perform in the first division of the National Basketball 

league in Turkey. The model combines the Analytic Hierarchy Process 

(AHP) and the VIKOR (Vise Kriterijumska Optimizacija I Kompromisno Resenje) 

methodologies. In the study, players were ranked in accordance with the 

evaluation criteria such as shooting, free throw, assist, block, steal, turnover and 

rebound. The findings showed that the novel methodology can be successfully 

applied by basketball managers to evaluate player performance. 

3 - Was Joe Girardi Foolhardy? An Analysis of the New 

York Yankee Manager’s Decision to use only Three 

Starting Pitchers in the 2009 World Series 

James Cochran, Department of Marketing and Analysis, 

Louisiana Tech University, PO Box 10318, 71272, Ruston, LA, 

United States, jcochran@cab.latech.edu 

The NY Yankees followed baseball convention by using a rotation of 5 starting 

pitchers in the 2009 regular season, allowing each pitcher 4 days of rest 

between starts. This is done because most baseball experts believe a pitcher’s 

performance (and his team’s probability of winning) deteriorates drastically if 

he has less than 4 days of rest between starts. However, the team deviated from 

this strategy in the 2009 World Series, using only its top 3 starting pitchers 

and allowing each only 3 days of rest between starts. We use classic decision 

analysis to assess the wisdom of this strategy. 

4 - An Inference Approach for Assessing Competitive Balance 

in Team Sports 

Dennis Van den Broeck, MOSI, VUB, P.P. Rubensstraat 9, 1880, 

Kapelle-op-den-Bos, Belgium, devdbroe@vub.ac.be 


Measures of competitive balance aim to assess the equality of playing strengths 

among teams in a sports tournament. Examples include deviations on winratios 

and the HH-index. These measures are in fact sample statistics and follow 

a probability distribution. Using concepts of graph theory, this distribution 

is derived assuming that all teams have equal playing strengths. Given this distribution, 

the deduction of critical values is straightforward and inferences on 

team parity are possible. A numerical example discussing a simplified tournament 

is presented. 

� TD-11 


8.2.38 

Various New OR Tools and Technologies I 






Chair: Tatiana Tchemisova, Departmento of Mathematics, University 

of Aveiro, Campus Universitario de Santiago, 3810-193, Aveiro, 

Portugal, tatiana@ua.pt 

1 - To develop an e-commerce website template for undergraduates 

Özge Nalan Alp, Yildiz Technical University, Turkey, 

ozgenalan@gmail.com, Nurgül Demirta¸s, Hayri Baraçlı 

E-commerce is a term meaning buying and selling on internet and its applications 

are increasing rapidly. To develop e-commerce applications, firstly the 

scope of the website must be determined. Data mining is a technique which 

finds important information in data warehouses and data mining helps to make 

predictions. We aim to analyze undergraduates’ purchasing tendency and demands 

which changes upon their demographic information by using data mining 

techniques and to prepare a new e-commerce website design template. 

2 - Measuring Virtual Experience in 3-D VR Interactive Simulator 

Environment: A Structural Equation Modeling 

Approach 

Pao-Long Chang, Department of Business Administration, Feng 

Chia University, 100, Wenhwa Road, Seatwen,Taichung, Taiwan, 

R.O.C, 407, Taichung, Taiwan, paolong_chang@yahoo.com.tw, 

Ming-Hua Chieng 

The study aims to measure the consumer’s virtual experience in the formation 

of flow in a 3-D VR environment via a 3-D interactive motion simulator. Study 

findings revealed that stages in the formation of flow encompass the intrinsic 

characteristic manifestation of the mediated environment; the premise and perception 

prior to entering the flow state; the stage of entering a flow state; and 

lastly the consequences of the formation of flow. 

3 - Exploring the Evolutional Life of White LED Lighting 

Technology with Technology Prediction 

Yuanchau Liour, Logistics Management, Takming University of 

Science and Technology, 11451, Taipei, Taiwan, 

ycliour@takming.edu.tw, Po-Liang Chao, Chie-bein Chen 

This study bases on patents of white LED lighting technology from the US 

open patent database, it imports the logistic and Gompertz technology forecasting 

models to explore the life cycle of white LED lighting technology and 

further investigates the strength and weakness in the global level of white LED 

lighting industry in Taiwan. The result shows that the number of patents doubled 

in 2003-2008. Through the analysis by Logistic and Gompertz models, 

the white LED lighting technology is growing to the mature stage of long-term 

life cycle. 

4 - Models and Techniques of Recourse Allocation in Controlling 

the Semi-structured Social and Economic Objects 

Nikolai Korgin, Lab of Active Systems, Institute of Control 

Sciences of Russian Academy of Sciences, 65„ Profsoyuznaya 

st., 117997, Moscow, Russian Federation, nkorgin@ipu.ru, 

Dmitry Makarenko 

175


In controlling the complex socio-economic systems (territories, branches) one 

of the key problems is the optimal resource allocation. This problem is solved 

under incomplete awareness of decision makers, and the object of control is 

semi-structured in terms of correlation between the impact of separate activities 

and their contribution to the whole object’s development goals. To solve this 

problem the paper suggests the combination of the following techniques: cognitive 

mapping, comprehensive evaluation, network programming, constructing 

of non-manipulable mechanisms for resource allocation. 

� TD-12 


8.2.39 

ANP 04 


Processes 


Chair: Mei-Chen Lo, National United University, Taiwan, 

meichen_lo@nuu.edu.tw 

1 - A model for equipment selection: An integrated approach 

with DEMATEL and ANP 

Metin Dagdeviren, Department of Industrial Engineering, 

Engineering and Architecture Faculty, Celal Bayar Bulvar, Gazi 

Unv. MMF, Endustri Muh., 06570, Ankara, Turkey, 

metindag@gazi.edu.tr, Erdem Aksakal, Ihsan Yüksel 

With the affect of the globalization and competition, selecting proper equipment 

becomes a very important activity due to the fact that improper equipment 

selection can negatively affect the overall performance and productivity of the 

manufacturing systems. Selecting the proper equipment includes many alternatives 

on the behalf of the manufacturing systems and makes the equipment 

selection process as a multicriteria decision making problem. In this study, a 

hybrid model which employs DEMATEL and Analytic Network Process together, 

is proposed for the equipment selection problem. 

2 - Improving the assessment of railway projects through 

new performance indicators 

Stefano Strami, DICA - Dept of Civil and Environmental 

Engineering, University of Trieste, Piazzale Europa 1, 34127, 

Trieste, Italy, stefano.strami@phd.units.it, Giovanni Longo, Elio 

Padoano 

In the railway field, multi-criteria decision-making methods are commonly applied 

in order to assess and compare alternative projects. In this work, some 

useful indicators are proposed to quantify the performances and positive effects 

of different interventions. These can be used together with the well-known parameters 

related to the negative drawbacks (which mainly explain the reasons 

not to intervene), aiming at a thorough description of each considered project. 

The new indicators were applied to a real case study in Italy that is described 

in the paper. 

3 - Decision-making by "Minor ANP’ for Selecting the Best 

Business Partner 

Mei-Chen Lo, National United University, Taiwan, 

meichen_lo@nuu.edu.tw, Ozaki Toshimasa, Gwo-Hshiung 

Tzeng 

The ANP is a network structure with alternatives and criteria for clarifying 

complex decision-making problems. It is important to obtain the consent from 

both criteria and alternatives because the evaluation is relative and reciprocal. 

From our study, the evaluation with only the alternatives consists of the missing 

values or the non-square matrix, these irregular alternatives is defined as 

"Minor ANP’ which describes these methods of making the priority of only 

the alternative’s matrix values by using the ANP. An empirical case of the best 

business partner selection is illustrated. 

176 

� TD-13 


2.2.21 

Emergency facilities location 



Chair: Sibel Salman, Industrial Engineering, Koc University, Sariyer, 

34450, Istanbul, Turkey, ssalman@ku.edu.tr 

1 - Emergency Response Facility Location in Istanbul for 

Effective Distribution of Relief Aid 

Selin Ozdinc, industrial engineering, koc university, koc 

university, rumelifeneri yolu, sariyer, 34450, istanbul, Turkey, 

selinozdinc@gmail.com 

We model the Emergency Response Facility location problem for the Disaster 

Coordination Center of Istanbul Municipality by a two-stage stochastic program 

that maximizes the expected post-disaster demand coverage in a network 

whose links may fail.We develop a Tabu Search algorithm that uses sampling 

to evaluate the expected coverage.We analyze the solutions under various cases 

and provide guidelines for earthquake-preparedness. 

2 - Pre-Positioning of Emergency Items Worldwide for 

CARE International 

Serhan Duran, Industrial Engineering, Middle East Technical 

University, Orta Dogu Teknik Universitesi, Endustri Muh. 

Bolumu, 06531, Ankara, Turkey, sduran@ie.metu.edu.tr, Marco 

Gutierrez, Pinar Keskinocak 

The most vital issues in responding to such disasters are the agility in mobilizing 

the supplies and the effectiveness in distributing them. In an effort 

to improve disaster response, a research group from Georgia Tech, in collaboration 

with CARE, developed a model to evaluate the effect of relief items 

pre-positioning on CARE’s average emergency response time to provide relief 

aid to people affected by natural disasters. The model’s results helped CARE’s 

managers to determine a desired configuration of three facilities around the 

world for the pre-positioning network. 

3 - A humanitarian relief logistics model and Lagrangeanbased 

solution approach 

Alper Döyen, Industrial Engineering, Bo˘gaziçi University, 

Department of Industrial Engineering, Bo˘gaziçi University, 

Bebek, 34342, ˙Istanbul, Turkey, alper.doyen@boun.edu.tr, 

Necati Aras, Gülay Barbarosoglu 

We develop a two-stage stochastic programming model for a humanitarian relief 

logistics problem where decisions are made for pre- and post-disaster rescue 

centers, the amount of relief items to be stocked and the flow amounts at 

each echelon . The objective is to minimize the total cost of facility location, 

inventory holding, transportation and unsatisfied demand. The deterministic 

equivalent of the model is formulated as an MIP for which a Lagrangean relaxation 

based algorithm is proposed. Test results show that proposed algorithm 

shows good performance for a wide range of problem instances. 

4 - Locating Emergency Response Facilities on a Network 

with a Linear Reliability Order of Links 

Sibel Salman, Industrial Engineering, Koc University, Sariyer, 

34450, Istanbul, Turkey, ssalman@ku.edu.tr, Refael Hassin, R. 

Ravi 

We study the problem of locating emergency response facilities to maximize 

the expected demand serviced in a network with unreliable links. Under a linear 

reliability ordering of links,that models dependencies among link failures in 

a disaster situation, we give two polynomial time exact algorithms by a greedy 

approach and by dynamic programming. When two disaster scenarios with different 

linear orderings are given, we prove the total unimodularity of a linear 

programming formulation and NP-hardness for three orderings. When a demand 

point can be covered only if a facility exists within a distance limit, we 

show that the problem is NP-hard even for a single ordering.

� TD-14 


2.2.15 

Supply Chain Design and Sustainability 



Chair: Charles Corbett, UCLA Anderson School of Management, 

110 westwood plaza, box 951481, 90095-1481, Los Angeles, CA, 

United States, charles.corbett@anderson.ucla.edu 

1 - Modeling and Design of Cross-Docking Systems 

Georgios Sgoutas, Business Administration, University of 

Aegean, N. Haviara 14, Evaggelistria, 82100, Chios, Greece, 

george.chios@gmail.com, Michael Vidalis, Vassilis Angelis 

Cross docks are modelled as queuing network with parallel servers. First station 

represents strip doors, 2nd station represents stack doors and the intermediate 

buffer models the width of the dock. Markov processes are used to 

model random arriving, servicing, repairing times and to calculate Performance 

measures for different dock configurations. Cross docking system is assumed 

with and without waiting space and with breakdowns in loading and unloading 

operations. Structure of transition matrices is examined for each case and 

computational algorithms are developed for generating them. 

2 - Supply Chain Network Design with Sustainability Perspective 

Hazal Karaman, Industrial Engineering, Kadir Has University, 

Zeytinoglu st. Lale C, 33/8 Akatlar, 34335, Istanbul, Turkey, 

hazalkaraman@gmail.com, Cagatay Mekiker, Sinan Mercan, 

Orhan Feyzioglu 

While supply chains have become globalized, environmental concerns due to 

global warming and associated security risks have drawn the attention of numerous 

constituencies. Companies are increasingly being held accountable for 

their own environmental performance, and also for that of their suppliers, distribution 

facilities and even for the disposal of their products. We present a 

supply network design model that aims to locate centers and assign material 

flow optimally while considering multiple sustainability related objectives, and 

provide economic insights with an illustrative example. 

3 - Carbon footprinting and labeling in a supply chain 

Charles Corbett, UCLA Anderson School of Management, 110 

westwood plaza, box 951481, 90095-1481, Los Angeles, CA, 

United States, charles.corbett@anderson.ucla.edu, Chien-Ming 

Chen, Rob Zuidwijk 

When firms plan to put carbon footprint labels on their products, it is often not 

unambiguous how those carbon footprints should be determined. Current standards 

for carbon footprinting also leave room for ambiguity. This gives firms 

some flexibility in how to allocate carbon emissions to different products. In 

this paper we examine conditions under which that flexibility in fact helps to 

reduce the firm’s total carbon footprint. 

� TD-15 


2.2.12 

Routing Optimization 



Chair: Christos Tarantilis, Department of Management Science & 

Technology, Athens University of Economics and Business, 28 

Hydras st, Athens, Greece, Management Science Laboratory (MSL), 

9th Floor, Room 913, 113-62, Athens, Greece, tarantil@aueb.gr 

1 - Solving the Weekly Log-Truck Routing and Scheduling 

Problem 




Canada, michel.gendreau@cirrelt.ca, ’Nizar El Hachemi, 

Louis-Martin Rousseau 

We address the problem of routing and scheduling log-trucks over a week in 

a forestry environment. This problem includes features such as multiple products, 

inventories, and lunch breaks for drivers. The objective is to minimize 

the overall cost including waiting times, loaded travel and empty driven distance. 

Our solution approach is based on solving in sequence two integer linear 

programs that determine respectively the destinations of full truckloads and the 

routing and scheduling of trucks at a minimum cost. Results of computational 

experiments using CPLEX 11 will be reported. 

2 - A branch-and-price algorithm for the capacitated team 

orienteering problem 

M. Grazia Speranza, Dept. of Quantitative Methods, University 

of Brescia, C.da Santa Chiara, 50, 25122, Brescia, Italy, 

speranza@eco.unibs.it, Claudia Archetti, Nicola Bianchessi, 

Nicola Bianchessi 

The capacitated team orienteering problem (CTOP) is the problem to select a 

subset of customers among a set of potential customers with the objective to 

maximize the total profit collected. Each customer has an associated demand 

and profit. A fleet of capacitated vehicles is available. In this paper we present 

a column generation approach for the solution of the CTOP that solves to optimality 

many previously unsolved benchmark instances. 

3 - Partial Path Column Generation for the Elementary 

Shortest Path Problem with Resource Constraints 

Mads Kehlet Jepsen, Managment Engineering, Technical 

University of Denmark, Produktionstorvet, bygn. 424, 2800, Kgs 

lyngby, Denmark, makj@man.dtu.dk, Bjørn Petersen 

This paper introduces a decomposition of the Elementary Shortest Path Problem 

with Resource Constraints, where the path is combined by smaller sub 

paths. The approach can be seen as an idea to extent the bidirectional labeling. 

We present a Danzig-Wolfe Decomposition algorithm for the problem and compare 

the dicuess the computational results by comparing different approaches 

for the decomposition. Finally we compare the algoritmh with existing algorithms. 

4 - Towards an Exact Approach for the Capacitated Vehicle 


Panagiotis Repoussis, Management Science & Technology, 

Athens University of Economics & Business, Evelpidon 47A, 

11362, Athens, Greece, prepousi@aueb.gr, Chrysanthos 

Gounaris, Christos Tarantilis, Christodoulos Floudas 

In this presentation we propose advances towards a two-level exact solution 

framework for the Capacitated Vehicle Routing Problem. During the first level, 

a soft branching decomposition scheme based on Hamming distances is iteratively 

applied to confine and partition the solution space. At the second level, 

each solution subspace is solved via a new branch-and-cut algorithm that uses 

an adaptive memory programming metaheuristic to inform branching decisions 

and accelerate the search process. Computational experiments on benchmark 

data sets illustrate the performance of the proposed approach. 

� TD-16 


2.2.14 

Strategic decisions and infrastructure 



Chair: Konstantinos Gkoumas, DITS (Dipartimento di Idraulica 

Trasporti e Strade), Sapienza Universita’ di Roma, Via Eudossiana 

18, 00184 , Rome, Italy, konstantinos.gkoumas@uniroma1.it 

Chair: Federico Perea, Matemática Aplicada 2, Universidad de 

Sevilla, 41092, Seville, Spain, perea@us.es 

1 - Exact algorithms for Location Routing Problems 

Claudio Contardo, Département d’ informatique et de r.o., 

University of Montreal, 2920 chemin de la tour, bureau 3484, 

H3T 1J4, Montreal, Québec, Canada, ccontard@crt.umontreal.ca 

177


Location routing problems usually occur whenever strategic decisions must be 

taken considering operational consequences. We review the state of the art of 

exact algorithms for solving Location Routing Problems. We describe algorithms 

based on cutting planes strategies as well as decomposition methods 

solved by means of Lagrangean relaxation or column generation. We present 

computational results and discuss the strengths of every approach as well as 

directions of future research 

� TD-17 


1.3.14 

Container Terminal Planning II 



Chair: Frank Meisel, Martin-Luther-University Halle-Wittenberg, Gr. 

Steinstr. 73, 06108, Halle, Germany, frank.meisel@wiwi.uni-halle.de 

Chair: Christian Bierwirth, Martin-Luther-University 

Halle-Wittenberg, 06108, Halle, Germany, 

christian.bierwirth@wiwi.uni-halle.de 

1 - Effect of Resource Allocation Rules in Different Layout 

Types of Seaport Container Terminals 

Olcay Polat, Industrial Engineering, Pamukkale University, 

Pamukkkale Üniversitesi Endüstri Mühendisli˘gi Bölümü No:437, 

Kinikli Kampusu, 20100, Denizli, Turkey, olcaay@gmail.com, 

Osman Kulak, Mustafa Egemen Taner, Hans-Otto Guenther 

Due to the long-lasting increase in global trade worldwide until the recent economic 

crisis container traffic has grown dramatically. As a result, new terminals 

have opened and existing terminals face a much higher container turnover than 

before. The goal of this study is to propose strategies for allocating resources 

in different layout alternatives and evaluating impact of these strategies rules in 

these layout types. With the help of the simulation model potential layout and 

allocation alternatives can be evaluated for the container terminals. 

2 - Terminal planning and optimising via distributed emulation 

Holger Schuett, Optimisation and Simulation, ISL - Institute of 

Shipping Economics and Logistics, Barkhausenstrasse 2, 27568, 

Bremerhaven, Germany, schuett@isl.org 

The paper presents an emulation system based on a worldwide accepted simulation 

tool for container terminals. By combining the Terminal Operating System 

to models of the terminal devices (emulators) the TOS may be tested before going 

alive. Moreover emulation will allow the terminal operator to fine-tune the 

strategies and historical data may be imported to train the staff. The emulation 

system is based on a distributed communication network, which allows the 

integration of external emulators. This will be shown in an example, where a 

Java based emulator connects to the C# based network. 

3 - An integrated simulation model for quay and yard operations 

at gioia tauro terminal 

Roberto Trunfio, Dipartimento di Elettronica, Informatica e 

Sistemistica, Università della Calabria, Via P. Bucci, 41C, 87036, 

Rende, Cosenza, Italy, rtrunfio@deis.unical.it, Pasquale Legato 

Discrete-event simulation (DES) models are powerful tools for representing 

maritime container terminals in a stochastic-dynamic context. They are able of 

giving a valuable contribution in estimation of entire terminal performance by 

considering all the logistical processes involved in system dynamics. Hence, 

container terminal companies recognize that simulation cover a lead role in 

supporting the decisional process. For this reason, the MCT company at Gioia 

Tauro terminal, asked us for an integrated DES model focused on quay and yard 

operations. The model depicts in detail both vessel and container life cycles. 

Moreover, quay crane allocation and yard block policies are considered. 

4 - Analysis and Simulation of a Port Container Terminal 

Operating in Naples (Italy) Harbor 

178 

Gennaro Improta, Dipartimento di Ingegneria 

Economico-Gestionale, Università Federico II di Napoli, Via 

Claudio, 21, 80125, Napoli, Italy, improta@unina.it, Francesco 

Gargano, Giuseppe Bruno, Andrea Genovese, Cinzia Vinti, 

Pietro Averaimo 

We implemented, in Arena environment, a Discrete Event Simulation model of 

the Co.Na.Te.Co. Container Terminal operating in Naples, Italy, harbor. In a 

first phase we reproduced its current situation. Obtained results showed a good 

correspondence between real and simulated data. Then, in order to evaluate 

possible performance improvements, we have experimented different decision 

rules for the Berth Allocation phase considering it as a dynamic parallel machines 

scheduling problem. First simulated results seems to be promising. 

� TD-18 


1.3.15 

Decision Analysis in Marketing and 

Financial Modeling 











1 - Applying Rough Set Theory Constructing Customer Relationship 

Model for Pharmaceutical Agents 

Tsau-Tang Yang, Graduate Institute of Management Sciences, 

Tamkang University, 8F, No.43, Lane 115, Sec.2, Chung Shan 

N.Rd. Taipe, Taiwan(R.O.C), 10448, Taipei, Taiwan, 

giddi.john@msa.hinet.net, Hsu-Hua Lee, Chie-Bein Chen, 

Mei-Hsiu Huang, Wei Ting Cho, Chung-Chang Lien 

Business need take targeted marketing to capture the customers in promotion 

activity, since modern marketing paradigm rapidly shifting. Recently, pharmaceutical 

agents face challenges, such as foreign pharmaceutical companies authorizing 

the rights to the agents’ pharmaceutical divisions or hospitals taking 

a total payment system. Therefore, pharmaceutical agents should adjust their 

marketing strategy to keep long-relationship with existed channels. It plays an 

important role for finding unobvious but valuable promotion knowledge to support 

making marketing related decisions in customer relationship management. 

This study use rough set theory to analyze actually the transaction records between 

a specific pharmaceutical agent company and its channels, and then build 

customer relationship model. The result of this study can provide the decision 

maker of pharmaceutical agent company as a reference for making marketing 

strategy. 

Keywords: pharmaceutical agents, customer relationship management, rough 

set theory 

2 - Strategic investment timing under asymmetric access 

charge regulation in telecommunications 

Takashi Shibata, Graduate School of Social Sciences, Tokyo 

Metropolitan University, 1-1, Minami-osawa„ Hachioji, 

192-0397, Hachioji, Tokyo, Japan, tshibata@tmu.ac.jp 

In a liberalized telecommunications market, an incumbent has several advantages 

over any entrant. An asymmetric access charge regulation for two such 

asymmetric firms stimulates competitive investment. We show that an entrant 

with a cost disadvantage has an incentive to invest as a leader under an asymmetric 

access charge regulation. These results fit well with the findings of 

previous empirical work. Moreover, we also investigate the effects of an asymmetric 

access charge regulation on competitive investment strategies. 

3 - Combination strategies in predictive-prescriptive setups 

Paulo Freitas, CIO and University of Madeira, Centro de 

Ciencias Exactas e da Engenharia, Campus Universitario da 

Penteada, 9000-390, Funchal, Portugal, paulo@uma.pt, Antonio 

Rodrigues

In decision problems involving asymmetric loss functions, the optimal decision 

should be based on a quantile of a density forecast, rather than simply given by 

an expected-value forecast. Also, the combination of forecasts produced by 

different models may improve forecasting accuracy, as much as the combination 

of decision proposals from different sources may reduce the expected 

regret. We thus investigate and compare two possible approaches: either inferring 

decisions from combined predictive estimates, or combining prescriptive 

solutions derived from different forecasting models. 

4 - Classification by Archipelago Subgraph 

Bela Vizvari, Industrial Engineering, Eastern Mediterranean 

University, Gazimagusa, Mersin 10, Turkey, 

vizvaribela@gmail.com 

An ideal clustering of stocks is that there are positively correlated ones in each 

cluster and among stocks of different clusters the correlation is negative. This 

state can be achieved only if correlations close to zero are neglected. The ideal 

case is an archipelago. Clustering is done by deleting edges from a signed 

graph to get an archipelago. Algorithms are provided to determine that how 

high weights must be deleted and to find an archipelago subgraph with maximal 

total weight. Joint work with P. L. Hammer, P. Majlender, and B. Simeone 

� TD-19 



Dynamical Systems 



Chair: João Almeida, Instituto Politécnico de Bragança, 5301-857, 

Bragança, jpa@ipb.pt 

Chair: Flávio Ferreira, Mathematics, ESEIG - Instituto Politécnico do 

Porto, R. D. Sancho I, 981, 4480-876, Vila do Conde, Portugal, 

flavioferreira@eu.ipp.pt 

1 - Pinto’s Golden Tilings 

João Almeida, Instituto Politécnico de Bragança, 5301-857, 

Bragança, jpa@ipb.pt 

We present the definition of a golden sequence. These golden sequences are Fibonacci 

quasi-periodic and determine a tiling of the real line. We prove the existence 

of a natural one-to-one correspondence between: (i) Golden sequences; 

(ii) Smooth conjugacy classes of circle diffeomorphisms with golden rotation 

number that are smooth fixed points of renormalization, and (iii) Smooth conjugacy 

classes of Anosov diffeomorphisms that are topologicaly conjugate to 

the toral automorphism G_A=(x+y,x). The Pinto-Sullivan tilings of the real 

line relate smooth conjugacy classes of expanding circle maps with 2-adic sequences. 

2 - Dynamic Thresholds in Biology 

Alberto A. Pinto, Departamento de Matematica, University of 

Minho, Escola de Ciências, Universidade do Minho, 4710-057, 

Braga, Portugal, aapinto1@gmail.com 

3 - Leadership in a differentiated and uncertain market 

Flávio Ferreira, Mathematics, ESEIG - Instituto Politécnico do 

Porto, R. D. Sancho I, 981, 4480-876, Vila do Conde, Portugal, 

flavioferreira@eu.ipp.pt, Fernanda A. Ferreira, Alberto A. Pinto 

We present a brief study of the effects of product differentiation in a Stackelberg 

model with demand uncertainty. We do an ex-ante and ex-post analysis, in 

terms of product differentiation and of the demand uncertainty, of the profits of 

the leader and of the follower firms. We show that even with small uncertainty 

about the demand, the follower firm can achieve greater profits than the leader, 

if their products are sufficiently differentiated. 

4 - Effects of Individual Dynamic Complexity Elements on 

the Overall Complexity of a Simulation Game 

Onur Ozgün, Department of Industrial Engineering, Bogazici 

University, Bebek, 34342, Istanbul, Turkey, 

onur.ozgun@boun.edu.tr 

EURO 24 - Lisbon 2010 TD-20 

System dynamics based simulation games are commonly used to assist learning 

about complex systems. Though, it’s unclear whether players can acquire 

transferrable learning by playing games. The main goal of this research is to assess 

the extent of learning under different game complexity levels. As the first 

part of this research, this paper focuses on explaining the overall game complexity 

in terms of dynamic complexity factors: delay, nonlinearity, stock and 

feedback. Using a two-stage experimental procedure, a complexity measure is 

sought, based on player performances and their assessments 

� TD-20 


1.3.33A 




Chair: J. M. Valério de Carvalho, Departamento de Produção e 


vc@dps.uminho.pt 

1 - A constructive Heuristic for the Minimization of Open 

Stacks Problem 

Marco Carvalho, Computer Science Department, Technological 

Institute of Aeronautics, Praça Marechal Eduardo Gomes, 50, 

Vila das Acácias, CEP 12.228, São José dos Campos, São Paulo, 

Brazil, mamc@ita.br, Nei Yoshihiro Soma 

We present a constructive heuristic for the minimization of open stacks, a sequencing 

problem related to cutting stock. The method uses a quantity of techniques: 

breadth-first search, permutations generation and an anticipatory closing 

stack heuristic rule. Computational experiments encompassed a benchmark 

and a random set of instances. The gap to the first set is of 1.82% over almost 

six thousand instances, and a comparison with two other heuristics from the 

literature using the second set shows that the constructive heuristic has a better 

performance, opening far fewer stacks on average. 

2 - Sequencing bidimensional guillotine cutting operations 

considering lateness and raw material utilization 

Felipe Lemos, Production Engineering Department, 

Universidade de São Paulo, Av. Prof. Almeida Prado, Travessa 2, 

N o 128, Cidade Universitária, 05508-070, São Paulo, São Paulo, 

Brazil, felipeklemos@gmail.com, Miguel Cezar Santoro 

The paper analyses sequencing on guillotine cutting aiming to improve lateness 

and raw material utilization. It studies the unexplored intersection of 2 classic 

themes, cutting & packing and sequencing. Ten constructive heuristics were 

tested on different conditions of items sizes, lateness intensity and decision criteria 

weighs. The results showed the advantage of probabilistic heuristics; the 

importance of the chosen bias; and also of coherency between initial order and 

the weights of the objectives aimed. The implantation on an aircraft factory 

resulted better raw material utilization. 

3 - Pattern Cutting and Sequencing in a Multi-Stage Production 

System 

Rifat Gürcan Özdemir, Industrial Engineering Department, 

Istanbul Kültür University, Atakoy Campus, Atakoy-Bakirkoy, 

34156, Istanbul, Turkey, rg.ozdemir@iku.edu.tr, Tülin Aktin 

In this study, an integrated mathematical approach is proposed for the pattern 

cutting and sequencing problem in a multi-stage production system. The completion 

times of final products are dependent upon the cutting and sequencing 

of patterns based on which raw materials are cut in the first stage. The developed 

ILP model with the objective of minimizing WIP and tardiness costs, 

determines the number of hardboards that will be cut and obtains the sequence 

of these cutting patterns simultaneously. The model is then implemented in a 

furniture manufacturer operating on a make-to-order basis. 

4 - Sequencing cutting patterns in order to minimize the 

stack occupation 

Isabel Cristina Lopes, Dep. Matematica, ESEIG - Instituto 

Politecnico do Porto, Rua D.Sancho I, 981, 4480-876, Vila do 

Conde, Portugal, tulicreme@netcabo.pt, J. M. Valério de 

Carvalho 

179


We address a situation where machines process one cutting pattern at a time 

and equal pieces cut from the patterns are piled in stacks in the work area. The 

sequence in which preset cutting patterns are processed can affect the flow and 

total completion time, so it is desirable to optimize the occupation of the stacks 

to eliminate unnecessary dispersion. A solution can be modelled by an interval 

graph exhibiting a set of intervals that match the duration of stacks. We propose 

an IP model that reduces the occupation of the stacks, by adding the least 

number of edges to the graph. 

� TD-21 


6.2.47 

Software for OR/MS II - Open Source 



Chair: Vinícius Armentano, Faculdade de Engenharia Elétrica e de 

Computação, Universidade de Campinas, FEEC-UNICAMP- CP 

6101, Av. Albert Einstein 400, 13083-970, Campinas, São Paulo, 

Brazil, vinicius@densis.fee.unicamp.br 

1 - Open-Source Software in OR Education 

Jorge Santos, Matemaatica, Univ Evora, Rua Romão Ramalho, 

59, 7000-671 Évora, 7000-671, Évora, Portugal, 

jmas@uevora.pt, Luís Cavique, Armando Mendes 

This contribution will focus on Computational Tools of Open-Source Software 

in OR Education. Some educational experiences in the area of Forecasting; 

Simulation; Graphs and Networks; Decision Theory and Linear Programming 

based on: R 2.10.0, Scilab 5.1.1 and an Open Source Spreadsheet will be illustrated, 

with a brief reference to the acceptance of pupils and colleagues. 

2 - Parallel Machine Scheduling Using Free Software: an 

Application 

António Duarte, Instituto Politécnico de Bragança, Campus de 

Santa Apolónia, Apartado 1038, 5301-854, Bragança, Portugal, 

aduarte@ipb.pt, J. M. Valério de Carvalho 

We will show how to implement large scale optimization by only using freely 

available software tools. We solve exactly a parallel machine scheduling problem 

with identical parallel machines and malleable tasks, subject to arbitrary 

release dates and due dates. The objective is to minimize a function of late work 

and setup costs. We use the COIN-OR BCP framework to implement column 

generation to solve a model that results from a Dantzig-Wolfe decomposition, 

and also CRIFOR MCFZIB to solve an equivalent network flow model. Computational 

results are presented. 

3 - Consumers and Suppliers Decision Process in Information 

System Markets 

João Rosário, Marketing, Escola Superior de Comunicação 

Social-Instituto Politécnico de Lisboa, Portugal, 

jrosario@escs.ipl.pt, António Palma dos Reis 

The objective of this presentation is to discuss the consumer decision process 

in the Information Systems markets and how these markets evolve influenced 

by these consumer decisions and the suppliers’ decisions and business models. 

Will be discussed the factors that have influence on the consumers buying process 

decision and the option between Proprietary Software and Open Source 

Software in the Operating Systems and Office Suites categories; the gratuity 

degree of Open Source Software; and also on the supply side the advantages 

and disadvantages regarding innovation, software developing organization and 

market survival of Open Source Software versus Proprietary Software business 

models. 

4 - Fleet Deployment Optimization for Tramp Shipping 

Vinícius Armentano, Faculdade de Engenharia Elétrica e de 

Computação, Universidade de Campinas, FEEC-UNICAMP- CP 

6101, Av. Albert Einstein 400, 13083-970, Campinas, São Paulo, 

Brazil, vinicius@densis.fee.unicamp.br, Rodrigo Branchini 

We address a tactical planning problem faced by many tramp shipping companies 

that have cargo contracts which they are committed to carry, while trying 

to serve optional spot cargoes to increase their revenue over medium-term 

horizon. The decisions include the number and type of vessels deployed, the 

assignment of vessels to contractual and spot voyages and the determination 

of vessel routes and schedules in order to maximize the profit. This problem 

is modeled as a mixed integer programming which is solved using COIN-OR 

open source platform. Computational results are reported. 

180 

� TD-22 



Health Care Policy Making II 



Chair: Jose luis Andrade, Industrial Management, School of 

Engineering, University of Seville, Camino de los Descubrimientos 

s/n, 41092, Seville, Spain, jlandrade@esi.us.es 

1 - Innovative Approach to Design Decisions on a Regional 

Network of Clinical Laboratories 

Jose luis Andrade, Industrial Management, School of 

Engineering, University of Seville, Camino de los 

Descubrimientos s/n, 41092, Seville, Spain, jlandrade@esi.us.es 

We present a model supporting design and operation decisions on a regional 

network of clinical laboratories (able to reassign resources, redistribute workloads 

and reorganize flows), modelled as a multicommodity flow problem 

which conducts to a MIP formulation. We build a prototype tool for Health 

Care Managers to analyse scenarios and select the best options to improve the 

behaviour of the whole network. It is a graphical tool that launches an optimization 

process and supports: zoom, iconographical presentation of solutions 

and on-click access to input data or output solution, among others. 

2 - A Case-based Reasoning System for Radiotherapy 

Treatment Planning in Brain Cancer 

Rupa Jagannathan, Computer Science, University of Nottingham, 


Nottinghamshire, United Kingdom, rxj@cs.nott.ac.uk, Sanja 

Petrovic, Angela Mckenna, Louise Newton 

A decision support system for brain cancer radiotherapy treatment planning 

is presented. The aim of treatment planning is to attain a uniform tumouricidal 

dose for the tumour cells while minimizing the damage caused to adjacent 

healthy tissue and organs. This is a complex decision-making process that relies 

on subjective experience and expert clinical domain knowledge. We develop a 

case-based reasoning system that generates treatment plans for new patients 

based on the plans for previous similar patients. Our experiments, which use 

real brain cancer patient cases, show promising results. 

3 - Decision Support System for Warfarin Therapy Management 

Barbaros Yet, Technology Management and Economics, 

Chalmers University of Technology, Sweden, 

barbaros@student.chalmers.se, Kaveh Bastani, Hendry Raharjo, 

Svante Lifvergren, Bo Bergman 

Warfarin therapy is known as a complex process due to variation in the patients’ 

response.Failure to deal with such variation may lead to thrombosis or 

bleeding.There have been studies done on investigating the sources of variation, 

such as alcohol consumption and interacting drugs. However, this knowledge 

is, unfortunately, often used loosely by the physicians.This paper proposes a 

decision support system to integrate experts’ knowledge in a systematic way 

using Bayesian Network.The model is built upon literature review in medical 

fields and interviews with doctors in a Swedish hospital. 

� TD-23 


6.2.49 

MOO: Nonlinear Multi-Objective 

Optimization Techniques in Action 



Chair: Jussi Hakanen, Dept. of Mathematical Information 

Technology, University of Jyväskylä, P.O. Box 35 (Agora), FI-40014, 

University of Jyväskylä, Finland, jussi.hakanen@jyu.fi 

Chair: Kaisa Miettinen, Dept. of Mathematical Information 

Technology, University of Jyvaskyla, P.O. Box 35 (Agora), FI-40014, 

University of Jyvaskyla, Finland, kaisa.miettinen@jyu.fi

1 - A synchronous Reference Point scheme for Stochastic 

Multiobjective Problems 

Mariano Luque, Applied Economics (Mathematics), University 

of Malaga, Ejido, 6, 29071, Malaga, Spain, mluque@uma.es, 

Jose Manuel Cabello, Ana Belen Ruiz, Francisco Ruiz 

In this work, we propose an interactive synchronous approach for stochastic 

multiobjective programming, which is based on the achievement scalarizing 

function approach, and where the reference point philosophy it adapted to the 

nature stochastic in several ways. At each iteration, the DM must provide a reference 

level for each objective and the risk level (probability) that he is prepared 

to assume for each objective. We show that three different scalarizing functions 

can be used for each iteration, depending on which infor-mation is regarded as 

reference point. Theoretical results are given which prove the efficiency of the 

optimal solutions obtained in the three cases. 

2 - Gemstone cutting - a tradeoff between beauty and volume 

Karl-Heinz Küfer, Optimization, Fraunhofer ITWM, 


kuefer@itwm.fhg.de, Anton Winterfeld 

Production of gemstones has in all times been a compromise between best possible 

volume exploitation of expensive rough materials and producing a beautiful 

jewel. Fraunhofer ITWM has deviced a mathematical model for decision 

support of this multicriteria problem for the gemindustry. The code based on a 

semiinfinite programming approach is working in daily business. The talk will 

survey mathematical models used and will demonstrate the functionality of the 

system with screen shots. 

3 - An interactive multi-objective approach to heat exchanger 

network synthesis 

Timo Laukkanen, Energy Technology, Aalto University School 

of Science and Technology, PO Box 14400, FI-00076 Aalto, PO 

Box 14400, FI-00076 Aalto, Finland, timo.laukkanen@tkk.fi, 

Tor-Martin Tveit, Vesa Ojalehto, Kaisa Miettinen, Carl-Johan 

Fogelholm 

The heat exchanger network synthesis problem is a very important problem for 

designing energy-efficient industrial processes. One approach in solving this 

synthesis problem has been to use single-objective MINLP models where the 

objective is to design a heat exchanger network that minimises total annualised 

cost, i.e. investment and energy costs. In this work we use a multi-objective 

approach to heat exchanger network synthesis. The approach solves a modified 

version of a known MINLP model using an interactive multi-objective optimisation 

method, NIMBUS, which is partly implemented in GAMS. 

4 - Interactive Wastewater Treatment Plant (WWTP) Design 

by Using Pareto Navigator 

Jussi Hakanen, Dept. of Mathematical Information Technology, 

University of Jyväskylä, P.O. Box 35 (Agora), FI-40014, 

University of Jyväskylä, Finland, jussi.hakanen@jyu.fi, Kaisa 

Miettinen 

In WWTP design, conflicting objectives need to be considered simultaneously, 

e.g. operational requirements and economical efficiency. Finding the most 

preferred design involves solving a multiobjective optimization (MO) problem 

where objective values are obtained via computationally costly numerical 

simulation. We apply an interactive MO method Pareto Navigator utilizing 

an approximation of the Pareto optimal set instead of costly simulation when 

searching for the best compromise design. The benefits include a reduced computational 

time and interactive design process enabling learning. 

� TD-24 


6.2.50 

Bioinformatics V 


Medicine 


Chair: Metin Turkay, Department of Industrial Engineering, Koc 

University, Rumelifeneri Yolu, Sariyer, 34450, Istanbul, Turkey, 

mturkay@ku.edu.tr 


1 - Cancer classification using hyper-box enclosure 

method 

Onur Dagliyan, Chemical and Biological Engineering, Koc 

University, Rumelifeneri yolu, Sariyer, 34450, Istanbul, Turkey, 

odagliyan@ku.edu.tr, Halil Kavakli, Metin Turkay 

One of the applications of microarray experiments is the classification of tumor 

types with respect to their gene profiles.In this work, we employed a mixed 

integer programming based "hyper-box enclosure method’ for the classification 

of some cancer types such as leukemia, prostate cancer, lymphoma, small 

round blue cell tumors with a minimal set of predictor genes. Due to its nonparametric 

structure and accurate prediction rate, hyper-box enclosure method 

is a robust and convenient tool for tumor prediction with a small number of 

gene markers. 

2 - Epidemic modeling by mixtures of binomial distributions 

with susceptible/exposure/infection dependence 

Eva Maria Ortega, CIO, University Miguel Hndez, Spain, 

evamaria@umh.es, Isabel Ortega 

Stochastic SIR (susceptible-infective-removed) models for the mathematical 

modelling of the spread of an infectious disease, have received considerable 

attention. To incorporate individual heterogeneity, or structure mixing patterns 

and randomness for some parameters constitute key issues. Uncertainties are 

due to either environmental, clinical and exposure or social, geographical and 

personal features. Correlations arise from several factors influencing the epidemics 

in the heterogeneous populations. The impact of these dependencies 

has been studied by different approaches, mainly from simulation and semiparametric 

modelling. Recently, Ortega and Alonso (2010) provided a general 

framework and analytical results to study the dependence for several epidemic 

models. We provide stochastic ordering results for the analysis of some epidemics 

where the number of infectives (susceptibles) is a mixture of a binomially 

distributed population conditioning upon certain amounts. Some measures 

for the risk assessment are compared. We consider a general stochastic model 

for the population of interest, with some random parameters, that leads to analyze 

some models in recent literature: randomized carrier-borne epidemics, 

infector or/and exposure dependent severity models, branching-type representations. 

3 - Assignment problems for medical image registration 

Michael Stiglmayr, Institute of Mathematics, University of 

Wuppertal, Gaußstr. 20, 42119, Wuppertal, Germany, 

stiglmayr@math.uni-wuppertal.de, Kathrin Klamroth 

Image registration has become an indispensable tool for diagnosis and operation 

planning. Combinatorial, feature-based registration approach have the 

advantage that global optima can be determined by applying well-researched 

Branch and Bound algorithms, however, at a potentially high cost. The application 

of a generalized, mixed-integer linear and a quadratic assignment problem 

is discussed in the context of feature-based image registration. Computational 

results as well as a multicriteria parameter analysis of the model are presented. 

4 - Protein domain networks: analysis of error and attack 

tolerance under different circumstances 

Saziye Deniz Oguz, Institute of Applied Mathematics, Scientific 

Computing Program, Middle East Technical University, Orta 

Dogu Teknik Universitesi, Inonu Bulvari, Cankaya, 06531, 

Ankara, Turkey, doguz@metu.edu.tr, Hakan Oktem 

The stability or survivability of some complex networks under different circumstances 

has received a growing interest among scientists. In this work we study 

the protein domain networks generated with data from the ProDom and Pfam 

domain databases with an essential interest in both error and attack tolerance. 

The topology of a complex network can be characterized by distribution of the 

connectivity. We first find out whether networks of same connectivity distribution 

have same robustness to error and attack. Secondly we explore whether 

there are evolutionary means to create such networks. 

181


� TD-25 


6.2.48 


Optimization III 






1 - Fuzzy portfolio selection using a genetic algorithm 

Enriqueta Vercher, Spain, enriqueta.vercher@uv.es, José D. 

Bermúdez, José Vicente Segura 

We consider the problem of finding the efficient frontier associated with the 

fuzzy mean-downside risk portfolio selection model. The return and risk of the 

portfolio are measured through the value and ambiguity of the fuzzy number 

that represents the uncertainty of the return. The feasible set includes the diversification 

and cardinality constraints. We develop a strategy for fuzzy ordering 

the portfolios using a genetic algorithm that provides a set of acceptable portfolios. 

Illustrative examples are provided for the historical returns of a set of 

assets. 

2 - Maxentropic Estimation of the Failure Probability Function 

Alin Rusu, Faculty of Mathematics and Computer Science, 

University of Bucharest, Academiei Street, no. 14, sector 1, 

010014, Bucuresti, Romania, rusualinmarian@yahoo.com 

In this paper we will describe the failure probability function and its properties 

and we will estimate the failure probability function. The failure probability 

function is a function that depends on the design variables. The maximum 

entropy principle will be used in estimating the failure function. Also the confidence 

interval of the failure probability function will be found. 

3 - S&P 500 Stocks Screening Using Multiple Criteria 

Mohsen Bahramgiri, Graduate School of Management and 

Economics, Sharif University of Technology, Azadi St„ Tehran, 

Iran, Islamic Republic Of, bahramgiri@sharif.edu 

In this work, an optimizing method based on the multiple criterion decision 

making for ranking the firms’ stocks with the most correlation between past 

performance and future success of the firms have been proposed. Ten financial 

rations of the firms were treated as: "past performance" and the second one "future 

success" has been built based on the two concepts: the maximum lost due 

to both buying and selling of the stocks of a single firm, and the maximum revenue. 

Finally three aggregation modes have been used to bring out two general 

criteria. 

4 - Possibilities of Calibrating Low-Default Portfolios with 

Parametric Approaches 

Vesna Bogojevic Arsic, management, Faculty of organizational 

sciences, Jove Ilica 154, 11000, Belgrade, Serbia, 

bogojevic@fon.rs 

Development of internal rating based models is very difficult for banks because 

of insufficient data of low-default portfolios that are used for backtesting these 

models. The paper considers possibilities of calibrating low-default portfolios 

using parametric approaches. One approach is to calibrate these portfolios using 

so called power curve, while the second uses receiver operating characteristic. 

Both approaches provide statistical measures to assess the discriminatory 

power of various rating models and can be used for rating model comparison. 

� TD-26 


3.1.11 

Dynamics, statistical inference and 

algorithms of cooperative game theory 



Chair: Marco Dall’Aglio, Dept of Economics and Business, LUISS 

University, Viale Romania 32, Rome, Italy, mdallaglio@luiss.it 

182 

1 - Bayesian inference for weighted voting games 

Pierpaolo Brutti, Department of Economics and Business, 

LUISS Guido Carli, Viale Romania 32, 00157, Roma, Italy, 

pierpaolo.brutti@gmail.com, Marco Scarsini 

Weighted voting games are ubiquitous models which are used in many fields 

ranging from economics and political science to bioinformatics and machine 

learning. They model scenarios where agents with associated a non-negative 

weight, vote in favour or against a decision. The decision is made if and only if 

the total weight of those voting in favour of the decision is equal to or greater 

than a given quota. In this work we consider a pre-electoral setup, so to speak, 

where the weight configuration has to be learnt from sampling data. Tackling 

this statistical problem from a Bayesian viewpoint, we place a Dirichlet prior 

over the weight vector to obtain – by coniugacy – a Dirichlet posterior distribution 

that can then be used to make inference on various solution concepts of 

interest. In particular we consider the posterior behaviour of the Shapley value, 

a key index used to quantify the "political" power of each agent, and show how 

useful some tools from computational and discrete geometry are in this respect. 

Finally we briefly touch upon an extension of our framework where also the cooperation 

structure is modeled as a random (weighted) graph with connectivity 

estimated from (historical) data. 

2 - Financial games with external opportunities 

Camilla Di Luca, Scienze, University "G. D’Annunzio" 

Chieti-Pescara, Viale Pindaro 42, 65127, Pescara, Italy, 

love_saved@hotmail.it, Carles Rafels, Josep M Izquierdo 

In financial games, from the point of view of cooperative game theory, the proportional 

distribution arises as the most prominent allocation method and the 

core of the associated cooperative game shrinks into it if and only if there are 

no decisive investors. We consider the possibility of an external investor and 

discuss the problem within the framework of coalition structures. Generally, 

the core of the game turns out to be empty, so we focus on the bargaining set 

and give a sufficient condition to guarantee that the bargaining set only includes 

the proportional distribution. 

3 - Algorithms in Cooperative Fair Division 

Marco Dall’Aglio, Dept of Economics and Business, LUISS 

University, Viale Romania 32, Rome, Italy, mdallaglio@luiss.it, 

Camilla Di Luca 

We considers the situation in which several agents attend the apportionment of 

a completely divisible good (such as a piece of land, or a cake) and are willing 

to cooperate to achieve a better division. We provide algorithms for finding 

values of the cooperative games associated to the division. These algorithms 

are based on a geometric model for fair division as well as on duality considerations 

over the same model. 

� TD-27 


8.2.06 

URBAN TRAFFIC CONTROL 



Chair: Tobias Pohlmann, Institut für Verkehr und Stadtbauwesen, 

Technische Universität Braunschweig, Rebenring 31, 38106, 

Braunschweig, Germany, t.pohlmann@tu-braunschweig.de 

1 - Online Offset Optimization of Traffic Lights using 

Vehicle-to-Infrastructure Communication 

Daniel Schmidt, Technical University Braunschweig, Institut of 

Transportation and Urban Engineering, Rebenring 31, 38106, 

Braunschweig, Germany, dan.schmidt@tu-braunschweig.de, 

Tchouankem Hugues 

Optimization of traffic lights increases road capacity and traffic flow, and can 

prevent traffic congestions in urban areas. Through the dynamic exchange 

of messages between vehicles and infrastructure online offset optimization 

for constructing a progressive signal system is possible. Within the project 

PLANETS we study methods for the offset optimization using realistic V2Icommunications. 

To validate our research, we connect the traffic simulator 

AIMSUN with a model that implements the traffic light control algorithm as 

well as the communication between vehicles and traffic lights.

2 - Decentralized Decision Support and Coordination of 

Autonomous Vehicles Based on Online Data Mining 

Techniques 

Maksims Fiosins, Department of Informatics, Clausthal 

University of Technology, Julius-Albert-Strasse 4, D-38678, 

Clausthal-Zellerfeld, Germany, maksims.fiosins@tu-clausthal.de, 

Jana Görmer, Jan Fabian Ehmke 

We present an approach for decision making and coordination of vehicles in 

urban traffic systems. Our concept is based on the application of multiagent 

systems by representing vehicles as agents in a centrally controlled environment. 

They gain recommendations based on reinforcement learning methods, 

represented in the form of (decentralized) Markov Decision Processes. In order 

to reduce the complexity of joint learning, we organize agents in groups of 

limited size. Methods from the field of online data mining provide a consistent 

information models for learning and group formation processes. 

3 - Optimisation of signal change intervals from a capacity 

point of view 

Axel Wolfermann, Transport Planning and Traffic Engineering, 

Technische Universität Darmstadt, Petersenstr. 30, 64287, 

Darmstadt, Germany, wolfermann@verkehr.tu-darmstadt.de 

To improve the efficiency of signal change intervals at signalised intersections 

it is important to connect the safety evaluation with a capacity evaluation. The 

capacity impact of signal change intervals is determined by the traffic flow 

characteristics, traffic volumes, and intersection design. A methodology is presented 

to comprehensively analyse the factors leading to capacity reductions by 

signal change intervals. This methodology is the basis for an optimisation of 

signal change intervals from a capacity point of view. 

4 - Model-based Adaptive Traffic Control employing Genetic 

Algorithms 

Tobias Pohlmann, Institut für Verkehr und Stadtbauwesen, 

Technische Universität Braunschweig, Rebenring 31, 38106, 

Braunschweig, Germany, t.pohlmann@tu-braunschweig.de, 

Bernhard Friedrich 

A new adaptive traffic control strategy for urban networks with several signalized 

intersections has been developed. It optimizes signal plans for consecutive 

time intervals of 15 minutes based on forecasted traffic demands. Its main feature 

is a heuristic model-based offset optimization. Three different algorithms 

are proposed: Parallel and Serial Genetic Algorithm and Sequential Enumeration. 

An extended Cell Transmission Model is used to calculate the fitness in 

terms of delay of possible solutions. A comprehensive microsimulation study 

has been conducted to assess the new strategy. 

� TD-28 



Scheduling in Health Care 

Stream: Scheduling 


Chair: Dobrila Petrovic, Faculty of Engineering and Computing, 

Coventry University, Priory Street, CV1 5FB, Coventry, United 

Kingdom, D.Petrovic@coventry.ac.uk 

1 - Scheduling Treatment at a Rehabilitation Hospital 

Richard M Wood, School of Mathematics, Cardiff University, 

Senghennydd Road, Cardiff, United Kingdom, 

woodrm@cardiff.ac.uk, Jeff Griffiths, Janet Williams 

A multi-objective hierarchical scheduling problem will be presented with application 

to a major UK rehabilitation hospital. The project will investigate the 

weekly allocation of physiotherapy treatment slots to patients. The aims are 

to reduce the time spent on timetable production and to improve the timetable 

quality. An automated program is therefore developed comprising a series of 

greedy heuristics and steepest descent based local search algorithms. The terminating 

criterion will also be addressed. 

2 - Heuristics for planning elective surgeries 

Inês Marques, DEIO - CIO, Faculdade de Ciências da 

Universidade de Lisboa, Campo Grande - Ed. C6 - 4 o Piso, 


1749-016, Lisbon, Portugal, ines.marques@fc.ul.pt, Maria 

Eugénia Captivo, Margarida Pato 

Resources rationalization is an imperative of our times, also present in health 

care services. In Portugal, the hospital surgery suite also has the important objective 

of reducing waiting lists for surgery. Both objectives are outlined in the 

National Health Plan for 2004-2010. Heuristics developed to address those objectives 

in the elective surgeries planning, as well as results obtained from their 

application to real data from a hospital in Lisbon, are presented. These results 

are also compared with results obtained through an integer linear programming 

model approach. 

3 - Heuristics for radiotherapy scheduling 

Dobrila Petrovic, Faculty of Engineering and Computing, 

Coventry University, Priory Street, CV1 5FB, Coventry, United 

Kingdom, D.Petrovic@coventry.ac.uk, Truword Kapamara 

Novel scheduling heuristics for booking appointments for radiotherapy patients 

in Arden Cancer Centre, UK were developed. The heuristics involved the following 

parameters: a) maximum allowed breach of the patients’ waiting time 

targets, b) number of reserved slots on treatment machines for different types 

of treatments, and c) number of overtime slots. Various tests were carried out 

to analyse the impact of these parameters on the schedule performance. 

4 - Operating room planning and scheduling: a multiobjective 

model solved via a genetic algorithm 

Rosita Guido, of Electronics, Computer Science and Systems, 

University of Calabria, Ponte Pietro Bucci, Arcavacata di Rende 

(CS), 87036, Rende, Italy, rguido@deis.unical.it, Domenico 

Conforti, Francesca Guerriero 

In recent years, an increasing interest of Operations Research is on the domain 

of operating room (OR) planning and scheduling. In this complex context, 

known the conflicting nature of several goals, we have developed a multiobjective 

model whose aim is to assign surgeries of different surgical specialties 

to multiple operating rooms in a block scheduling system. The set of efficient 

solutions have been determined on several scenarios (real-life and randomly 

generated data) by adopting a genetic algorithm. Preliminary results confirm 

the importance of multiobjective model in OR management. 

� TD-29 


8.2.11 

Preference Learning I 

Stream: Preference Learning 


Chair: Roman Slowinski, Institute of Computing Science, Poznan 

University of Technology, Laboratory of Intelligent Decision Support 

Systems, Street Piotrowo 2, 60-965, Poznan, Poland, 

roman.slowinski@cs.put.poznan.pl 

1 - Learning of rule preference models for a mixture of nonordinal 

and ordinal attributes 

Jerzy Blaszczynski, Institute of Computing Science, Poznan 


jerzy.blaszczynski@cs.put.poznan.pl, Salvatore Greco, Roman 

Slowinski 

We induce from data decision rule models that represent preferences in terms 

of logical statements. In case of multi-attribute classification their syntax is: "if 

performance of action a is better (or worse) than given values of some attributes, 

then a belongs to at least (at most) given class’, and in case of multi-attribute 

ranking: "if action a is preferred to action b in at least (at most) given degrees 

with respect to some attributes, then a is preferred to b in at least (at most) given 

degree’. The input data are examples of classification or ranking decisions described 

by a mixture of non-ordinal and ordinal attributes. The above rules are 

induced within the Dominance-based Rough Set Approach (DRSA). While the 

syntax of DRSA rules involves dominance relation in the condition part, such 

rules are also useful when non-ordinal attributes are considered, because they 

provide comprehensible pros and cons arguments for a decision. Moreover, 

they permit to avoid discretization of numerical attributes prior to induction of 

rules. We give examples of induction of rule preference models and we prove 

their usefulness in a comparative experiment. 

183


2 - Conditional Ranking on Relational Data 

Willem Waegeman, Department of Applied Mathematics, 

Biometrics and Process Control, Ghent University, Coupure links 

653, 9000, Ghent, willem.waegeman@ugent.be, Tapio 

Pahikkala, Antti Airola, Tapio Salakoski, Bernard De Baets 

In domains like bioinformatics, information retrieval and social networks analysis, 

one can find learning tasks where the goal consists of inferring a ranking of 

objects, conditioned on a target object. We present a general kernel framework 

for learning conditional rankings from various types of relational data, where 

rankings can be conditioned on unseen data objects. Symmetric and reciprocal 

relations can in this framework be treated as two important special cases, for 

which the task simplifies to a traditional ranking task when the relations satisfy 

certain restrictive properties. 

3 - How to establish a mathematical relationship between 

temperament and competences ? From a classification 

problem to a label ranking problem. 

Massimo Gurrieri, University of Mons, Rue du Houdain 9, 7000, 

Mons, Belgium, Massimo.Gurrieri@umons.ac.be, Philippe 

Fortemps 

In the context of recruitment, the recruiter is often interested in establishing a 

ranking among competences profiles of the candidates. But such a ranking of 

the different competences (e.g. soft skills) of one candidate is hard to establish. 

We are looking for rules of the following kind: "if candidate is more temperament 

T1 than T2, then his competence C1 is better than competence C2’. Rules 

are generated by means of DRSA for pair-wise comparisons of competences, 

and then more general rules are induced by merging previous ones on a logical 

basis. 

4 - Evolutionary Fuzzy Rules for Binary Classification 

Problems with Preferences 

Christian Moewes, Faculty of Computer Science, University of 

Magdeburg, Universitaetsplatz 2, 39114, Magdeburg, Germany, 

cmoewes@ovgu.de, Roman Slowinski, Izabela Szczech, Rudolf 

Kruse 

We present an approach to learn fuzzy binary classifiers from ordinal data. We 

assume that one class is preferred to the other, e.g., it must not be misclassified. 

Hence we use Dominance-based Rough Set Approach to select relevant fuzzy 

rules. Fuzzy partitions are tuned by an evolutionary algorithm trying to minimize 

classification accuracy of the current rule base. The model complexity 

in terms of number of rules is controlled by fuzzy confirmation measures. We 

discuss similarities to interval-based fuzzy classifiers and compare their performances 

to our approach on benchmark data sets. 

� TD-30 


8.2.13 

MCDM 1 


other Issues [c] 


Chair: Denis Yannacopoulos, Department of Business 

Administration, Technological Educational Institute of Piraeus, 

12244, Egaleo, Greece, dgian@teipir.gr 

1 - Predictive Maintenance Program Audit by means of a 

multicriteria decision support methodology 

Maria Carmen Carnero, Technical School of Industrial 

Engineering, University of Castilla-La Mancha, Avda. Camilo 

José Cela s/n, 13071, Ciudad Real, Ciudad Real, Spain, 

Carmen.Carnero@uclm.es, Mónica Oliveira, Carlos Bana e 

Costa 

Auditing tools have a key role in promoting continuous improvement in a 

predictive maintenance program (PMP). This study proposes (and applies) 

a framework for auditing a PMP based on a multicriteria decision support 

methodology that consists on: a hierarchical model identifying and describing 

the auditing criteria in which the performance of the PMP should be appraised; 

an additive value model based on the MACBETH methodology to evaluate the 

level of accomplishment of the PMP; and a classification procedure to identify 

the dimensions and areas that demand for higher managerial attention. 

184 

2 - A Multicriteria Ordinal Measurement Method to Customer 

Satisfaction Evaluation: MOMENTS 

Isabel M. Joao, DEQ-ISEL, CEG-IST, R. Conselheiro Emidio 

Navarro„ 1959-007 , Lisbon, Portugal, ijoao@deq.isel.ipl.pt, 

Carlos Bana e Costa 

Understanding interactions between customer and the provided service, the crucial 

moments that lead to customer satisfaction (CS), is the key to improve the 

service. CS surveys ask questions about several criteria and typically use ordinal 

scales. To deal with the ordinal property we propose a multicriteria ordinal 

measurement method based on the additive value model. The value functions 

and criteria weights proposed by MOMENTS are obtained by an ordinal regression 

method with constraints. MOMENTS is illustrated with a real application 

in the hotel industry for guests satisfaction evaluation 

3 - Citizens’ Satisfaction from the Public Services: A Multicriteria 

Approach 

Denis Yannacopoulos, Department of Business Administration, 

Technological Educational Institute of Piraeus, 12244, Egaleo, 

Greece, dgian@teipir.gr, Panagiotis Manolitzas, Nikos Tsotsolas, 

Dimitrios Drosos 

Studying the administrative reforms that have taken place the last twenty years 

in the European Union we can assume that the basic principle of these reforms 

led to major changes in citizens’ expectations, organizational structures and 

working processes. With the byway of years in the public sector many practices 

have been adopted like Total Quality Management, and the replanning of 

business procedures. These practices have been adopted in order to help the 

public sector to work more effectively and efficiently. In this paper we measure 

the citizen’s satisfaction using an ordinal regression model based on the 

principles of multicriteria decision analysis. 

� TD-31 


8.2.15 

OR based Decision Support by Fraunhofer 

Society 



Chair: Sascha Herrmann, Decision Support Systems, Fraunhofer 

SCS, Nordostpark 93, 90411 , Nuremberg, Germany, 

sascha.herrmann@scs.fraunhofer.de 

1 - System-Assisted Cooperation of Independent Freight 

Forwarding Offices 

Jens Wollenweber, Decision Support Systems, Fraunhofer SCS, 

Nordostpark 93, 90411, Nürnberg, Germany, 

jens.wollenweber@scs.fraunhofer.de 

The talk gives an overview of the central ideas and the preliminary findings 

of a publicly funded research project in the road transport sector. The project 

deals with non-specialized partial and full truckload (FTL) tramp transports 

by non-specialized less-than-truckload (LTL) forwarders. This business sector 

is characterized by individual and manual planning processes, the lack of 

computer-aided dispatching, and virtually no information interchange between 

local offices. The purpose of the project is the improvement of the planning 

of FTL road transports in independent, decentral forwarding offices by means 

of a decision support system (DSS) for collaborative planning. The two main 

challenging research issues in the development such a DSS were (i) to find a 

practically viable process of collaboration, i.e., a scheme for trading transport 

requests and vehicles, and (ii) to devise a heuristic algorithm capable of solving 

large instances of multi-period, multi-depot pickup and delivery problems with 

heterogeneous fleet, time windows and driver rules. The central aspects of the 

solution approaches developed in the project are presented and discussed, and 

some new research issues that have arisen during this work are pointed out. 

2 - A Decision Support System for waste management organizations 

Sascha Herrmann, Decision Support Systems, Fraunhofer SCS, 

Nordostpark 93, 90411 , Nuremberg, Germany, 

sascha.herrmann@scs.fraunhofer.de

By directive of the EU existing substances like plastics or glass waste have to 

be collected, recycled and commercialized as secondary raw materials. While 

the value of these substances is small, the expenses for recycling and transportation 

are considerable. Therefore waste management organizations are interested 

in minimizing their expenses for this process. In this talk we present a 

generalized two-stage assignment and transportation model for this real-world 

problem, which can be solved within seconds. The model is integrated into a 

custom-made DSS, and generates substantial savings. 

3 - Optimal Scheduling of Transports with Interdependent 

Resources for the Advanced Truckload Concept 

Bettina Berning, Decision Support Systems, Fraunhofer SCS, 

Nordostpark 93, 90411, Nürnberg, Germany, 

bettina.berning@scs.fraunhofer.de 

The Advanced Truckload Concept is a new business model for the truckload 

transport market where drivers and vehicles should operate decoupled in an 

area wide network of driver stations and cargo occurrence. This enables more 

flexibility in scheduling. Particularly with regard to smart driver changes this 

may result in less operating time and cost. However the economic benefit is in 

contrast to the mathematical complexity this problem causes. As proof of the 

mathematical feasibility this talk will provide a mathematical formulation as 

well as optimal results for small instances. 

4 - CARP of Urban Municipal Solid Waste Collection 

Volker Engels, Fraunhofer IML, 44227, Dortmund, Germany, 

volker.engels@iml.fraunhofer.de 

This Article tackles a specification of real world urban Capacitated Arc Routing 

Problems (CARP). It deals with the collection of solid household waste in 

heterogeneous containers (bins, bags) with different service frequencies which 

belong to directed arcs of streets considering area structures and service levels. 

We face the challenge to find daily tours starting at depots traveling over arcs 

to be serviced and ending at waste plants. We consider real heterogeneous collection 

vehicles with fixed teams and some special constraints. Following the 

description of this specific problem this article proposes a multi-stage solution 

idea. 

� TD-32 


8.2.17 

OR in Forest Management 



Chair: Raul Brás, Matemática, ISEG/UTL - CEMAPRE, Rua Miguel 

Lupi, 20, Lisboa, 1249-078, Lisbon, Portugal, rbras@iseg.utl.pt 

1 - Dynamic log yard designs for an improved coordination 

of sawmill and log yard operations 

Luc LeBel, Sciences du bois et de la foret, université laval, 2405 

de la terrrase, g1v 0a6, Quebec, Quebec, Canada, 

luc.lebel@sbf.ulaval.ca 

Sawmills have traditionally kept high level of inventory in their yards for seasonal 

considerations. We attempt to find the optimal assortments to store 

through a Forward-Reserve Problem (FRP). The classical FRP was extended 

to a multi-period FRP in the context of a divergent process industry. The multiperiod 

context allows for changes in assignments to the forward area as market 

conditions change. In order to account for the divergent nature of the industry, 

the FRP formulation has been extended to anticipate production at the mill 

based on known demands and market anticipation functions. 

2 - On the validation of an occurrence forest fire index on 

a Mediterranean region 

Fortunato Crespo Abril, Estadística e Investigación Operativa 

Aplicadas y Calidad, Universidad Politécnica de Valencia, 

Camino de Vera, s/n, 46022, Valencia, Spain, 

fcrespo@eio.upv.es, Javier de Vicente 

This paper describes the methodology used to asses the capabilities of a new 

fire danger rating system. This index is used to map the wildland fire occurrence 

and risk on a Mediterranean area in Spain. A generalized estimating 

equations (GEE) model, which takes into account the correlated nature of responses 

is used to validate the occurrence fire index and to obtain a different 

mapping representation. 


3 - Two variable selection methods for fuzzy regression 

model 

Barbara Gladysz, Institute of Organization and Management, 

Wroclaw University of Technology, Wybrzeze Wyspianskiego 

27, 50-370, Wroclaw, Poland, barbara.gladysz@pwr.wroc.pl, 

Dorota Kuchta 

In the paper we propose two variable selection methods for a fuzzy regression 

model. The first one is a graph method and the second one is a backward selection 

technique. The main criterion of both methods is the significance of 

the regression coefficients. In the paper we also propose a criterion of fuzzy 

regression coincidence. In all the proposed techniques the possibility theory is 

applied. An example of an energy load model construction is presented. The 

set of a priori plausible variables consists of past levels of energy load variables 

and weather data. 

4 - A modified Steiner forest problem with applications in 

conservation biology 

Raul Brás, Matemática, ISEG/UTL - CEMAPRE, Rua Miguel 

Lupi, 20, Lisboa, 1249-078, Lisbon, Portugal, rbras@iseg.utl.pt, 

J. Orestes Cerdeira, Diogo Alagador, Maria Triviño, Mar 

Cabeza, Miguel Araújo 

The Steiner forest problem is a generalization of the Steiner tree problem where 

sets of terminal vertices have to be connected at minimum cost. We consider 

the connection of different types of environmental similar protected areas (PA), 

to allow free flow of species between PA of the same type. The paths should 

not include sites which are environmental too different from those of the PA 

they are linking. We designed an algorithm to deal with very large graphs, and 

applied it to all the PA of the Iberian Peninsula clustered in four different types. 

Computational outcomes are reported. 

� TD-33 


8.2.19 

Modelling of uncertainties in the energy 

sector (short-term planning) 



Chair: Christoph Weber, Universität Essen, Universitätsstr. 11, 

45117, Essen, Germany, Christoph_Weber@uni-duisburg-essen.de 

1 - Efficient treatment of the Stochastic Unit Commitment 

problem for long-term planning studies 

Alexander Sturt, Electrical and Electronic Engineering, Imperial 

College, London, SW7 2AZ, London, United Kingdom, 

alexander.sturt07@imperial.ac.uk, Goran Strbac 

We describe a new tool for simulating simplified, large-scale power systems 

with a significant renewables component. It solves the stochastic unit commitment 

problem at each timestep using a MILP solver. We attack the computational 

problem using a multi-pronged approach, including parallelisation, efficient 

modelling of many identical units, parsimonious construction of scenario 

trees, and importance sampling. These techniques reduce the computational 

burden to tolerable levels while still allowing us to draw meaningful, quantitative 

assessments of potential future power systems. 

2 - Spot-based storage valuation using a multi-factor price 

process 

Alexander Boogert, KYOS Energy Consulting / Birkbeck 

College, University of London, Netherlands, boogert@kyos.com, 

Cyriel de Jong 

In this talk we discuss a popular valuation method for gas storage. Our starting 

point is the spot-based approach as introduced in Boogert & De Jong (Journal 

of Derivatives, 2008). In this talk we extend the approach to multi-factor price 

processes. Such price processes can capture more realistically the actual price 

behavior. First, we compare two ways to incorporate multi-factor price process 

into the optimization. Next, we study the impact of using multi-factor price 

processes on different aspects of the valuation such as convergence, average 

storage value and distribution of storage values 

185


3 - Modelling Electricity Prices in the Selected Regions: 

Linear vs Non-linear Auto-Regressive Models 

Jiri Valecky, Finance, VSB-TUO, Sokolska tr. 33, 701 21, 

Ostrava, Czech Republic, jiri.valecky@vsb.cz, Miroslav Culik 

Because of the features occurring in the time series of electricity prices, we 

focused on more sophisticated models for the purposes of their modelling. We 

aim to the proposing the empirical models for modelling daily electricity prices 

in three selected regions (California, North Europe and Austria). We place the 

emphasis on the application of non-linear regime switching models and we also 

verify whether the proposed non-linear models give better results then the linear 

AR models and whether these results are satisfactory in the sense of data 

fitting and diagnostic checks. 

4 - Evaluation of energy technologies based on stochastic 

price simulations 

Dogan Keles, Institute for Industrial Production (IIP), Karlsruhe 

Institue for Technology (KIT), Germany, dogan.keles@kit.edu, 

Dominik Möst, Wolf Fichtner 

The liberalisation of the energy market requires new methods for the assessment 

of energy technologies. Therefore a new modelling approach based on 

stochastic methods is introduced, considering different market parameters, such 

as price expectations on the spot or reserve power markets. The electricity 

spot prices are simulated with appropriate time-series models and a regimeswitching 

approach. Based on these simulations a specific energy technology 

investment is analysed. A main outcome is that the regime-switching approach 

delivers well-fitting price paths for the investment evaluation. 

� TD-34 


8.2.23 

Generalized Convexity II 



Chair: Zsolt Pales, Institute of Mathematics, University of Debrecen, 

Egyetem tér 1, 4032, Debrecen, Hungary, pales@math.klte.hu 

1 - Geometric approach to approximating the solution of a 

class of generalized convex optimization problems 

Gabriela Cristescu, Department of Mathematics and Computer 

Science, Aurel Vlaicu University of Arad, Aurel Vlaicu 

University of Arad, Department of Mathematics and Computer 

Science„ Str. Revolutiei, No. 77„ 310130, Arad, Arad, Romania, 

gcristescu@inext.ro, Mihail Găianu 

The classes of g-convex sets and g-convex functions are described from the 

shape point of view. The g-convexity is a particular type of invexity, which 

brings useful properties in connection with approaching the vectorial optimization 

problems with g-convex objective functions on g-convex domain. A procedure 

of geometrical representation of the objective function by maps leads to 

a sequential approximation of the solution. A method of estimating the error 

is given. The paper is supported by the Education and Research Ministry of 

Romania, within the research project ID-1239 PN II 2007 CNCSIS. 

2 - Sufficient efficiency conditions for multiobjective fractional 

programming with generalized type-I-univexity 

Andreea Madalina Stancu, Institute of Mathematical Statistics 

and Applied Mathematics, The Romanian Academy, Calea 13 

Septembrie, nr 13, RO-050711, Bucharest, 

andreea_madalina_s@yahoo.com, Ioan Stancu-Minasian 

In the last time, important results in multiobjective programming involving 

type-I functions were obtained. In this paper, we consider the n-set functions 

multiobjective fractional programming problem (Problem (P)). The problem 

is to find the collection of (properly) efficient sets. We present a few global 

semiparametric sufficient efficiency conditions under various generalized type- 

I univexity hypotheses for Problem (P). 

3 - New method of Homotopy Approximation for Solving 

Linear and Nonlinear problems 

186 

Mohamad Hosein Kafash, Mechanical Engineering, Islamic 

Azad University, No.23-12 daneshjoo-daneshjoo Blv.-vakilabad 

Blv., No.23-12 daneshjoo-daneshjoo Blv.-vakilabad Blv., 

+98511, Mashhad, Iran, Islamic Republic Of, 

m_h_k2001@yahoo.com 

The homotopy perturbation method is a powerful devise for solving functional 

equations. This method was introduced by He in the year 1998. In this method, 

the solution is considered as the summation of an infinite series that converges 

rapidly. This technique is used for many other subjects such as optimization of 

linear and nonlinear systems. nonlinear oscillators with discontinuities, nonlinear 

wave equations, boundary value problems. In this article, an analytic 

approximation to the solution of Blasius equation is obtained by using a new 

modification of homotopy perturbation method. The Blasius equation is a nonlinear 

ordinary differential equation which arises in the boundary layer flow. 

The comparison with Howarth’s numerical solution shows that the new homotopy 

perturbation method is an effective mathematical method with high 

accuracy. 

� TD-35 


6.2.46 

Facilitated Discrete-Event Simulation 



Chair: Stewart Robinson, University of Warwick, Warwick Business 

School, Coventry, United Kingdom, 

stewart.robinson@warwick.ac.uk 

1 - Challenges for Discrete-Event Simulation as a Facilitative 

Modelling Tool 

Stewart Robinson, University of Warwick, Warwick Business 

School, Coventry, United Kingdom, 

stewart.robinson@warwick.ac.uk 

Modelling studies involving discrete-event simulation (DES) typically require 

detailed models and associated with them protracted periods of model building, 

data collection, data analysis and experimentation. Much of this work is 

carried out in the back office by expert modellers with occasional client interaction. 

This presents a challenge to using DES for facilitative modelling which by 

nature requires rapid turnaround and involvement of the clients in group modelling 

sessions. This paper explores these challenges for DES as a facilitative 

modelling tool and suggests some ways in which it might be possible to use 

DES for this purpose. The ideas will be illustrated by two example projects. 

2 - A participative modelling framework for health care 

simulation studies 

Antuela Tako, Warwick Business School, University of Warwick, 

CV4 7AL, Coventry, United Kingdom, antuela.tako@wbs.ac.uk, 

Kathy Kotiadis, Christos Vasilakis 

This study sets out a framework and accompanying tools for developing models 

and conducting health care simulation studies with the active involvement 

of the project stakeholders. So far discrete-event simulation studies in health 

care have not engaged stakeholders in a structured and participative mode in the 

development of models. In this paper we propose a framework called PartiSim 

that aims to engage the stakeholders of simulation studies by using a set of 

tools as part of a series of facilitated workshops. The PartiSim framework is illustrated 

by an example drawn from our experience in conducting a simulation 

study of a bariatric care system with healthcare professionals. 

3 - Participative simulation as a pragmatic tool to facilitate 

problem solving 

Tillal Eldabi, Brunel University, UB8 3PH, Uxbridge, United 

Kingdom, Tillal.Eldabi@brunel.ac.uk 

The traditional approach to modelling and simulation modelling has been criticised 

as too rigid and irrelevant to the complexity and fluctuating dynamism 

of business and public sectors. In this presentation we explore the use of modelling 

and simulation as a facilitative tool to support problem understanding in 

a more pragmatic way with less emphasis on the scientific rigour of unnecessary 

outputs. We propose a modelling approach that sheds more light on the 

intangible outcomes of the modelling process. A number of examples are presented 

to examine the proposed approach and lay out some opportunities for 

further work.

� TD-36 


3.1.05 

Fuzzy Decision Making 




Chair: Miroslav Hudec, Information Systems and Applications, 

Infostat, Dubravska cesta 3, 845 24, Bratislava, Slovakia, 

hudec@infostat.sk 

1 - Paraconsistent Annotated Evidencial Logic Et and Decision 

Making Method in Fuzzy Set Theory 

Fábio Romeu de Carvalho, Paulista University, Rua Dr Bacelar 

1212, 04026-002 , São Paulo, SP, Brazil, fabioromeu@unip.br, 

Jair Minoro Abe 

Decision making is a matter of attention by experts since long time ago. Fuzzy 

set theory has provides a satisfactory application mainly in control systems, but 

it lacks some details mainly involving conflicting situations. On the other hand, 

Paraconsistent logic supplies the concept of inconsistency in a non-trivial way. 

This work presents a first attempt to cope ideas developed with paraconsistent 

annotated logic Et and apply them to a decision making technique for Fuzzy 

set theory. 

2 - Introducing and Managing Uncertainty in Performance 

Measurement Systems 

Sergio Sousa, Production and Systems, University of Minho, 

Campus de Gualtar, 4710-057, Braga, Portugal, 

sds@dps.uminho.pt, Eusebio Nunes 

Many Performance Measurement Systems (PMSs) are available nowadays. 

Some Performance Measures (PMs) have intrinsic uncertainty, such as customer 

satisfaction that is usually expressed through natural language. Most 

PMSs use ’crisp’ numbers that have limitations to deal with such uncertainty. 

The authors, inspired by the ’measurement system analysis’, present an alternative 

approach based on the definition of a scale of uncertainty and its propagation 

through the PMS. The testing of such method is currently being performed 

to ascertain its feasibility, usability and utility. 

3 - Decision Makers Weighting in Fuzzy Multiple Attributes 

Group Decision Making 

Mohsen Rostamy-Malkhalifeh, Mathematics, Science and 

Research Branch, Islamic Azad University(IAU), Hesarak, 

poonak, 1477893855, Tehran, Iran, Islamic Republic Of, 

rostamy@srbiau.ac.ir, Fateme Ghaemi Nasab, Razieh Mehrjoo 

In this paper we introduce a method for modifying data in Fuzzy multiple attributes 

group decision making problems with respect to the weight (reliability) 

of decision makers based on trapezoidal interval type-2 Fuzzy sets. A modified 

decision making method is introduced by combining our proposed algorithm 

with the method which was proposed by Shyi-Ming Chen. 

4 - A fuzzy model for single-period inventory problem with 

pre-season extension 

Hülya Behret, Industrial Engineering Dept., Istanbul Technical 

University, I.T.U. Faculty of Management, Industrial Engineering 

Dept., Macka., Istanbul, Turkey, hulbeh@yahoo.com 

In this study a fuzzy model for an extended single-period inventory problem is 

proposed. The problem is extended with multiple pre-season periods. As the 

selling season becomes closer, the production cost increases and demand fuzziness 

decreases. The demand is believed to be a discrete normal fuzzy number, 

the holding and shortage cost parameters are considered as imprecise and represented 

by triangular fuzzy numbers. The optimum order period and optimum 

order quantity minimize the fuzzy total cost. The model is experimented with 

an illustrative example and supported by sensitivity analyses. 


� TD-37 


3.1.09 

Applications in Decision Making & Decision 

Analysis 

Stream: Decision Support Systems 


Chair: Pascale Zaraté, Institut de Recherche en Informatique de 

Toulouse, Toulouse University, 118 route de NarBonne, 31062, 

Toulouse, France, zarate@irit.fr 

Chair: Fatima Dargam, SimTech Simulation Technology, Ries Strasse 

120, 8010, Graz, Austria, F.Dargam@SimTechnology.com 

1 - Survey forest fire detection systems via multicriteria 

disaggregation — aggregation approach: the case of 

Lycabettus Hill 

Athanasios Spyridakos, Mathematics, TEI of Piraeus, P. Ralli 

and Thivon 250, 12244, Aigaleo, Athens, Greece, 

tspyr@teipir.gr, Yannis Psaromiligkos, Lazaros Vryzidis, Maria 

Litsardaki, Ioannis Salmon 

Allocation of fire detection units in forests is influenced by a set of conflicting 

parameters representing effectiveness (degree of screening, degree of plating), 

cost of installing and maintenance, reliability and security. This research work 

proposes a multictriteria methodological frame based on disaggregation — aggregation 

UTA II method and MIIDAS system, in order to support the selection 

of positions and the arrangement of the fire detection units in the forest. The 

methodological frame was utilized for the design of Lycabettus Hill fire detection 

system. 

2 - The PROBE Decision Support System and Capital Budgeting: 

an application to the electricity sector 

João Carlos Lourenço, Centre for Management Studies of IST 

(CEG-IST), Instituto Superior Técnico / Technical University of 

Lisbon, Av. Rovisco Pais, 1049-001, Lisbon, Portugal, 

joao.lourenco@ist.utl.pt, Carlos Bana e Costa, João Oliveira 

Soares 

Every year EDPD, the main electricity distribution company in Portugal, selects 

a portfolio of projects to expand and modernize its electricity network 

infrastructures. The portfolio must comply with budget and dependence constraints 

concerning different types of projects and organizational units. We 

present how the decision support system PROBE (Portfolio Robustness Evaluation) 

may be used to help the portfolio decision analysis process when in 

presence of uncertainty in the benefits and in the costs of the projects. 

3 - Search Analysis to Locate a Sustainable Data Center in 

Portugal using Geographic Information Systems 

Luis C. Dias, Faculdade de Economia / INESC Coimbra, 

University of Coimbra, Av Dias da Silva 165, 3004-512, 

Coimbra, Portugal, lmcdias@fe.uc.pt, Miguel Covas, Carlos 

Santos Silva 

Data Centers (DC) are buildings constructed primarily to house, power and 

cool computer equipment, consuming as much energy as 25.000 households 

on average. The increasing demand for computer resources has led to a growth 

of DC’s over the world along with the increase of energy used by these servers 

and the power and cooling infrastructure to support them. A search analysis 

to locate a sustainable data center in Portugal is performed using Geographic 

Information Systems, explicitly taking into account multiple criteria. 

4 - How does the EURO Working Group on DSS interact? 

A social-academic network analysis 1989-2008 

Fatima Dargam, SimTech Simulation Technology, Ries Strasse 

120, 8010, Graz, Austria, F.Dargam@SimTechnology.com, Rita 

Ribeiro, Pascale Zaraté 

This work reports the advances on the analysis of the social-academic network, 

created for the EURO Working Group on Decision Support Systems 

(EWG-DSS), representing the various relationships that link academically its 

105 members; as well as its collaboration dynamics’ evolution since 1989 up 

to 2008. The EWG-DSS social-academic network has already shown to encourage 

new research and promote further collaboration among its members in 

common projects and joint-publications. In this paper further network analysis 

is presented, to better explain some concluding factors about the group. 

187


� TD-38 


6.2.44 

Stochastic Valuation of Energy Prices and 

Derivatives 



Chair: Nadia Maïzi, Center for Applied mathematics, MINES 

ParisTech, Rue C. Daunesse, BP 207, 06904, Sophia-Antipolis, 

France, nadia.maizi@mines-paristech.fr 

Chair: Pourtallier Odile, Sophia Antipolis, INRIA, 2004 rte des 

lucioles, 06902, Sophia Antipolis, odile.pourtallier@sophia.inria.fr 

Chair: Martin Rainer, Inst. Applied Mathematics, METU Ankara, 

ENAMEC Inst., Glockengasse 15, D-97070 Würzburg, 97070, 

Würzburg, Germany, martin.rainer@enamec.de 

1 - Robust pricing in electricity markets with a variable demand 

Eugene Zak, Areva T&D Inc., 10865 Willows Road NE, 98052, 

Redmond, WA, United States, eugene.zak@areva-td.com, Kwok 

Cheung, Ricardo Rios-Zalapa 

Clearing prices in a linear dispatch problem come from a dual solution. The 

situation is getting more complicated when a power demand is not constant but 

depends on current prices. Following Bulavskii [Soviet Math. Dokl., Vol 23 

(1981), No. 2] a corresponding primal-dual model incorporates both primal 

and dual variables. Under a linearity assumption for the demand-price function 

with a semi-definite matrix a resultant quadratic program yields stable prices 

and optimal levels of the demand. 

2 - Determining market clearing prices for coupled dayahead 

energy markets 

Johannes Müller, Discrete Optimization, TU Darmstadt, 

Dolivostr 15, 64293, Darmstadt, Germany, 

jmueller@mathematik.tu-darmstadt.de, Alexander Martin, 

Sebastian Pokutta 

The European power grid can be divided into several market areas where the 

prices of electricity are determined in a day-ahead auction. These market areas 

are connected by power lines whose transmission capacity is restricted. Hence 

supply and demand of the different areas can not be balanced in all cases. The 

goal of the auction is to determine prices and cross border flow, so that the social 

welfare gets maximized and the electricity is transmitted in general from a 

low price area to an high price area. Therefore a MIQP has to be solved. 

3 - Indifference price for emission permits 

Olivier Davidau, Center for Applied Mathematics, Mines 

Paristech, Rue Claude Daunesse - B.P. 207, 06904, Sophia 

Antipolis Cedex, France, olivier.davidau@cma.ensmp.fr, 

Mireille Bossy, Nadia Maïzi, Pourtallier Odile 

Agents on the emission markets have to assess their subjective value for emission 

permits. This value depends on expected emissions and thus on production. 

We focus on the case of an electricity producer and use a stochastic control 

framework, taking into account the electicity market uncertainties. Solving this 

stochastic control problem gives the price the agent is ready to pay for emission 

permits. This so called indifference price gives information on opportunity to 

reduce emission or invest in clean technology. 

� TD-39 


6.2.45 

Optimal Control: Recent Advances II 



Chair: Gustav Feichtinger, Institute for Mathematical Methods in 

Economics, Vienna University of Technology, Argentinierstr. 

8/105-4, 1040, Vienna, Austria, gustav@eos.tuwien.ac.at 

188 

1 - Local ramp metering control based on a coupling of 

continuum traffic flow models with reinforcement learning 

Salissou Moutari, Applied Mathematics and Theoretical Physics, 

Queen’s University Belfast, David Bates Building, Queen’s 

University Belfast, University Road, BT7 1NN , Belfast, United 

Kingdom, s.moutari@qub.ac.uk, Barry McCollum 

In order to increase the operational efficiency of transportation systems, nowadays, 

freeway networks are equipped with ramp metering systems which aim to 

regulate, in an appropiate way, the inflow from on-ramps to main carriageways. 

The effectiveness of a ramp metering system in improving traffic conditions, 

e.g. by minimizing traffic congestion, depends highly upon the efficiency of its 

operational strategy. In this work, we introduce a new strategy of addressing the 

issue of coordinating local ramp metering actions by coupling the continuum 

traffic flow model with reinforcement learning. 

2 - Optimal Control for Spacecraft Orbital Transfer Maneuver 

M. Navabi, Shahid Beheshti University, Iran, Islamic Republic 

Of, civil.space.edu@gmail.com, M. Sanatifar 

The spacecraft orbital transfer is significant in view of time or energy optimality. 

In this paper, the optimal control for spacecraft orbital transfer using 

impulsive thrust is studied. Impulsive maneuvers could be categorized based 

on the geometry of initial and final orbits, number of impulses, etc.. Our paper 

is a comprehensive investigation of existing approaches and also to propose a 

numerical method based on minimization of total delta V for problem of optimal 

transfer maneuver. Some numerical examples are presented to demonstrate 

the accuracy of the method. 

3 - A new Branch and Bound based method dedicated to 

minimize the energy consumption of an electric vehicle. 

Merakeb Abdelkader, IRIT-ENSEEIHT, 2, rue Camichel, 31000, 

toulouse, France, abdelkader.merakeb@n7.fr, Frederic Messine 

This work consists to elaborate a strategy of control in fixed time of an electric 

vehicle which must perform a given displacement. Moreover, a constraint 

about a state variable has to be taken into account. The minimization of the 

energy consumption with a fixed time makes it possible to use a mesh of time 

and control. According to assumptions about the terminal constraints imposed 

on the trajectories, an algorithm based on Branch and Bound techniques is proposed. 

This one is dedicated to solve the studied problem but this approach is 

sufficiently general for some extensions. 

� TD-40 


6.2.52 

Port Simulation and Optimization 

Stream: Container Terminal Operations 


Chair: Yue Wu, School of Management, University of Southampton, 

University Road„ SO17 1BJ, Southampton, United Kingdom, 

Y.Wu@soton.ac.uk 

Chair: Jing-An Li, MADIS, Institute of Systems Science, Academy 

of Mathematics and Systems Science, No. 55, Zhongguancun East 

Road, Haidian District, 100190, Beijing, China, ajli@amss.ac.cn 

1 - Modeling of vehicle scheduling and storage problems 

for import containers 

Jiabin Luo, School of Management, University of Southampton, 

School of Management, University of Southampton, University 

Road, Highfield, SO17 1BJ, Southampton, United Kingdom, 

j.luo@soton.ac.uk, Yue Wu, Dali Zhang 

This paper discusses optimization problems for unloading import containers. A 

mixed integer programming model is presented for scheduling different types 

of container handling equipments and selecting storage locations in an integrated 

way. The objective is to minimize the container handling time at quay 

side and yard side as well as the transportation time by vehicles. Several constraints 

are considered as well, such as the block size, the sequence of vehicles 

and the handling time of cranes. A series of numerical experiments has been 

carried out to test the efficiency of the proposed model.

2 - Optimal pricing strategies for empty container management 

Jing-An Li, MADIS, Institute of Systems Science, Academy of 

Mathematics and Systems Science, No. 55, Zhongguancun East 

Road, Haidian District, 100190, Beijing, China, ajli@amss.ac.cn 

One assumption in most research about empty container management is that, 

the demand should be satisfied by companies’ owned containers or leased containers, 

where empty containers are always available for the lease. However, it 

is not true in the real world. And companies always face the shortage of empty 

containers to such an extent that more and more excess demand are delayed. 

In this paper, we analyze the pricing strategies for empty container management. 

Especially, we analyze the strategy ‘high price for short delivery time 

while low price for long delivery time’. It is shown that, this strategy can partially 

relax the tension of the limited empty containers. On the basis of ensuring 

the regular profit, this strategy also provides one flexible mechanism for empty 

container management. 

3 - Simulation based evaluation of container terminal yard 

layouts 

Jörg Wiese, DS&OR Lab, University of Paderborn, Warburgerstr. 

100, 33098, Paderborn, Germany, wiesej@upb.de, Leena Suhl, 

Natalia Kliewer 

The structure of a container terminal yard layout is significantly influenced 

by the equipment type used for stacking operations. Based on these equipment 

types different yard layout categories can be distinguished, such as parallel rubber 

tired gantry crane based layouts or perpendicular rail mounted gantry crane 

based layouts. We present a simulation study in which we evaluate if these yard 

layout categories (respectively different yard layouts) are adequate for several 

possible scenarios, e.g. for scenarios with varying ratios of transshipment and 

import/export containers. 

4 - Generic Port Operations Simulator 

Rui Carlos Botter, Logistics Systems, University of São Paulo, 

Av. Prof. Mello Moraes, no. 2231, 05356000, São Paulo, SP, 

Brazil, rcbotter@usp.br, Afonso Celso Medina 

The research presents a generic simulator for iron ore port operations that allow 

the analysis of current and future operations of bulk terminals, aiming identification 

of bottlenecks, operational alternatives and needs for infrastructure, as 

well as estimating the equipments required annual productivity rates, the storage 

capacity requirements of all different products handled, according to the 

operational restrictions; evaluation of the nominal capacity and operational efficiency 

of the handling equipments and the selection of the export plans in the 

short, medium and long time horizon. 

� TD-41 


3.1.06 


III 



Chair: Gisele Bosso de Freitas, Department of Physics, Sao Paulo 

State University, Brazil, freitas_gb@yahoo.com.br 

1 - Diffusion equation, growth and diagnosis of tumor 

agressiveness 

Gisele Bosso de Freitas, Department of Physics, Sao Paulo State 

University, Brazil, freitas_gb@yahoo.com.br, Elso Drigo Filho 

In this work we study a model for growth of tumor cells based on a diffusion 

equation, which uses a 2D model on the growth of cells "in vitro". The solutions 

obtained can be compared to the ways that tumor cells take during growth, 

which may indicate tumor aggressiveness. Thus, comparing the solutions obtained 

through the model to the tumor forms "in vitro" could be inferred, the 

solutions obtained through the model, which more aggressive tumors. We suggest 

that this identification is related to the number of basis functions required 

to simulate the contour of the tumor. 

2 - Optimization of resources for a behavior controlled of 

Petri nets with multipliers in dioid algebra 

Samir Hamaci, Productique, EPMI, 13, boulvard de l’Hautil, 

95092, Cergy, France, S.HAMACI@EPMI.FR, Rahim Benfkir 


We are particularly interested to the problem of allocating an initial marking in 

a Timed Event Graphs with Multipliers for a desired cycle time. For this, for 

define the marking of some places, we proceed by linearization of the mathematical 

model reflecting the behavior of a TEGM in order to obtain a (min, 

+) linear model. From the latter, we determine the marking which satisfies the 

desired cycle time. 

3 - Collective animal manure management simulation and 

environmental impact assessment 

Francois Guerrin, CA, Inra/Cirad, Station de la Bretagne - BP 

20, 97408, Saint-Denis, France, francois.guerrin@cirad.fr 

This contribution deals with simulation modeling to help manage agricultural 

production systems. It describes the use of a Systems Dynamics complex 

model to simulating the functioning of pig slurry collective spreading plans in 

Brittany (Northwestern France), where intensive livestock farming has a wellknown 

harmful environmental impact, namely on groundwater and coastal waters. 

The model dynamically simulates the slurry stock evolutions at pig farms 

and the spreading fluxes on crops, both at the pig farms and the remote crop 

farms. 

� TD-42 


3.1.07 

Decison Making 1 

Stream: Decision Making 


Chair: Alan French, The Business School, Loughborough University, 

Ashby Road, LE11 3UT, Loughborough, United Kingdom, 

A.P.French@lboro.ac.uk 

1 - Tail Restricted Stochastic Dominance (TRSD): A Class 

for Quasi Stochastic Dominance 

Edgar Elias Osuna, Centro de Finanzas, Instituto de Estudios 

Superiores de Administracion (IESA), Av. IESA, San 

Bernardino, 1010, Caracas, DC, Venezuela, eosuna@iesa.edu.ve 

We introduce a class for quasi stochastic dominance between probability distributions. 

It will be concerned with a dominance restricted to a range of values 

of the random variables which disregards very unlikely outcomes; specifically 

those in the tail(s) of the distributions. We have named the class Tail Restricted 

Stochastic Dominance (TRSD) to differentiate it from two similar concepts: 

the almost stochastic dominance for investment decisions (ASD) introduced 

by Leshno and Levy (2002) and the restricted stochastic dominance used by 

Davidson and Duclos (2000) for poverty studies. We illustrate the concept for 

the case of two empirical distributions and apply it to decision making under 

risk. 

2 - The effects of emotions on decision making 

Mohammad Emambocus, Business, London South Bank 

University, 103 Borough Road, SE1 0AA, London, United 

Kingdom, emambocm@lsbu.ac.uk 

Decision making occurs in the light of uncertainty between different alternatives 

depending on the expected outcomes. These rational decisions are taken 

through the application of conditional probability as elaborated by Bayes’ Theorem. 

However, there are lots of evidences that have proven that investors 

violate the Bayes rules and this is mainly due to the effects of emotions which 

affects the rational decision-making process. This paper investigates at this 

issue and analyses whether emotions are harmful to decision making or is it 

enhancing the process. 

3 - Measuring Changes in Brand Knowledge/Perception 

Using Brand Concept Mapping 

Alan French, The Business School, Loughborough University, 

Ashby Road, LE11 3UT, Loughborough, United Kingdom, 

A.P.French@lboro.ac.uk, Gareth Smith 

The brand is an associative network of interconnected information about an 

object, held in memory and accessible when stimulated from the memory of 

a consumer. This network is dynamic and subject to change. We show how 

Brand Concept Mapping can be used to capture individual maps of consumers 

perception of brands at a point in time, which are then aggregated to produce 

one consensus map. In this way, changes in perception of a brand can be tracked 

over time. We treat the leaders of the main political parties in the UK as brands 

and illustrate the approach in the run-up to the general election. 

189


� TD-43 


8.2.02 

Global Optimization 2 



Chair: Mirjam Duer, Johann Bernoulli Institute of Mathematics and 

Computer Science, Rijksuniversiteit Groningen, P.O. Box 407, 9700 

AK, Groningen, Netherlands, M.E.Dur@rug.nl 

1 - An approach to constrained global optimization based 

on exact penalty functions 

Francesco Rinaldi, Dipartimento Informatica e Sistemistica, 

Sapienza, via Ariosto, 25 Roma, 00185, Rome, Italy, 

rinaldi@dis.uniroma1.it, Gianni Di Pillo, Stefano Lucidi 

In the field of global optimization many efforts have been devoted to solve unconstrained 

global optimization problems. The aim of this paper is to show 

that unconstrained global optimization methods can be used also for solving 

constrained optimization problems, by resorting to an exact penalty approach. 

In particular, we make use of a non-differentiable exact penalty function. We 

show that, under weak assumptions, there exists a threshold value of the penalty 

parameter such that any global minimizer of the penalty function is a global solution 

of the related constrained problem and conversely. On these bases, we 

describe an algorithm that, by combining an unconstrained global minimization 

technique for minimizing the penalty function and an automatic updating 

of the penalty parameter that occurs only a finite number of times, produces 

a sequence of solutions such that any limit point of the sequence is a global 

solution of the related constrained problem. In the algorithm any efficient unconstrained 

global minimization technique can be used. In particular, we adopt 

an improved version of the DIRECT algorithm. Some numerical experimentation 

confirms the effectiveness of the approach. 

2 - Test-Problem Generator for Unconstrained Global Optimization 

Chi-Kong Ng, Systems Engineering & Engineering 

Management, Chinese University of Hong Kong, Shatin, N.T, 

Hong Kong, ckng@se.cuhk.edu.hk, Duan Li 

A novel test-problem generator for benchmarking unconstrained GO algorithms 

is discussed. By combining n sophisticated univariate problems and 

applying linear transformation of variables, a class of inseparable test-problems 

with 2n local minima is obtained. The generator is realized by a set of MAT- 

LAB programs, and is capable of generating test-problems in MATLAB and 

GAMS. A standard set of test problems is produced. Computational experiments 

have demonstrated the stability of the generating process and the controllability 

of assigning the difficulty level to the test problems. 

3 - Certificates for copositive programming 

Immanuel Bomze, ISDS, University of Vienna, Bruenner Str. 72, 

A-1210, Vienna, Austria, immanuel.bomze@univie.ac.at 

Many global and some combinatorial optimization problems have a copositive 

reformulation, shifting complexity entirely to the question whether a given matrix 

is copositive, i.e. generates a quadratic form taking no negative values over 

the positive orthant. Either an affirmative answer or a negative certificate (a 

violating vector) is required. The dual involves completely positive matrices, 

so-called because they are symmetric-factorizable with a nonnegative rectangular 

matrix. Knowledge of this factor gives the explicit solution of the given 

problem, which is a positive certificate. 

4 - Some new results on Copositive Programming 

Mirjam Duer, Johann Bernoulli Institute of Mathematics and 

Computer Science, Rijksuniversiteit Groningen, P.O. Box 407, 

9700 AK, Groningen, Netherlands, M.E.Dur@rug.nl 

We present some new results on copositive programming. Copositive programs 

are optimization problems over the cone of copositive matrices, or its dual cone 

of completely positive matrices. Copositive programs are of interest because 

they model both quadratic and binary problems and hence provide a unified 

framework for certain mixed-integer nonlinear problems. 

190 

� TD-44 


8.2.03 

Business Process Modelling and Simulation 



Chair: Miro Gradisar, Univ. of Ljubljana, Faculty for economy, 

Kardeljeva ploscad 17, 1000, Ljubljana, Slovenia, 

miro.gradisar@ef.uni-lj.si 

1 - Probability of default simulation: managing predictive 

analytics processes in a multi-model system 

Igor Perko, Faculty of Economics and Business, Razlagova 14, 

2000, Maribor, Slovenia, igor.perko@uni-mb.si, Miro Gradisar, 

Samo Bobek 

For the probability of default simulation to be successful it needs to include all 

customers in the credit portfolio and continually deliver accurate results. To 

resolve these issues we propose a multi-model evaluation system using transparent 

selection logic and an active predictive analytics (PA) processes management. 

The multi-model system knowledge is stored in reasoning capable 

knowledge management structures while the involved PA processes are actively 

managed using a multi agent system. The proposed solution is applied in a scenario, 

based on a commercial bank customer portfolio data. 

2 - Business process simulation for solving large cutting 

problems 

Mihael Cesar, Business Informatics, Faculty of Economics, 

Kardeljeva pl. 17, 1000, Ljubljana, mcesar@gmail.com, Jure 

Erjavec, Miro Gradisar 

Reasons for solving large one-dimensional cutting stock problems (CSPs) in 

the context of business process as a whole are outlined. The solution considers 

characteristics of different units (logistic, production). Large CSPs are solved 

through optimization on two levels. Individual order lengths are cut from the 

best suiting stock lengths. The whole order is divided into groups consisting of 

certain amount of different order lengths, which can appear only in one group. 

Groups enable better handling of material during the logistic process and make 

entire business process more effective. 

3 - Cutting stock problem as a part of a business process 

Jure Erjavec, Informatics, Faculty of Economics, Kardeljeva 

plosèad 17, 1000, Ljubljana, jure.erjavec@ef.uni-lj.si, Miro 

Gradisar 

Renovation of business processes leads to shorter lead times and lower inventory 

levels. The size of inventory is one of the key factors when addressing 

the cutting stock problem. Low inventory size with regard to order size means 

less cutting possibilities which can lead to higher trim loss. Therefore the company 

has to optimize its inventory size when seeking balance between inventory 

costs and costs of trim loss. A simulation model for determining the optimal 

inventory size is developed. 

4 - Cash flow modeling and optimization in road network 

building 

Marko Šetinc, Geopolis, d.o.o., 1000, Ljubljana, 

marko.setinc@guest.arnes.si, Miro Gradisar 

An efficient system was made that allows the cash flow simulation and optimization 

in the road building process. In the simulation a model for the calculation 

of cash flows was created, which is based on the National program for 

the construction of motorways in Slovenia. The optimization of cash flows was 

carried out with a genetic algorithm. It is presented, analyzed, and used several 

criteria for optimization as maximization of net present value and minimization 

of time deviation from the original plan. The obtained result is an optimal value 

of selected financial and time parameters.

� TD-45 


8.2.12 

OR in Military II 

Stream: OR in Military 


Chair: Ana Isabel Barros, Information & Operations, TNO - Defense, 


Netherlands, ana.barros@tno.nl 

1 - A Robust Approach to the Missile Defence Location 

Problem 

Lanah Evers, Defence, Security and Safety, TNO, P.O. Box 

96864, 2509 JG, The Hague, Netherlands, lanah.evers@tno.nl, 

Axel Bloemen, Ana Isabel Barros, Herman Monsuur, Albert 

Wagelmans 

This paper proposes a model for determining a robust defence strategy against 

ballistic missile threat. Two problem variants are formulated. In the first one, 

the number of ballistic missile interceptor systems is minimised, such that a 

predetermined defence level is achieved. In the second variant, the defence 

level is maximised for a given number of available interceptor systems. To 

solve these problems we applied both a heuristic and an exact solution method. 

We used a fictive, but realistic data set to illustrate the differences between both 

variants and their use in practice. 

2 - Development of a Joint Picture for Domestic Domain 

Awareness 

Mark Gammon, DRDC Atlantic, Defence R&D Canada, 9 Grove 

Street, PO Box 1012, B2Y 3Z7, Dartmouth, Nova Scotia, 

Canada, mark.gammon@drdc-rddc.gc.ca 

Canada Command requires a Joint Picture that can combine maritime, air, and 

land pictures on one system to enhance their Domestic Domain Awareness in 

their Area of Responsibility. The development of a Joint Picture is a stated objective 

of Canada Command that requires more definition in order to progress 

from the conceptual stage to a Joint Picture that is relevant in the context of 

Domestic Domain Awareness. This document proposes an approach based on 

examination of Canada Command exercises that will produce requirements for 

a Joint Picture. These requirements should be consistent with and relevant to 

the Canada Command mission and evolve over the life of the project to accommodate 

changes in the CF/DND. This approach will allow the Joint Picture to 

progress from the conceptual stage to a Joint Picture that is relevant for Domestic 

Domain Awareness. The initial analysis performed and reported here shows 

that there is sufficient information available to support the proposed methodology. 

3 - UAV’s MISSION OPERATION SIMULATION 

Dusan Starcevic, Faculty of Organizational Sciences, University 

of Belgrade, Jove Ilica 154, 11000, Belgrade, Serbia, 

starcev@fon.rs 

Unmanned Aerial Vehicles (UAV) system require complex hardware and software 

components and highly time constrained coordination between these components. 

The software component consists of three grouped concurrent tasks 

(hard real-time tasks, soft real-time tasks and non-real-time tasks). Given the 

high cost of fielding physical UAVs, computer simulation has been used extensively 

to test the solutions that incorporate aforementioned challenges for 

UAVs. In this paper we present a simulation environment for visualizing, controlling 

and simulating UAVs data acquisition operations. 

4 - A search game with incomplete information on a 

searcher’s detection capability 

Hideki Higuchi, Dep. of Computer Science, National Defense 

Academy, Japan, g48037@nda.ac.jp, Ryusuke Hohzaki, Toru 

Komiya, Emiko Fukuda 

This presentation deals with a search game, in which a stationary hider selects 

a cell in a discrete search space to hide himself/herself, and a searcher that distributes 

his/hers search resouces in the space to detect the hider. The searcher’s 

detection capability usually depends on the searcher’s or the hider’s property 

which is unknown to the other. We model two search games with asymmertric 

information in which the seacher or the hider has private information affecting 

the detection capability, and show how their private information change their 

optimal strategies and the value of the game. 


� TD-46 


8.2.14 

Agent-based Modeling II 

Stream: Agent-Based Modeling [c] 





1 - An agent-based electricity demand model for the distribution 

grid. Case study PREMIO-Capenergies project 

Enrique Kremers, EIFER, Universität Karlsruhe, 

Emmy-Noether-Strasse 11, 76131, Karlsruhe, Germany, 

kremers@eifer.org, Alvaro Gómez Ibañez, Pablo Viejo, Sophie 

Chartres 

To analyze demand at substation level, static, top-down based methods (load 

profiles or statistical analysis) are the state-of-art. The model presented here 

has been developed in the PREMIO-Capenergies project (dynamic load management 

at local scale). A three level, bottom-up approach is presented. 

Agents representing customers at the lowest level generate the demand based 

on stochastic models. Groups of customers are modelled by different agent 

classes. Transformers represent the aggregation points for the intermediate 

level and the global aggregation point stands for the substation. 

2 - Stable Distributions in Agent-based Model and Asset 

Liability Management 

Vadym Omelchenko, Faculty of Mathematics and Physics, 

Charles University in Prague and Institute of Information Theory 

and Automation, Ke Karlovu 3, 121 16 Praha 2, Prague, Czech 

Republic, so8859@mail.ru, Vladimir Gurzhiy 

Stable distributions can serve as the best candidate to replace the normal distribution 

due to the fact that they possess most of the properties of the normal 

distributions representing a flexible family cabable of adjusting to many constraints 

and fitting larger classes of data. In my research I am going to use 

them for modeling asset prices for the agent-based model and for Asset Liability 

management. The latter will be conducted in cooperation with Vladimir 

Gurzhiy. 

3 - Supply Chains and Multiple Marginalization: A Study 

Using Multi-Agent Simulation 

Jan Zouhar, Department of Econometrics, University of 

Economics, nám. W. Churchilla 4, Prague 3, 13067, 12000, 

Prague 2, zouharj@vse.cz 

As described by multiple marginalization models, with a lack of cooperation 

the prices within a supply chain tend to spiral up to an inefficient level (equilibrium 

prices) where both the consumer surplus and the SC’s profit are diminished. 

The focus of our recent research was the process of convergence 

of prices inside the SC to their equilibrium levels. The analysis was carried 

out using computer experiments with a multi-agent simulation model of a SC 

with limited information. Two different types of supply chain architecture were 

studied — the serial and the parallel type. 

4 - A generic agent-based framework for cooperative hybrid 

meta-heuristic search 

Simon Martin, Mathematics, University of Portsmouth, Lion 

Building, Lion Gate, PO1 3HE, Portsmouth, Hampshire, United 

Kingdom, simon.martin@port.ac.uk, Djamila Ouelhadj, Patrick 

Beullens, Ender Ozcan 

Different (meta/hyper-)heuristics have strengths and weaknesses. This study 

aims to build a framework where different agents performing different 

(meta/hyper-)heuristics cooperate and investigate its performance over some 

OR problems. We have implemented an island model in which the agents run 

different (meta/hyper-)heuristics. The agents cooperate through asynchronous 

communication by exchanging solutions during the search process. Preliminary 

experiments have been conducted on TSP. We plan to apply this agent 

based cooperative approach to other OR problems. 

191


� TD-47 


8.2.16 

Optimization in Water Systems I 

Stream: OR in Water Management 


Chair: Mathias Höfer, Heidelberg University, 69120, Heidelberg, 

hoefer_m@web.de 

Chair: Halil Önder, Civil Engineering, Middle East Technical 

University, Middle East Technical University, Department of Civil 

Engineering, 06531, Ankara, Turkey, onde@metu.edu.tr 

1 - Optimization of Booster Disinfection Station Scheduling 

Caglayan Sert, Civil Engineering, T.C. Atılım University, 

˙Incek/Ankara, 06836, Ankara, Turkey, caglayansert@gmail.com, 

Ay¸se Burcu Altan Sakarya 

Chlorine is mostly used as disinfectant in water systems. Since disinfectants 

are reactive and decay in the system, in order to satisfy bounds on residual 

concentration required for water quality, chlorine added at source may not be 

enough. Hence booster stations are also needed. The optimization problem is 

formulated as having an objective to minimize the injected mass dosage rate 

subjected to the provision of adequate and more uniform residual concentration. 

A C++ code interfacing with EPANET is developed to find optimum 

scheduling and injection rates of the booster disinfection stations. 

2 - Optimum Design of Slurry Pipelines 

Burhan Yıldız, Civil Engineering, Middle East Technical 

University, Ankara, Turkey, burhan@metu.edu.tr, Ay¸se Burcu 

Altan Sakarya, Metin Ger 

There exist various applications of slurry transportation through pipelines. This 

transpotation problem is solved to determine the pipe diameters and the transported 

slurry amounts from the demand points to the processing points. The 

minimization of the pipe and energy costs is considered as the objective function. 

The optimization method used is Genetic Algorithms. The proposed 

methodology to solve this nonlinear programming problem is applied to a transportation 

system and it is seen that the methodology made the complex, labor 

intensive equation solution process very convenient to use. 

3 - Online Pressure Optimization in Water Distribution Networks 

Frederik Blank, ABB Corporate Research, Germany, 

frederik.blank@de.abb.com, Mathias Höfer, Rüdiger Franke, 

Markus Gauder 

In this presentation we provide an insight into how to improve energy efficiency 

and to reduce the leakage flows in water distribution networks by calculating 

optimal pressure set-points of the main pressure reduction valves and thus to 

minimize the pressure levels and along with it the leakage rate. In a first step, 

we present a methodology to reduce the number of model equations and constraints 

of a large hydraulic nonlinear water distribution network simulation 

model to an online suitable optimization model. In a second step the formulation 

of the NLP problem and results are presented. 

4 - Optimal Management of Coastal Aquifers by Using Genetic 

Algorithm 

Korkut Demirba¸s, Middle East Technical University, Turkey, 

korkut.demirbas@tubitak.gov.tr, Ay¸se Burcu Altan Sakarya, 

Halil Önder 

Overexplotation of groundwater in coastal aquifers results in extraction of salt 

water from wells near the coast. In this work, optimal management of a coastal 

aquifer is studied where groundwater flow is modeled with single potential formulation 

of Strack (1976). Genetic algorithm (GA) is used to optimize the 

pumping scheme. Different seawater prevention methods such as injection 

wells are added to the model to further test the optimal management. 

192 

� TD-48 


8.2.04 

Multi-Objective Optimization 

Stream: Nonlinear Programming [c] 


Chair: Lino Costa, Dept. Production and Systems, University of 

Minho, School of Engineering, Campus de Gualtar, 4710-057, Braga, 

Portugal, lac@dps.uminho.pt 

1 - Nonlinear Multi-Objective Integer Programming: A General 

Approach For Optimization 

Melih Özlen, School of Mathematical and Geospatial Sciences, 

RMIT University, 3000, Melbourne, Australia, 

melih.ozlen@rmit.edu.au, Meral Azizoðlu 

In this study we develop a general approach to solve Nonlinear Multi-Objective 

Integer Programming problems. Our approach is based on identifying the objective 

efficiency and nonlinear optimality ranges. The objective efficiency 

ranges are found by solving Multi-Objective Integer Programming problems 

with fewer objectives. To find the nonlinear objective optimality ranges, we 

use a set of already known nondominated solutions and defined nonlinear objective 

function. 

2 - Multiobjective optimization based on the cone of efficient 

directions and multiobjective golden section algorithm 

Douglas Vieira, ENACOM, Brazil, 

douglas.vieira@enacom.com.br, Adriano Lisboa, Rodney 

Saldanha 

This work introduces a method to solve nonlinear multiobjective optimization 

problems based on the cone of efficient directions and on a multiobjective line 

search. The cone of efficient directions is derived from the KKT conditions. 

Any direction in it guarantees a dominating point (at least in the vicinity of 

the evaluation point). A multiobjective golden section was formulated in to 

find such a point in a finite number of iterations. The resulting method inherits 

an asymptotical convergence to a point with necessary optimality conditions, 

which is a notable property in real world problems. 

3 - Integrating Taylor’s expansion on the lexicographic approach 

for unconstrained optimization 

Christina D. Nikolakakou, Department of Mathematics, Patras 

University, Greece, University Campus, 26504, Rio, Greece, 

christina.nikolakakou@gmail.com, Theodoula N. Grapsa, 

George S. Androulakis 

Recently, we proposed a novel technique for transforming the unconstrained 

optimization problem to an equivalent lexicographic optimization one. This 

was achieved by constructing objective functions using appropriate partition of 

the original objective function. In this paper we propose objective functions, 

via Taylor expansions, to constitute the corresponding partition of the objective 

function. The proposed methodology is tested in optimization problems using 

different order of Taylor expansions with promising results. 

4 - Optimality in multiobjective programming under generalized 

invexity 

Hachem Slimani, Laboratory of Modeling and Optimization of 

Systems LAMOS, Operational Research Department, University 

of Bejaia, 06000, Bejaia, Algeria, haslimani@gmail.com, 

Mohammed Said Radjef 

We introduce new classes of invex and weakly pseudo-invex vector functions, 

where every component is considered with respect to its own function eta. A 

new Kuhn-Tucker type necessary condition is established, for differentiable 

multiobjective problems, without using any constraint qualification and any alternative 

theorem. Sufficient conditions for a feasible point to be weakly or 

properly efficient are obtained under weak invexity assumptions. The obtained 

optimality conditions allow to prove that a feasible point is weakly or properly 

efficient even if it is not a vector Kuhn-Tucker point.

Tuesday, 15:40-17:00 

� TE-01 


Aula Magna 

Plenary Talk 2 



Chair: Silvano Martello, DEIS, University of Bologna, Viale 

Risorgimento 2, 40136, Bologna, Italy, silvano.martello@unibo.it 

1 - A Tale of Three Eras: The Discovery and Rediscovery 

of the Hungarian Method 

Harold W. Kuhn, Department of Mathematics, Princeton 

University, Fine Hall, Washington Road, NJ 08544-1000, 

Princeton, United States, kuhn@math.Princeton.EDU 

In the Fall of 1953, a translation of a paper of Jeno Egervary from Hungarian 

into English combined with a result of Denes Konig provided the basis of 

a good algorithm for the linear assignment problem. To honor the Hungarian 

mathematicians whose ideas had been used, it was called the Hungarian 

Method. In 2005, Francois Ollivier discovered that the posthumous papers of 

Carl G. J. Jacobi contain an algorithm that, when examined carefully, is essentially 

identical to the Hungarian Method. Since Jacobi died in 1851, this 

work was done over a hundred years prior to the publication of the Hungarian 

Method in 1955. This lecture will provide an account of the mathematical, 

academic, social and political worlds of Jacobi, Konig/Egervary, and Kuhn. As 

sharply different as they were (Prussian monarchy, Hungary under the Nazis 

and the Communists, and the post-war USA), they produced the same mathematical 

result. The lecture will be self-contained, assuming little beyond the 

duality theory of linear programming. The Hungarian Method and Jacobi’s 

algorithm will be explained at an elementary level and will be illustrated by 

several examples. 

EURO 24 - Lisbon 2010 TF-03 


� TF-03 


3.2.15 

Location problems 



Chair: Manuel Laguna, Leeds School of Business, University of 

Colorado at Boulder, 80309-0419, Boulder, Colorado, United States, 

laguna@colorado.edu 

Chair: Javier Alcaraz, Dept. Estadística, Matemáticas e Informática, 

Universidad Miguel Hernández de Elche, Av. Universidad s/n, 03202, 

Elche, Alicante, Spain, jalcaraz@umh.es 

1 - Approximation Algorithms for the Competitive Facility 

Location Problem 

Vladimir Beresnev, Operation Research, Sobolev Institute of 

Mathematics, pr.Academica Koptyuga,4, 630090, Novosibirsk, 

Russian Federation, beresnev@math.nsc.ru 

We consider the competitive facility location problem, where two rival firms ( 

Leader and Follower) open facilities sequentially and each client selects one of 

the open facilities according to his preferences. The problem is to find a facility 

location for the Leader which maximized his profit taking into account the 

best answer of the Follower. We formulate model as bilevel integer programming 

problem. The way of construction of an upper bound for optimal values 

of the Leader’s profit is proposed. Together we obtain an initial approximation 

solution of the problem. We present local search algorithms for improving 

the initial solution. Our computation results illustrate the good quality of the 

obtained solutions. 

2 - Solving a location problem with advanced metaheuristics 

Javier Alcaraz, Dept. Estadística, Matemáticas e Informática, 

Universidad Miguel Hernández de Elche, Av. Universidad s/n, 

03202, Elche, Alicante, Spain, jalcaraz@umh.es, Mercedes 

Landete, Juan Francisco Monge 

In this work we propose two new advanced hybrid metaheuristics to solve the 

Reliability P-Median Problem (RPMP) in which the objective is to minimize 

the operation costs but also hedging against cost failures within the system. 

The algorithms we have designed, a scatter search technique and a genetic 

algorithm, have been hybridized with procedures which incorporate problem 

specific knowledge in order to improve the efficiency of the techniques and 

the quality of the solutions found. The computational experiment carried out 

shows the excellent performance of these metaheuristics. 

3 - A grouping genetic algorithm for the capacitated facility 

location problem 

Diptesh Ghosh, Production and Quantitative Methods, Indian 

Institute of Management, Ahmedabad, Vastrapur, Ahmedabad, 

Gujarat, India, diptesh@iimahd.ernet.in, Megha Sharma 

We present a grouping genetic algorithm for the capacitated facility location 

problem (CFLP). We first use Monte Carlo simulations to ascertain a group of 

facilites with a high probability of being opened in a good solution along with 

another group of facilities with a low probability of being opened in such a solution. 

We also find groups of facilities which co-exist in a good solution, and 

other groups which do not co-exist in such a solution. We use these two groupings 

to develop a genetic algorithm for the CFLP. We present our computational 

experience on benchmark CFLP instances. 

193

TF-04 EURO 24 - Lisbon 2010 

� TF-04 


3.2.13 

Airline applications 



Chair: Elaine Wong, Innovation Works, European Aeronautic 

Defence and Space Company, 41 Science Park Road #01-30, The 

Gemini, Science Park II, 117610, Singapore, -, Singapore, 

elaine.wong.kl@eads.net 

Chair: Pasu Ratprasert, aeronautics, NCKU, NCKU Department of 

Aero neutics & Astroneutics, university Road, 701, tainan, Taiwan, 

cessna-skyline@hotmail.com 

1 - Ant colony optimization applied on weekly fleet assignment 

with time window model 

Pasu Ratprasert, aeronautics, NCKU, NCKU Department of 

Aero neutics & Astroneutics, university Road, 701, tainan, 

Taiwan, cessna-skyline@hotmail.com 

Fleet assignment consists of deciding on the type of aircraft that will operate 

each specific flight. The objective is to maximize the total profit. The inherent 

complexity of fleet assignment problem has normally resulted in development 

of integer programming based model and various heuristic procedures.This paper 

we introduce the Max-Min Ant System to solve weekly fleet assignment 

with time window.Using data from major airlines,the result from the implementation 

and evaluation confirm that the proposed ACO is suitable for the 

airline fleet assignment problem with good performance. 

2 - Exploiting metaheuristics to optimize performancebased 

logistics contracts for MRO services 

Elaine Wong, Innovation Works, European Aeronautic Defence 

and Space Company, 41 Science Park Road #01-30, The Gemini, 

Science Park II, 117610, Singapore, -, Singapore, 

elaine.wong.kl@eads.net, Zhichao Zheng, Arnd Schirrmann 

A key challenge of using Performance-Based Logistics contracts for aircraft 

MRO services is pricing competitively and profitably in face of operational 

uncertainties. Previous work to solve this contracting problem adopted the 

principal-agent model. This work extends existing models by incorporating 

integrality constraints. Using Simulated Annealing, we show how the MIP 

problem can be solved efficiently by appropriate (a)initial solution derivation, 

(b)neighbour generation, and (c)solution acceptance. Graphical representations 

of results from a real test case will be provided and analyzed. 

3 - Stochastic Airport Gate Assignment Problem 

Merve ¸Seker, Manufacturing Systems/Industrial Engineering, 

Sabanci University, Sabanci University, Orhanli, Tuzla, 34956, 

istanbul, Turkey, merveseker@su.sabanciuniv.edu, Nilay Noyan 

We consider the flight-gate assignment problem in the presence of uncertainty 

in arrival and departure times of the flights. We develop stochastic programming 

models incorporating alternate robustness measures to obtain assignments 

that would perform well under potential random disruptions. In particular, we 

focus on the buffer and idle times, and the number of conflicting flights as robustness 

measures. The proposed problems are formulated as computationally 

expensive large-scale mixed-integer LPs. In order to find good feasible solutions 

in short CPU times, we employ tabu-search algorithms. 

� TF-05 


3.2.16 

Meet the Editors of EJOR 

Stream: EJOR 


Chair: Gerard Wanrooy, Economics and Decision sciences, Elsevier 

Science, P O Box 1991, Amsterdam, Netherlands, 

g.wanrooy@elsevier.nl 

1 - Meet the Editors of EJOR 

194 

Roman Slowinski, Institute of Computing Science, Poznan 

University of Technology, Laboratory of Intelligent Decision 

Support Systems, Street Piotrowo 2, 60-965, Poznan, Poland, 

roman.slowinski@cs.put.poznan.pl, Jesus Artalejo, Jean-Charles 

Billaut, Robert Dyson, Lorenzo Peccati 

During this session, the editors of the European Journal of Operational Research, 

Jesus Artalejo, Jean-Charles Billaut, Robert Dyson, Lorenzo Peccati 

and Roman Slowinski, will present the editorial policy (increase of selectivity, 

invited reviews, special issues, plagiarism problems, etc.) and production 

report of the journal (bibliometry, rejection rate, geographical breakdown, ScienceDirect 

downloads, etc.). They will also explain their approach to evaluation 

and selection of articles, and will point out topics in which methodological 

papers, invited reviews and application papers are particularly welcome. Some 

general questions will be welcome during this session, but it is not the intention 

to discuss individual articles (published, rejected or in process). A drink 

offered by Elsevier will be served to the participants. 

� TF-06 


8.2.30 

DEA Application IV — Quality of Life and 

development 



Chair: Juan Ventura, Administración de Empresas, University of 

Oviedo, Av. Cristo s/n, 33071, Oviedo, Spain, ventura@uniovi.es 

1 - Evaluation of performance of european cities with the 

aim to promote quality of life improvements 

Paulo Morais, Engenharia Industrial e Gestão, Faculdade de 

Engenharia da Universidade do Porto, Rua Roberto Frias, 

4200-465, Porto, Portugal, pauloteixeirademorais@gmail.com, 

Ana Camanho 

This paper explores the possibilities presented by DEA to assess quality of life 

and evaluate performance of city managers in the promotion of urban quality 

of life. Using the data provided by the Urban Audit program from the EU we 

defined the profile for 206 cities. Two approaches are presented: a composite 

indicator of quality of life and the performance of local management, contextualised 

by the GDP per capita to take into account national economic conditions. 

The results identify the cities with urban best practices and present a model of 

intervention for the cities considered inefficient 

2 - Efficiency of public expenditure: a two-stage dea approach 

Janaina Tenorio, of Economics, Federal University of 

Pernambuco, R. Prof. Marculino Botelho, 835, ap 401, 

53130-150, Olinda, Pernambuco, Brazil, 

janainaratis@gmail.com, Francisco Ramos 

The importance of public expenditure policies for economic development is 

widely recognized. It is argued that the local government has superior advantage 

in providing such services because its proximity to users facilitates the 

identification of the social needs. This paper aims to measure the efficiency 

of the expenditures made by the local governments in a Brazilian state using 

two-stages DEA estimators. The results showed that, in general, the smaller 

is the municipal the lower is its efficient index, characterizing the existence of 

economies of scale. 

3 - Optimal National Resource Allocation for Multi-Factor 

Development: Cross-Country Analysis Based on DEA 

Abdel Latef Anouze, Busienss School, American University of 

Beirut, Bliss Street, 1107 2020, Beirut, Lebanon, 

aamajed2001@hotmail.com, Neil Yorke-Smith 

Development, economic and other, depends on prudent allocation of national 

resources. Recent works that rank countries using DEA approaches, taking 

Human Development Factors as the primary model input and outputs, offer 

classification of countries but fail to provide strategies to improve a country’s 

rank. We propose such strategies, combining DEA with Classification and Regression, 

in order to suggest rules that indicate the most effective allocation of 

resources. By considering a wider range of national performance indicators, 

we aim toward applicability and acceptability of the results.

4 - Assessment of quality of life using value efficiency 

analysis: the importance of the geographic level of 

analysis 

Eduardo Gonzalez, Business Administration, University of 

Oviedo, Av Cristo s(n, 33071, Oviedo, Asturias, Spain, 

efidalgo@uniovi.es, Ana Cárcaba, Juan Ventura 

Here we attempt to quantify the relative importance of three different geographic 

levels of analysis in assessing the quality of life (QoL) of the Spanish 

population. We evaluate how much QoL is explained by the province and 

region in which the municipality is located. We first construct a composite indicator 

of QoL for the 643 largest municipalities of Spain using 19 variables 

which are weighted by means of VEA, a refinement of DEA. The results show 

that the municipal level is the most important accounting for a 52% of the variance 

in QoL. 

� TF-07 


8.2.47 

Scheduling Applications 



Chair: Malgorzata Sterna, Institute of Computing Science, Poznan 


Malgorzata.Sterna@cs.put.poznan.pl 

1 - Scheduling MapReduce Computations 

Maciej Drozdowski, Dept. of Computer Science and 

Management, Institute of Computing Science, Poznan Univeristy 

of Technology, Piotrowo 2, 60-965, Poznan, Poland, 

Maciej.Drozdowski@cs.put.poznan.pl, Joanna Berlinska 

We analyze MapReduce distributed computations. MapReduce can be understood 

as two divisible applications with precedence constraints. This poses 

a scheduling problem of transferring load between two parallel applications 

using network with limited bandwidth. Schedule dominance properties are analyzed. 

We investigate performance limits of MapReduce computations. To 

our best knowledge this is the first time that processing divisible loads with 

precedence constraints is considered on the grounds of divisible load theory. 

2 - Scheduling re-entrant jobs in the car factory paint shop 

Grzegorz Pawlak, Institute of Computing Science, Poznan 


Poland, grzegorz.pawlak@cs.put.poznan.pl 

The car factory paint shop production system was considered in the paper. Low 

cost and efficient production are always the car makers’ goals. The scheduling 

jobs problem for the practical car factory paint shop was considered. The 

planned car sequence is disturbed by the repainted cars. The reduction of the 

negative impact of that process is the purpose of the work. Proposed mathematical 

models take into account the real production system and real data 

constraints. For these models the solution algorithms are developed and computational 

experiments are preformed. 

3 - Balancing and Equalizing Problem of Assembly Lines 

Waldemar Grzechca, Faculty of Automatic Control, Electronics 

and Computer Science, The Silesian University of Technology, 

Akademicka 2A, 44100, Gliwice, Poland, 

waldemar.grzechca@polsl.pl 

Assembly lines became one of the most popular production structures in manufacturing 

systems. The problem of balancing is connected with assigning all 

tasks to the workstations. The optimal solution is characterized by total zero 

idle time in whole system and 100% line efficiency. Mostly only feasible solutions 

are possible to obtain using heuristics methods. Some quality measures 

(line efficiency, station efficiency, smoothness index and time of line) allow us 

to estimate the final result. Very often instead of balancing problem the equalizing 

problem is considered. Equalizing allows us to avoid situation where one 

of the workstation is clearly loaded with low station time in comparison with 

other workstations. 

4 - Reflecting Scheduling Goals with Due Date Involving 

Criteria 

Malgorzata Sterna, Institute of Computing Science, Poznan 


Malgorzata.Sterna@cs.put.poznan.pl 


The performance measure is an important component of every scheduling problem. 

It allows for modeling different goals which should be achieved in constructed 

solutions. We collect some results obtained for a few criteria involving 

due dates and deadlines such as minimizing late work or maximizing revenue. 

Slight changes in the definitions of objective functions make it possible to cover 

different application fields. We present definitions of several scheduling problems 

with some results obtained for them. 

� TF-08 


6.1.36 

Scheduling with Uncertainties 



Chair: Andreas Ernst, Mathematics, Informatics and Statistics, 

CSIRO, Gate 5, Normanby Road, 3168, Clayton, Vic, Australia, 

Andreas.Ernst@csiro.au 

1 - Algebra of uncertain variables applied to task scheduling 

Maciej Hojda, Computer Science and Management, Wroclaw 

University of Technology, Wybrze˙ze Wyspiańskiego 27, 50-370, 

Wrocław, Dolnoslaskie, Poland, maciej.hojda@pwr.wroc.pl 

Uncertain variables, as introduced by Bubnicki, are a useful tool for dealing 

with non-determinism in operation complexes. Proposed is an approach 

that simplifies problem solving for a class of uncertain variables with (nonsymmetric) 

triangular certainty distributions. It is shown that uncertain variables 

form a linear algebra and an expression for a linear combination of a 

set of uncertain variables is derived. Application of the proposed approach is 

shown on a simple task scheduling problem with uncertain execution times. 

2 - Dynamic programming approximations for a stochastic 

machine scheduling problem 

Débora Ronconi, Production Engineering, University of Sao 

Paulo, Av. Prof. Almeida Prado,128, 05508070, Sao Paulo, Sao 

Paulo, Brazil, dronconi@usp.br, Warren Powell 

In this work we examine a scheduling environment where jobs arrive to the 

system at random times. Jobs have to be scheduled without knowing what jobs 

will come afterwards. We investigate the application of approximate dynamic 

programming in order to help decide whether a job should be scheduled for 

today or some day in the future to minimize the total tardiness. Computational 

experiments are presented in a set of 750 problems and the proposed approach 

was found to be superior to known myopic policies. 

3 - An exact solution algorithm for parallel machine 

scheduling with stochastic release dates 

Andreas Ernst, Mathematics, Informatics and Statistics, CSIRO, 

Gate 5, Normanby Road, 3168, Clayton, Vic, Australia, 

Andreas.Ernst@csiro.au, Gaurav Singh 

Consider a static stochastic machine scheduling problem where jobs with fixed 

duration and stochastic release dates are to be scheduled on parallel machines, 

subject to precedence constraints. Here a solution specifies both the order in 

which jobs are to be processed on each machine and a due date (or advertised 

completion date) for each job. The aim is to minimise the sum of exepected 

completion times and expected tardiness. We develop a branch and bound algorithm 

that can solve these kinds of problems exactly and demonstrate the 

computational effectiveness of the scheme. 

195


� TF-09 


6.2.53 

Various Aspects of Modern Mathematical 

Programming 








1 - A New Disaggregated Formulation of the Generalized 

Assignment Problem and Its Associated Inequalities 

Ishwar Murthy, QMIS, Indian Institute of Management 

Bangalore, NF103, IIM Campus, Bannerghatta Road, 560076, 

Bangalore, Karnataka, India, ishwar@iimb.ernet.in 

We present a new disaggregated formulation of the Generalized Assignment 

Problem, consisting of O(mn2) variables and constraints, where n denotes the 

number of jobs and m the number of agents. We show this formulation to 

be stronger than the traditional formulation. We then present generalizations 

of the Cover and (1,k)-Configuration inequalities whose presence in this disaggregated 

formulation is a lot more ubiquitous than their counterparts in the 

traditional formulation. Finally, we present inequalities that involve multiple 

agents that are specific to this formulation. 

2 - Feasibility of Integer Knapsacks and Lattice Problems 

Iskander Aliev, School of Mathematics, Cardiff University, 

Senghennydd Road, CF24 4AG, Cardiff, alievi@cf.ac.uk 

We will discuss recent results on feasibility of integer knapsacks, obtained by 

applying lattice techniques. The feasibility problem appears to be related to 

the geometric structure of a certain feasible set which, apart from a few special 

cases, remains unexplored. Known results suggest that the set may be decomposed 

into the set of all integer points in the interior of a certain translated 

feasible cone and a complementary set with complex combinatorial structure. 

We give an optimal estimate for the position of the feasible cone and show that 

a much stronger estimate holds on average. 

3 - Complexity of a Particular Class of 0-1 Linear Multiplicative 

Fractional Programming Problems 

Simranjit Kaur, Mathematics, University of Delhi, !00 Dr. 

Mukherjee Nagar, Delhi, 110001, Delhi, Delhi, India, 

simran_1209@rediffmail.com 

The paper deals with hardness of 0-1 Linear Fractional Multiplicative Programming 

Problem considered under the assumptions: each factor in numerator and 

denominator is positive; each decision variable in every factor in the denominator 

has a non zero coefficient. We have also discussed the complexity of 

checking whether the problem has a unique solution or not, hardness of solving 

the problem even in case of unique solution, complexity of local search and 

global verification and complexity of the problem with cardinality constraint. 

4 - Finding all the vertices of a convex polyhedral set 

Israfil Roshdi, Department of Mathematics, Science and 

Research Branch,Islamic Azad University, Tehran, Iran, Rajaei 

Aveune. Eslamabad alley. No 47. Marand,Iran, 5417733737, 

Marand, East-azerbaijan, Iran, Islamic Republic Of, 

i.roshdi@gmail.com, Mostafa Davtalab Olyaie 

A system of linear inequality constraints determines a convex polyhedral set of 

feasible solutions S. We consider the problem of finding all the extreme points, 

pay attention to redundancy and adjacency. We will construct a mixed integer 

programming problem and will give some structural theorems. By converting 

the main problem into two small sub problems; we will present an efficient 

algorithm for finding all extreme points of S. Our approach is totally different 

with the other methods in theory and practice. We will illustrate our algorithm 

by numerical examples. 

196 

� TF-10 


6.2.56 

OR in Sports 2 

Stream: OR in Sports 


Chair: Anastasios Oikonomidis, management, university of 

southampton, United Kingdom, tasos33bc@yahoo.co.uk 

1 - Subjective Judgment vs Market Idiosyncrasies; Exploring 

Influences on the Favourite Longshot Bias in European 

Betting Markets 

Anastasios Oikonomidis, management, university of 

southampton, United Kingdom, tasos33bc@yahoo.co.uk, 

Johnnie Johnson 

A dataset of 55,880 football events is analyzed to explore the favourite longshot 

bias in the bookmaker betting market. Limited dependent variable modelling 

of game outcomes is employed to identify sources of variation in the observed 

bias. It is concluded that it is consistently expressed through time, but varies 

significantly across leagues. It is shown that the bias is caused by bookmakers 

offering better odds for popular bets in order to increase their customer basis. 

Finally, it is proved that league specific fundamental information causes variation 

in the magnitude of the bias. 

� TF-11 


8.2.38 

Structural Equation Modelling Approach in 

User Acceptance of Information Technology 

II 






1 - An investigation into the acceptance of mobile banking 

by turkish consumers 

Gülgün Afacan, Information Systems, METU, METU 

Information Systems, 06531, Ankara, Ankara, Turkey, 

e157093@metu.edu.tr, Sevgi Ozkan 

Carrying out banking operations via mobile terminals is defined as mobile 

banking. Although number of mobile users is increasing throughout the world, 

this figure has not been observed in rate of mobile banking transactions. In 

this regard, major purpose of this study is to explore acceptance and adoption 

of mobile banking services by Turkish consumers. Theoretical framework 

of the study is based on Technology Acceptance Model (TAM) and also integrate 

trust, system quality, financial cost and self-efficacy constructs to fit the 

research model in mobile banking field. 

2 - Factors Influencing Users’ Adoption of Technology: 

Empirical Investigations in Different Contexts using 

Structural Equation Modeling Approach 

Sevgi Ozkan, Information Systems, Middle East Technical 

University, ODTU Enformatik Enstitüsü, Ismet Inönü Bulvari, 

06531, Ankara, Turkey, sozkan@ii.metu.edu.tr, Irfan Emrah 

Kanat, Emre Sezgin, Yasemin Çetin, Duygu Findik, Oguzhan 

Alasehir, Nurcan Alki¸S 

The factors that affect the user acceptance of technology related products and 

practices have attracted the attention of researchers. In this study, a technology 

acceptance model will be developed via Structural Equation Modeling approach. 

Technology Acceptance Model (TAM) will be extended in four different 

contexts, i.e. e-government, e-health, e-learning and e-commerce. The 

factors that affect the attitudes of users in each context will be investigated. The 

study will be implemented by collecting data within practices. This research is 

funded by TUBITAK project no: 109K394

3 - What is available about technology acceptance of elearning 

software and systems? A review and comprehension 

study 

Mustafa Degerli, Information Systems, Middle East Technical 

University, Informatics Institute, Dept. of IS, METU, Cankaya, 

Ankara, 06531, Ankara, Turkey, md.mustafadegerli@gmail.com, 

Sevgi Ozkan 

Applying technology by means of e-learning software and systems (e-LSS) 

to facilitate and support learning is an important and interested in application 

area recently. Yet, another important concern meant for this context is surely 

the technology acceptance of these e-LSS by the people. Although there are 

studies, but not many, conducted in this subject with respect to various contexts, 

there is lacking a study that reviews and summarizes previous studies 

and provides a comprehensive guide to let people know about the technology 

acceptance of e-LSS. This study aims to compensate this lack. 

4 - Evaluating User Acceptance of Internet Banking Service 

in Turkey 

Mustafa Aydin, Information Management, Banking Regulation 

and Supervision Agency, Ataturk Bulvari, No:191/B, 

Kavaklıdere Çankaya, 06680, Ankara, Turkey, 

maydin@bddk.org.tr, Sevgi Ozkan, Arif Yilmaz 

Online banking systems are beneficial both for banks as well as for users. 

Banks make an investment in internet based online banking systems to improve 

their operations and to reduce cost. The more users use online banking 

systems, the more they improve banks’ performance. Therefore banks require 

a better understanding of how users accept and use online banking services. 

The purpose of this study is to examine factors, especially perceived usefulness 

and perceived ease of use, that affect users to accept online banking based on 

Technology Acceptance Model (TAM) in Turkey. 

� TF-12 


8.2.39 

ANP 05 


Processes 


Chair: Yi-Chun Chen, Department of Leisure Management, Taiwan 

Hospitality & Tourism College, No 268 Chung-Hsing ST., Feng-Shan 

Village, Shou-Feng County, 974, Hualien, Taiwan, 

chen.vivien@gmail.com 

1 - Improving the Career-Change Tendency for Hot-Spring 

Hotels Using Dominance-base Rough Set Approach 





The stability of hotel employees is essential to overall management of a hotel. 

Yet it is reported that high turnover rate is found common in hotels and 

causes an increasing impact on hospitality industry. Given this, the study adopts 

Dominance-based Rough Set Approach (DRSA) to find a strategy helpful to 

improve the high career-change tendency in hot spring hotels. The model was 

implemented with survey data from a large sample of employees of the hot 

spring hotels in central Taiwan. 

2 - A decision Rules Approach for Marketing Improvement 

of European Art Tour Participants from Taiwan 





The goal of the research is to investigate the tour market of European Art-tour 

participants from Taiwan and learn how to improve this market for satisfying 

tourists’ needs of mental feeling-well in art enjoyment. This study is using a 

brand new method of rough set-based logics. It is Dominance-based Rough 

Set Approach (DRSA), a set of decision rules "if-then rule’ use in the preference 

model. The proposed method can provide practical information supposed 

helpful for tourism industry to develop the improvement strategies. 


3 - Constructing the Learning Criteria of Basic System for 

Community Industry Development Based on DEMATEL 

technique with MCDM Model 





This research aims to explore and establish the learning criteria of basic system 

for community industries development to prosper the core-values and happiness. 

To achieve such research purposes, the methodology includes Delphi 

technique, and then combined DEMATEL technique with new hybrid MCDM 

model to construct the learning criteria of basic system for community industry 

development. Most importantly, what learning criteria context of this study develops 

could be adopted as important references; then the best core-values and 

happiness of community industries can be achieved. 

4 - A Hybrid MCDM Model Combing DEMATEL-Based ANP 

with VIKOR for Establishing the Best Management Systems 

of Transformation-Learning Community Industry 





The purpose of this paper is to explore and establish the best management 

systems for transformation-learning community industries. To achieve such 

research purposes, a hybrid MCDM model combing DEMATEL-Based ANP 

(DANP) with VIKOR method is used to construct the learning criteria of the 

best management system for community industry transformation. Therefore, 

the best management systems of transformation-learning community constituted 

in this study could be adopted as important citation, and the best transformation 

criteria can be achieved. 

� TF-13 


2.2.21 

Competitive Location 



Chair: Dolores R. Santos-Peñate, Métodos Cuantitativos en 

Economía y Gestión, University of Las Palmas de Gran Canaria, 

Campus de Tafira. Edificio Dptal de Ciencias Económicas y 

Empresariales D-4-22, 35017, Las Palmas de Gran Canaria, Canarias, 

Spain, drsantos@dmc.ulpgc.es 

1 - A Competitive Location Model with Demand Uncertainty 

Tolga Seyhan, Industrial and Systems Engineering, Lehigh 

University, 8 Duh Drive Apt. 222, 18015, Bethlehem, PA, United 

States, ths207@lehigh.edu, Lawrence Snyder 

We consider a facility location problem under Stackelberg competition where 

two players sequentially locate their facilities in order to capture the maximum 

customer demand. We propose a model where the follower is assumed to employ 

a greedy add heuristic as his response, and formulate a mixed integer programming 

model that solves the leader’s problem under this assumption. Later, 

we bring demand uncertainty into the scene such that it exists as the leader 

makes her decision but is resolved when the follower makes his. We extend our 

first model into a two stage stochastic programming model. 

2 - Incorporating inventory decisions in competitive location 

models 

Francisco Silva, Economics and Business, University of the 

Azores, Rua da Mae Deus, Ponta Delgada, 9500, Ponta Delgada, 

Portugal, fsilva@uac.pt, Helena Ramalhinho Lourenço 

Competitive Location Models seek positions and prices which maximize the 

market captured by an entrant firm from previously positioned competitors. 

Nevertheless, strategic location decisions may have a significant impact on future 

inventory and shipment costs thus affecting the firm’s competitive advantages. 

In this paper we introduce a heuristic algorithm which considers both 

market capture and replenishment costs in order to choose the firm’s locations. 

Viswanathan’s (1996) algorithm is used to solve the replenishment problem 

whereas a Greedy Randomized Adaptive Search Procedure is used to solve the 

location problem. 

197


3 - A new method for the location and design of multiple 

new facilities in the competitive market 

Nasreddine Saidani, ICD, CNRS 2848 Université de Technologie 

de Troyes. France, 12, rue Marie Curie - BP 2060, 10010, 

TROYES, France, nasreddine.saidani@utt.fr, Feng Chu, Chen 

Haoxun 

This work investigates the location of multiple new facilities in a competitive 

market. The objective is to set up new facilities in a planar area, where similar 

facilities of competitors are already present. In EURO 2009 we proposed a new 

method with two stages for the location of a single facility in the market with a 

single facility of competitors. Different from the existing methods, our method 

takes in consideration of the reaction of competitors. In this work we generalize 

our method to the problem of locating multiple facilities with multiple facilities 

already presented in the market. 

4 - An LP-PSO matheuristic to solve the leader-follower 

problem 

Dolores R. Santos-Peñate, Métodos Cuantitativos en Economía y 

Gestión, University of Las Palmas de Gran Canaria, Campus de 

Tafira. Edificio Dptal de Ciencias Económicas y Empresariales 

D-4-22, 35017, Las Palmas de Gran Canaria, Canarias, Spain, 

drsantos@dmc.ulpgc.es, Clara M. Campos, José A. 

Moreno-Pérez, Rafael Suarez-Vega 

The leader-follower problem consists of determining optimal strategies for two 

competing firms, the leader and the follower. The follower maximizes its market 

share, given the locations chosen by the leader. The leader minimizes the 

maximum market share that the follower can get. We propose a resolution approach 

in which the leader strategies are represented by particles position in a 

swarm optimization procedure. Given the facility locations for the leader, the 

follower problem is solved using linear programming. 

� TF-15 


2.2.12 

Vehicle Routing Applications I 






1 - Periodic Vehicle Routing and Scheduling for End-of- 

Life Product Collection 

Yeliz Akca, Industrial Engineering, Koc University, Koc 

Universitesi, Rumeli Feneri Yolu„ Sariyer/Istanbul - Turkey, 

34450, Istanbul, Sariyer, Turkey, yakca@ku.edu.tr, Sibel Salman, 

Onur Kaya, Deniz Aksen 

We address the collection of end-of-life products to be used in a production 

process, such as used vegetable oil for bio-diesel production. We develop a 

MIP model to optimize the integrated production and periodic routing schedule 

for collection while satisfying operational constraints. The aim is to minimize 

the sum of transportation, inventory and outsourcing costs. We develop 

a Lagrangian Relaxation method to generate both lower bounds and feasible 

solutions. We also develop heuristics and compare their performance with the 

bounds found by Lagrangian Relaxation and MIP solutions. 

2 - Collection and recycling of WEEE facilities location 

Jorge Pereira, Production and Systems Department, University of 

Minho, Portugal, Universidade do Minho, Campus de Gualtar, 

4710-057, Braga, Portugal, jamilcar@gmail.com, Simão Ribeiro, 

Joel Carvalho, José Oliveira, Manuel Figueiredo, José Telhada, 

Luis Dias 

This project focuses on a network optimization for collecting waste of electrical 

and electronic equipment (WEEE). Waste recovery is part of the environmental 

policies, regulated by European Directives promoting the collection and 

recycling of such equipment (Directive 2002/96/EC) which has been in force 

since February 2003. This project addresses the facilities location issues and 

the related routing problems. It is expected that relevant economic and environmental 

benefits will be achieved, including increases in the quantities of WEEE 

collected and the reduction of operational costs. 

198 

3 - An Exact Algorithm for the Vehicle Routing Problem 

with Stochastic Demands 



Canada, michel.gendreau@cirrelt.ca, Walter Rei 

The Vehicle Routing Problem with Stochastic Demands (SVRPSD) consists in 

finding routes for a a fleet of capacitated vehicles delivering goods to customers 

with stochastic demands. When the cumulative demand of the customers assigned 

to a vehicle exceeds its capacity, the vehicle goes back to the depot to 

replenish its stock. The objective is to find the routes that yield the lowest 

expected total travel cost. We present an exact solution method which relies 

heavily on concepts of the local branching approach for mixed integer programs. 

Computational results will be reported. 

� TF-16 


2.2.14 

Optimization Methods for Railway Freight 






1 - A branch and generate heuristic (BANG) to solve largescale 

train timetabling problems 

Thomas Schlechte, Optimization, Zuse-Institute-Berlin, 

Takustrasse 7, 14195, Berlin, Berlin, Germany, 

schlechte@zib.de, Ralf Borndörfer, Steffen Weider 

We present a branch and generate heuristic to solve very large-scale instances 

of the macroscopic train timetabling problem (TTP). The TTP consists in finding 

a conflict free set of train routes of maximum value in a given railway 

network. We model the TTP as a large scale integer program and solve it by 

a column generation procedure using the bundle method. We show how our 

branch and generate heuristic performs in comparison to standard branch and 

bound approaches. Computational results are shown for a part of German network 

(available at tttplib.zib.de) and the Simplon corridor. 

2 - Modeling and Optimizing the Simplon Railway Corridor 

Elmar Swarat, Optimization, Zuse Institute Berlin, Takustr. 7, 

14195, Berlin-Dahlem, Germany, swarat@zib.de, Ralf 

Borndörfer, Thomas Schlechte 

We present a bottom-up approach of automatic network aggregation with application 

to the Simplon corridor, an important line in the European railway 

network. Our method reduces detailed microscopic infrastructure data from a 

simulation kernel to an aggregated, macroscopic level. We prove error estimates 

for the proposed transformation, which allows for an exact optimization 

of train schedules using integer programming. 

3 - Single Car Routing in Rail Freight Transport 

Armin Fügenschuh, Optimierung, Zuse Institut Berlin, 

Takustraße 7, 14195, Berlin, Germany, fuegenschuh@zib.de, 

Henning Homfeld, Alexander Martin, Hanno Schülldorf 

Cars in rail freight service follow prescribed routes from their origin via intermediate 

shunting yards to their destination. The goal in designing such routes 

is to reduce the number of trains and their travel distances. Various real-world 

hard constraints make the problem difficult to formulate and also to solve. We 

present mixed-integer linear and nonlinear programming formulations for this 

car routing problem arising at Deutsche Bahn, one of the largest European railway 

companies, and computational results using test- and real-world data. 

4 - Optimized fleet management of ballast supply railcars 

François Ramond, Innovation & Research, SNCF, 45 rue de 

Londres, 75008, Paris, France, francoisramond@gmail.com, 

Francis Sourd, Lionel Lagarde, Pierrick Vallat 

We propose a model of railcar fleet management. The railcars, used to deliver 

ballast to track renewal sites, follow the cycle: empty move, ballast loading 

within ballast quarry, loaded move towards a customer site, unloading of ballast, 

empty move towards a quarry, and so on. The objective is, given a fixed 

fleet size, to maximize demand satisfaction while minimizing the travelled distance 

and encouraging railcar reuse by the same customer. Demand is satisfied 

if ballast is delivered during a given time range, and quarries have capacity 

constraints on the number of in / out moves each day.

� TF-17 


1.3.14 

Technologies for Collaborative Planning 






1 - A Survey of Electronic Transportation Marketplaces in 

Germany: How can IT-Innovation promote Collaborative 

Transportation Planning 

Michael Schwind, IT-based Logistics, Goethe University 


schwind@wiwi.uni-frankfurt.de, Susanne Aponte, Andreas 

Stenger 

Electronic transportation marketplaces (ETM) provide a platform for the exchange 

of logistics services between shippers, carriers, and freight forwarders. 

ETMs are essential to foster collaborative transportation, but most ETMs have 

not reached a significant influence in logistics until now. In a survey of German 

logistics platforms, we identify and analyze success factors and shortcomings 

of the current ETMs. Finally, we propose functionalities such as integrated 

tour planning, tracking-and-tracing and route pricing based on combinatorial 

auctions in order to make the ETMs more appealing. 

2 - A proposal of autonomous agents’ implementation for 

the maritime logistics of manufactured products 

Vanina Macowski Durski Silva, Production Engineering, Federal 

University of Santa Catarina, Rua Apeninos n. 38, apto. 303, 

Bairro: Córrego Grande, 88037620, Florianópolis, Santa 

Catarina, Brazil, vaninadurski@gmail.com, Antonio G.N. 

Novaes, Antönio Coelho 

This beginner study aims to identify and analyze major strategic and operational 

parameters that work in maritime logistics of manufactured products 

for export. The work will be proceed on two ways, a theoretical and other 

numerical, in order to research the Collaborative Transportation Management 

to propose a model for the logistical problem under study. In the sequence, 

the autonomous agent theory in dynamic systems will be applied to solve the 

problem. Moreover, attempts to establish a criterion to distribute the costs and 

benefits among the key players in the transportation chain. 

3 - Minimizing CO2-Emmissions in Transportation Logistics 

via Autonomous Cooperation? Insights from a stylized 

Model 

Richard Colmorn, Jacobs University Bremen, 28759, Bremen, 

r.colmorn@jacobs-university.de, Olivier Gallay, Philip Cordes, 

Max-Olivier Hongler, Michael Hülsmann 

Awareness about global warming increases the importance of the concept of 

Green Logistics for logistics service providers. In that regard and focusing 

on transportation logistics, using autonomous cooperation technologies (e.g. 

RFID) might be a promising approach. In this paper, a stylized model is proposed 

that helps to answer the question whether decentralized or centralized 

organization of transportation networks should be favoured in order to minimize 

the CO-2 emissions produced by fuel consumption. Furthermore, possible 

extensions are introduced that enable to incorporate other potential sources 

of CO-2 emissions occurring in transportation processes. 

4 - Auction-based request re-allocation in large-scale 

transport cooperations considering cost and quality 

Tobias Buer, Dept. of Information Systems, FernUniversität - 

University of Hagen, Profilstr. 8, 58084, Hagen, Germany, 

tobias.buer@fernuni-hagen.de, Giselher Pankratz 

We examine the problem of re-allocating a large number of transportation requests 

among cooperating transportation firms using a combinatorial exchange 

mechanism. We model the problem as a bi-objective extension to the set covering 

problem, taking into account both minimization of total transportation 

cost and maximization of overall transportation service quality. For solving the 

problem, we propose a ’matheuristic’ which applies a problem specific branchand-bound 

algorithm to the path relinking phase of a bi-objective GRASP metaheuristic. 

Numerical results are given. 


� TF-18 


1.3.15 

Data Mining and Knowledge Representation 











1 - Visualization and Representation of Temporal Knowledge 

Jiang-Liang Hou, Dept. of Ind. Eng. & Eng. Mgt., National 

Tsing Hua Univ., No. 101, Section 2, Kuang-Fu Road, Hsinchu, 

Taiwan 30013, R.O.C., 30013, Hsinchu, Taiwan, Taiwan, 

adamhou@ie.nthu.edu.tw, Shih-Ting Yang, Wei-Ning Pi, 

Yuh-Liang Chen 

A model for temporal knowledge visualization is presented to enhance reuse 

of temporal knowledge. The free-form, text-based temporal knowledge can be 

converted into visualized illustrations via the proposed visualization approach. 

The model includes three main modules namely full text tagging (FTT), event 

sequence analysis (ESA) and temporal knowledge visualization (TKV). The 

model can improve the efficiency and effectiveness for the knowledge receivers 

to acquire the temporal knowledge in the text-based documents and can be applied 

in enterprises for employee training and other activities. 

2 - Comparing partitions: visual aids 

Margarida Cardoso, Métodos Quantitativos, ISCTE-IUL, Av. 

Forças Armadas, 1649-026, Lisboa, Portugal, 

margarida.cardoso@iscte.pt, Ana Alexandra Martins 

The focus of our analysis is the data of a contingency table depicting the association 

between two partitions. In the context of clustering evaluation, some 

authors solve an assignment problem to match the two partitions. In addition, 

there are several coefficients quantifying agreement between two partitions. We 

suggest considering the contingency table data as similarities and using Multidimensional 

Scaling to visualize the relationship between the partitions. The 

proposed approach simultaneously helps visualizing the match between partitions 

and the distance between the corresponding groups 

3 - On the efficiency of Spectral Clustering : interpretation, 

parallel computation and results 

Sandrine Mouysset, IRIT-ENSEEIHT, 2 rue Camichel, 31000, 

Toulouse, France, sandrine.mouysset@enseeiht.fr, Joseph 

Noailles, Daniel Ruiz 

Spectral Clustering (SC) is one of the most important method based on dimension 

reduction space in Data Mining. It consists in defining a low-dimensional 

data space in which data points are clustered by selecting dominant eigenvectors 

of a Gaussian affinity matrix. With a reformulation of SC algorithm as an 

eigenvalues problem, an interpretation on how this method works is given. By 

exploiting this theoretical material, we propose a domain decomposition strategy 

for parallel SC. Additionally, with a criterion for determining the number 

of clusters, this strategy becomes robust and efficient. 

4 - WhiBo - A platform for component-based design of partitioning 

cluster algorithms 

Kathrin Kirchner, Faculty of Business and Economics, Friedrich 

Schiller University Jena, Department of Business Informatics, 

Carl - Zeiss - Strasse 3, D - 07743, Jena, Germany, 

kathrin.kirchner@uni-jena.de, Boris Delibasic, Milos Jovanovic, 

Milan Vukicevic, Johannes Ruhland 

We suggest a process of designing new cluster algorithms by structuring existing 

ones as sets of reusable components (RCs). New algorithms can now be 

created as innovative component combinations. To allow for their evaluation, 

a RC repository has been created and is integrated into our WhiBo platform 

for RapidMiner. Very often, it is found that desired properties can literally be 

engineered into an algorithm through judicious combination of components. 

199

TF-20 EURO 24 - Lisbon 2010 

� TF-20 


1.3.33A 




Chair: A. Miguel Gomes, Fauculty of Engineering / INESC Porto, 

University of Porto, Rua Dr. Roberto Frias s/n, 4200-465, Porto, 

Portugal, agomes@fe.up.pt 

1 - Multilevel cutting optimisation with production costs in 

practice 

Helmut Schreck, TietoEnator Deutschland GmbH, 81379, 

München, Germany, Helmut.Schreck@tietoenator.com 

Solving cutting stock problems in industrial environment has to deal with a variety 

of production and handling costs besides standard goals like trim loss or 

knife changes. Cutting optimization in paper or steel industry is often one of 

the key points to reduce overall production costs. In our paper we give 2 examples 

from steel and paper production and discuss the requirements and solution 

approaches for cutting stock algorithms. 

2 - A Mathematical Programming Approach For Retail 

Space Planning 

Gabriel Tavares, FICO, 551 Openaki Road, 07834, Denville, NJ, 

United States, gabrieltavares@fico.com 

A MIP approach to automatically generate planograms based on presentation, 

inventory and assortment goals, is presented. The model considers a set of 

bounded knapsack constraints that restrict the item facings on every planogram 

fixture. For shelf fixtures, the packing is a 1-dimensional item tiling problem. 

For peg boards, the problem consists of tiling items (rectangles) into blocks 

(larger rectangles). Using the solver FICO-Xpress and the modeling environment 

Xpress-Mosel, this approach has been deployed for a large US retailer 

and can typically handle hundreds of items per request. 

� TF-21 


6.2.47 

Software for OR/MS III 



Chair: Gordon Dash, Finance and Decision Sciences, University of 

Rhode Island, 7 Lippitt Road, College of Business Administration, 

02881, Kingston, RI, United States, ghdash@uri.edu 

1 - SIOPRED Web Platform: a Web Application for automatic 

forecasting of time series 

Francesc Silva, University of Valencia, Valencia, Spain, 

Francesc.Silva@uv.es, José D. Bermúdez, José Vicente Segura, 

Enriqueta Vercher, Ana Corberan 

An automatic forecasting procedure based on exponential smoothing has been 

implemented as a web application using the kernel of SIOPRED. The procedure 

uses a multi-objective formulation that allows us to estimate jointly all 

the unknowns — smoothing parameters and initial conditions — of the forecasting 

method in a Soft Computing framework. The Web application enables 

to choose among three fitting measures and different exponential smoothing 

methods with additive or multiplicative seasonality. The application includes a 

client management suite and a SVG dynamic graphic generator module. 

2 - Software for bound-constraint minimization 

Ernesto G. Birgin, Dept. of Computer Science, University of São 

Paulo, Rua do Matão, 1010, Cidade Universitária, 05508-090, 

São Paulo, SP, egbirgin@gmail.com 

In this talk we present recent advances in the development of a software 

for bound-constraint minimization. A comparison with some state-of-the-art 

solvers is presented and analysed. The proposed method, embedded in an 

augmented Lagrangian framework, is also evaluated for solving nonlinear programming 

problems. 


3 - Optimizing Automated Share Trading Using WinORSe- 

AI: Cognitive Decision Theory and High Frequency Artificial 

Neural Networks 

Gordon Dash, Finance and Decision Sciences, University of 

Rhode Island, 7 Lippitt Road, College of Business 

Administration, 02881, Kingston, RI, United States, 

ghdash@uri.edu, Nina Kajiji 

Recent technological advances have helped to usher in an era where both automated 

and algorithmic trading have come to characterize current trends in 

real-time equity trading. This talk presents The WinORSe-AI stochastic price 

formation algorithm (WINKS). It is a cognitive decision making system based 

on twin radial basis function artificial neural networks. WINKS is a high frequency 

automated trading system for individual equity securities traded on 

U.S. exchanges. We report and demonstrate effective risk-adjusted trading that 

equals or exceeds results from a simple buy-hold strategy. 

� TF-22 



Health Care Policy Making III 



Chair: Amílcar Arantes, Civil Engineering-DECivil, IST/Technical 

University of Lisbon, Av. Rovisco Pais, 1049-001 , LISBOA, 

Portugal, amilcar.arantes@civil.ist.utl.pt 

1 - Procurement strategies in the health sector — How advanced 

are procurement processes in Portuguese hospitals? 

Stephan Messner, Universidade Nova de Lisboa, Rua Fernando 

Cabral N12/3e, 1750-329, Lisboa, Portugal, 

stephan.messner@gmx.at, Amílcar Arantes 

Applying advanced strategies for procurement can help hospitals to save substantial 

amounts of money. Group purchasing organisations, the use of consignment 

stock and supplier performance management are strategies especially 

suited to improve the procurement process in hospitals. Based on focus interviews 

with procurement managers of Portuguese hospitals it became clear 

that there is enormous room to improve the general way how procurement is 

done but also the application of the three strategies mentioned. Given the overall 

importance of hospital logistics, the hospitals’ performance could improve 

substantially when advancing the hospitals’ procurement processes. 

2 - Binary matrix decompositions without tongue-andgroove 

underdosage for radiation therapy planning 

Celine Engelbeen, Department of mathematics, Universite Libre 

de Bruxelles, Boulevard du Triomphe, CP 216, 1050 Brussels, 

Brussels, Belgium, cengelbe@ulb.ac.be, Antje Kiesel 

We consider the problem of decomposing a binary matrix into a nonnegative 

integer linear combination of some particular binary matrices satisfying the 

consecutive ones property as well as the tongue-and-groove constraint. We 

prove that this problem is polynomially solvable and provide an algorithm to 

solve it. This problem arises in the elaboration of radiation therapy plans. 

3 - Lean Healthcare in Portuguese Hospitals — a case 

study 

Luis Ferreira, DEGEI, Universidade de Aveiro, Campus 

Universitário Santiago, 3810-193, Aveiro, Portugal, 

lmferreira@ua.pt, Filipe Simões 

Nowadays the evolution of concepts and the growing need to present a rational 

and effective management of resources in an environment filled with difficulties, 

some of them chronic, has in the healthcare, and obliges to implement 

new management practices such as Lean Management. The purpose of this 

study is to present our findings in relation to how lean is being implemented 

in Portuguese public hospitals. Several hospitals were visited and we describe 

in detail one lean implementation project and the correspondent impacts in the 

organization.

� TF-23 


6.2.49 

MOO: Multiple Criteria Approaches in 

Mathematical Finance 



Chair: Frank Heyde, Institute of Mathematics, MLU 

Halle-Wittenberg, Theodor-Lieser-Str. 5, D-06099, Halle (Saale), 

Germany, heyde@mathematik.uni-halle.de 

1 - Modelling a private equity fund 

Johannes Tannert, Mathematics, 

Martin-Luther-University-Halle-Wittenberg, Theodor-Lieser-Str. 

5, 06120, Halle, Sachsen-Anhalt, Germany, 

johannes.tannert@mathematik.uni-halle.de 

In this talk, the focus is on modelling a private equity fund. Stochastic differential 

equations describe the draw downs, the return repayments and the 

investment value of the fund from the investors view. These are the restrictions 

of a multiobjective Markowitz optimization problem with expected return 

and variance as objective functions. Furthermore, several objective functions 

like short- and long-term profits and different risk measures are added. The 

multiobjective optimization problems are solved with MATLAB. 

2 - A set-valued approach to minimizing risks in markets 

with transaction costs 

Andreas Hamel, Operations Research and Financial Engineering, 

Princeton University, Sherrerd Hall 224, 08544, Princeton, NJ, 

United States, ahamel@princeton.edu 

A main source for incompleteness of a financial market is the presence of transaction 

costs. Major constructions for such markets (e.g. super-hedging prices, 

risk measures) are best understood as set-valued functions. We present an approach 

to set-valued optimization problems involving risk minimization of setvalued 

risk measures in conical market models: A solution concept is defined, 

optimality conditions and the construction of dual problems are discussed. As 

a particular example, a portfolio selection problem is discussed involving a setvalued 

variant of the average value at risk. 

3 - Duality for Bicriteria Portfolio Optimization using Average 

Value at Risk 

Frank Heyde, Institute of Mathematics, MLU Halle-Wittenberg, 

Theodor-Lieser-Str. 5, D-06099, Halle (Saale), Germany, 

heyde@mathematik.uni-halle.de, Andreas Löhne, Christiane 

Tammer, Mandy Werfel 

We consider a bicriteria Markowitz type of portfolio optimization problem 

where the risk is expressed by the Average Value at Risk. This problem can 

be approximated by a linear vector optimization problem. We introduce a setvalued 

and a geometric dual problem and give an interpretation of the dual 

problems for the portfolio problem. Moreover, we present numerical results 

using a primal and dual variant of the Benson algorithm. 

4 - Strong KKT conditions in convex vector optimization 

Joydeep Dutta, Math and Stat, Indian Institute of Technology, 

Office : Room No 575, Faculty Building, Academic Area, 

208016, Kanpur, Uttar Pradesh, India, jdutta@iitk.ac.in, Regina 

Burachik 

In this paper we provide a simple approach using the recent developments in 

scalar convex optimization to derive strong KKT conditions for vector optimization 

problems. 


� TF-24 


6.2.50 

Workforce Scheduling 1 

Stream: Timetabling and Rostering 


Chair: Joao Telhada, Departamento de Estatística e Investigação 

Operacional / Statistics and Operations Research Department, Centro 

de Investigação Operacional / Operational Research Centre - 

Faculdade de Ciências da Universidade de Lisboa / Faculty of 

Sciences - University of Lisbon, Bloco C6 - Campo Grande, 

1749-016, Lisboa, Portugal, joao.telhada@fc.ul.pt 

1 - Understanding an Operations Research Customer — Or 

Not 

Kimmo Nurmi, Research and Development, Satakunta 

University of Applied Sciences, Tiedepuisto 3, 28600, Pori, 

Finland, cimmo.nurmi@samk.fi, Jari Kyngas 

Good rosters have many benefits for an organization, such as lower costs, more 

effective utilization of resources and fairer workloads and shifts. The process 

of constructing optimized work timetables for the personnel is a demanding 

task. This paper presents a process of understanding and not understanding a 

customer. The presented staff scheduling problem can be divided into two separate 

sub-problems: days-off scheduling and shift scheduling. The algorithm 

used to solve the problem is a variation of the cooperative local search method. 

The generated software is currently in use. 

2 - Alternative MIP formulations for an integrated shift 

scheduling and rostering problem 

Joao Telhada, Departamento de Estatística e Investigação 

Operacional / Statistics and Operations Research Department, 

Centro de Investigação Operacional / Operational Research 

Centre - Faculdade de Ciências da Universidade de Lisboa / 

Faculty of Sciences - University of Lisbon, Bloco C6 - Campo 

Grande, 1749-016, Lisboa, Portugal, joao.telhada@fc.ul.pt, Ana 

Godinho 

A problem of personnel scheduling in a multiskilled environment is addressed. 

This problem is treated in an integrated manner, modeling shift scheduling and 

rostering as one problem. Additionally, the integrated approach allows also to 

better model intraday breaks and days-off scheduling. Alternative MIP formulations 

are presented which lead to optimal shift schedulings and task assignments. 

Improved models are obtained by deriving new block-indexed variables. 

Computational results show the improvement obtained by extended formulations. 

� TF-25 


6.2.48 

ROADEF/EURO challenge junior session 1 

Stream: ROADEF/EURO challenge 


Chair: Christian Artigues, LAAS, CNRS, 7 avenue du Colonel 

Roche, 31077, Toulouse Cedex 4, artigues@laas.fr 

1 - An Evolution Based Algorithm for a Large Scale Energy 

Management Problem 

Mustafa Kerim Yılmaz, Industrial Engineering, Kocaeli 

University, Umuttepe Campus, Engineering Faculty, 41000, 

Kocaeli, Turkey, m_kerim@hotmail.com, Mustafa Tacettin, 

Ahmet Cihan 

The large scale energy management problem is decomposed into two parts. 

First, we construct alternative outage schedules for achieving the maximum 

number of outages by using an evolutionary algorithm while satisfying the constraints. 

Then we calculate the production levels of both power plant types for 

required demand in each time step based on each alternative outage schedule 

while minimizing the cost by using a heuristic approach. 



2 - ROADEF/EURO Challenge 2010: A constructive heuristic 

for large-scale energy management problem with 

varied constraints 

Murat Firat, Mathematics and Computer Science, Eindhoven 

University of Technology, Postbus 513, 5600 MB, Eindhoven, 

Netherlands, m.firat@tue.nl 

The main strategy in my approach is to determine "lower" and "upper" bounds 

of both fuel amounts and starting times of type2 outages and to tighten the 

starting intervals of outages in order to schedule them with least possible conflicts. 

A constructive heuristics is designed to schedule the outages. Reducing 

the search space is achieved by determining the minimum cardinality subsets 

of outages subjected to the same constraint. The encountered conflicts among 

outages are resolved applying local search using these subsets. 

3 - An ACO/VNS Hybrid Approach for a Large-Scale Energy 

Management Problem 

Roman Steiner, Institute for Computer Graphics and Algorithms, 

Vienna University of Technology, Favoritenstraße 9-11, 1040, 

Wien, Austria, e0326433@student.tuwien.ac.at, Günther Raidl, 

Sandro Pirkwieser, Matthias Prandtstetter 

We use a combination of two different metaheuristics for solving the ROADEF 

2010 energy management problem. First, we search for good solutions with an 

ant colony optimization (ACO) with an integrated local search component. The 

best found solution is then passed to a variable neighborhood search (VNS). 

The available time is split approximately in half for both algorithms. It turned 

out that the ACO is good in finding some average feasible solution and the 

VNS is able to further improve it. This sequential collaboration yields better 

solutions than solely applying one of the metaheuristics. 

� TF-26 


3.1.11 

Cooperative situations on networks: 

algorithms and applications 



Chair: Stefano Moretti, UMR7024 - LAMSADE, CNRS - Université 

Paris-Dauphine„ Place du Maréchal de Lattre de Tassigny, F-75016, 

Paris, France, stefano.moretti@dauphine.fr 

1 - The G-Game: A Cooperative Game Approach for Resource 

Consolidation in Network Dimensioning 

Aruna Prem Bianzino, CNRS LTCI UMR 5141, Institut 

TELECOM, TELECOM ParisTech, 75013, Paris, France, 

bianzino@telecom-paristech.fr, Stefano Moretti, Dario Rossi, 

Jean-Louis Rougier 

Reduction of energy consumption is a major concern in wired networks, usually 

referred to as green networking. We propose a two-steps cooperative game 

to reduce the unnecessary energy consumption due to under-utilized network 

devices, and we discuss its feasibility and computational issues. As a first step, 

the importance of network devices is evaluated according to the Shapley value 

of coalitional games, where traffic and topological constraints are considered. 

Second, less important devices are switched off, up to the minimal sub-network 

satisfying the providers constraints. 

2 - Cost Sharing in Shortest Path Tree Games (SPTGs) 

Nayat Horozoglu, Operational Research Group, London School 

of Economics and Political Science, Houghton Street, WC2A 

2AE, London, United Kingdom, n.horozoglu@lse.ac.uk, 

Katerina Papadaki 

The SPTGs are cooperative linear programming games where the set of players 

are the nodes on an undirected graph. The aim of the players is to connect to a 

root node at minimum cost either directly or via other players who are willing 

to cooperate. In the current study, we derive structural properties of the SPTGs 

and present a polyhedral analysis of the core of the SPTGs which is nonempty. 

Furthermore, we reduce the number of inequalities needed to describe the core 

using dominance. Our motivation to define these games is to address the cost 

allocation problem in Wireless Multihop Networks. 


3 - Bargaining games with arbitration committee 

Vladimir Mazalov, Karelia Research Center of Russian 

Academy of Sciences, Institute of Appied Mathematical 

Research,Karelia Research Center, Pushkinskaya st. 11, 185910, 

Petrozavodsk, Karelia, Russian Federation, 

vmazalov@krc.karelia.ru, Julia Tokareva 

Here we consider two-person bargaining model based on arbitration procedure 

with juri. There are two players I and II. Player I makes an offer x, and II - 

an offer y. Offers are arbitrary real numbers. If x>y the players call in the juri 

which consists of some members. Each member of juri decides which offer 

must be selected. After that they vote and the majority rule works. The solutions 

of the arbitrators are presented by random variables with a distribution 

function. We obtain the equilibrium in this bargaining game and analyze the 

effect of correlation between the arbitrators. 

4 - Coalitional games on biological networks to measure 

the power of genes 

Stefano Moretti, UMR7024 - LAMSADE, CNRS - Université 

Paris-Dauphine„ Place du Maréchal de Lattre de Tassigny, 

F-75016, Paris, France, stefano.moretti@dauphine.fr, Vito 

Fragnelli, Fioravante Patrone, Stefano Bonassi 

The interpretation of gene interaction in biological networks generates the need 

for a meaningful ranking of network elements. We introduce a new approach 

using coalitional games to evaluate the centrality of genes in networks keeping 

into account genes interactions. The Shapley value for coalitional games 

is used to express the power of each gene in interaction with the others and to 

stress the centrality of certain hub genes in the regulation of biological pathways 

of interest. In addition, the new approach allows for the integration of a 

priori knowledge about genes. 

� TF-27 


8.2.06 

LOGISTICS 1 



Chair: Klaus-Christian Maassen, Mercator School of Management, 

University of Duisburg-Essen, Lotharstrasse 65, 47048, Duisburg, 

Germany, klaus-christian.maassen@uni-due.de 

1 - A multi-agent negotiation protocol for freight carrier coordination 

Gianluca De Pascale, Dipartimento di Ingegneria 

dell’Informazione, Università degli Studi di Siena, via Roma 56, 

53100, siena, Italy, luca1975@gmail.com, Alessandro Agnetis 

In our work we address the problem of coordinating several carriers, by different 

owners, in using limited and shared resources in a logistic network. 

For such purpose we devised a distributed negotiation protocol based on the 

autonomous-agent paradigm. Such protocol can be implemented as a supervised, 

distributed local search algorithm. Our protocol exhibits several desirable 

properties, such as truth enforcing, reduced information exchange, little 

computational burden and robustness to real-time changes in problem data and 

structure. 

2 - An Expanded Template for the Hosting of Contemporary 

Auctions for Freight Services 

Dimitrios Emiris, INDUSTRIAL MANAGEMENT & 

TECHNOLOGY, UNIVERSITY OF PIRAEUS, 80 KARAOLI 

& DIMITRIOU STREET, 18534, PIRAEUS, emiris@unipi.gr, 

Charis Marentakis 

The present article combines findings from previously published decision 

frameworks (Auctions Classification Ecosystem (ACE) and multi-dimensional 

auction modelling framework) for the customization of contemporary auctions 

focusing on suitability in the freight services marketplace, obeying to new 

types of constraints (temporal, spatial, etc.). The presented scheme is enriched 

to contain characteristics of auctions conducted electronically or over mobile 

networks setting the basics for the standardization of auctions in all available 

platforms beyond classic and the development of a parametric auction design 

model as an analysis tool for logistics practitioners.

3 - Pattern-Based Evacuation Planning for Urban Areas 

Sarah Bretschneider, Mercator School of Management - 

Fachbereich Betriebswirtschaft, Universität Duisburg-Essen, 

Lotharstraße 65, Gebäude LB, 47057, Duisburg, Germany, 

sarah.bretschneider@uni-due.de, Alf Kimms 

The population of an urban area may be in danger due to disasters like floods or 

chemical accidents. This requires decisions to protect lives of the affected population. 

One decision may be to evacuate the affected area. For the exceptional 

case of an evacuation an approach of the reorganization of the traffic routing 

of the endangered area is developed. In this paper a two-stage heuristic solution 

approach for a pattern-based mixed integer dynamic network flow model 

is presented. 

4 - A Fast Heuristic Approach for Large Scale Cell- 

Transmission-Based Evacuation Planning 

Klaus-Christian Maassen, Mercator School of Management, 

University of Duisburg-Essen, Lotharstrasse 65, 47048, 

Duisburg, Germany, klaus-christian.maassen@uni-due.de, Alf 

Kimms 

The basic ideas of the Cell-Transmission-Model (CTM) by Daganzo (1994) 

were picked up recently to formulate optimization models for evacuation planning, 

namely the CTEPM and ExCTEPM. These optimization models were 

able to generate high quality evacuation plans, but computational effort and requirements, 

especially in terms of real world applications were very high. In 

order to extend adaptability to much larger evacuation scenarios, a fast heuristic 

procedure for solving the ExCTEPM will be presented. In a computational 

study we will demonstrate the effectiveness of our approach. 

� TF-28 



Decentralized scheduling 



Chair: Hoong Chuin Lau, School of Information Systems, Singapore 

Management University, 80 Stamford Road, 178902 , Singapore, 

Singapore, hclau@smu.edu.sg 

1 - Decentralized Resource Allocation and Scheduling via 

Walrasian Auctions with Negotiable Agents 

Hoong Chuin Lau, School of Information Systems, Singapore 

Management University, 80 Stamford Road, 178902 , Singapore, 

Singapore, hclau@smu.edu.sg 

In this talk, we discuss decentralized resource allocation and scheduling problems. 

We present an auction approach with negotiable agents by allowing 

agents to switch their bid generation strategies within the auction process, such 

that a better system wide performance is achieved on average as compared to 

the conventional walrasian auction running with agents of fixed bid generation 

strategy. We propose a negotiation mechanism embedded in auctioneer 

to solicit bidders’ change of strategies in the process of auction. Finally we 

benchmark our approach against conventional auction subject to the real-time 

large-scale dynamic resource coordination problem to demonstrate the effectiveness 

of our approach. 

2 - Market-based Scheduling at Container Terminals 

Clemens van Dinther, Institute of Information Management and 

Systems, Karlsruhe Institute of Technology (KIT), Englerstr. 14, 

76131, Karlsruhe, Germany, clemens.vandinther@kit.edu, 

Andreas Hudelmaier 

This paper presents a distributed market-based scheduling method for a megacontainer 

port. The approach is based on work of Bertsekas (1990) and Lim 

et al. (2003). The problem is to dispatch transportation vehicles to quaycranes 

to (un)load vessels. A market-based approach can cope with ad hoc 

changes/events. We analyze the suggested mechanism and compare it to an 

alternative dispatching approach. In contrast to a classical scheduling problem, 

it is assumed that the information and optimization decisions in the container 

terminal are decentralized. 


3 - Operational Transport Planning and Scheduling 

Cees Witteveen, Software Technology, Delft University of 

Technology, Mekelweg 4, 26, Delft, Netherlands, 

C.Witteveen@tudelft.nl, Jonne Zutt 

We consider the problem of ensuring efficient operational transport planning 

and scheduling, where several vehicles share a common road infrastructure 

with limited capacity. We propose a distributed approach where infrastructural 

resources act as agents that are capable to locally schedule traffic at that 

resource, taking into account constraints (priorities) of individual requests of 

the vehicles. The outcome is a distributed and scalable, context-aware, operational 

transportation planning that is able to deal with congestion and incidents 

and outperforms traditional planning techniques. 

4 - Scalable decentralised approaches for job shop 

scheduling 

Patrick De Causmaecker, Computerscience/CODeS, Katholieke 

Universiteit Leuven, Campus Kortrijk, Etienne Sabbelaan 53, 

BE-8500, Kortrijk, Flanders, Belgium, 

Patrick.DeCausmaecker@kuleuven-kortrijk.be, Ann Nowe, 

Katja Verbeeck, Tony Wauters, Yailen Martinez 

We present two Reinforcement Learning approaches for the Parallel Machines 

Job Shop Scheduling Problem. The objective used is the minimization of the 

schedule makespan. We study two approaches, one where resources are modeled 

as intelligent agents and have to choose what operation to process next, 

and an other where operations themselves are seen as the agents that have to 

choose their mutual scheduling order. We use a value iteration method (Q- 

Learning) and a policy iteration method (Learning Automata). The results of 

both approaches improve on recently published results from the literature and 

we argue that they exhibit better scaling behaviour. We validate our approaches 

by applying them to the flexible job shop scheduling problem where operations 

can be executed on any of a number of available machines. 

� TF-29 


8.2.11 

Risk measurement and management 

Stream: Financial Modeling 


Chair: Umberto Triulzi, Department of Economic Theory and 

Quantitative Methods for Political Choices, University of Rome "La 

Sapienza ", P.le Aldo Moro 5, 00185, Rome, Italy, 

umberto.triulzi@uniroma1.it 

1 - Cds signaling 

Rita D’Ecclesia, Teoria Economica e Metodi Quantitativi per le 

Scelte politiche, Università di Roma, Piazza Aldo Moro 5, 

00185, Roma, Italy, rita.decclesia@uniroma1.it 

This paper studies the reactions of the Credit Default Swap (CDS) to rating announcements. 

Credit rating agencies make multiple announcements, some of 

which are intended to reflect the latest information available about a firm and 

others of which are intended to provide a stable signal of credit quality. Applying 

event study methodology to data on CDS, we examine whether and to what 

extent these markets respond to rating announcements. Given CDS quotes represent 

market price of credit risk. The CDS market behavior provide a crucial 

tool for risk measurement and management 

2 - Nonlinearity in electricity price series: a SETARX approach 

Carlo Lucheroni, School of Science and Technologies, 

University of Camerino, via M. delle Carceri 9, 62032, Camerino 

(MC), Italy, carlo.lucheroni@unicam.it 

Two TARX electricity price models are presented. In power systems, technical 

constraints introduce a threshold in the price formation mechanism. Below 

the threshold prices react smoothly to demand variations, above the threshold 

prices can react in a non-smooth way. A self-excited three or five regimes 

TARX, with one ARX sector set in the usual stable regime and two other sectors 

are set in unstable and metastable regimes in a specific sequence allows for 

nonlinear deviations from the stable regime, generating spikes and antispikes. 



3 - Using Euro zone sovereign debt ETFs as portfolio constituents 

Mikica Drenovak, Operational research, Statistics and IT, 

Faculty of Economics, Djure Pucara Starog 3, 34000, 

Kragujevac, Serbia, Serbia, mikicadrenovak@yahoo.com 

Performance of an ETF is dominantly affected by changes in underlying benchmark. 

European sovereign bond space is covered by wide range of indices. 

ETFs that track those indices provide very good base for building new type of 

debt portfolios. However different criteria for constructing underlying indices 

imply challenging task when attempting to recognize the most efficient solutions. 

We identify important characteristics of the main index families and also 

compare equivalent indices. The aim is to provide reference framework for 

constructing diversified sovereign debt portfolios of ETFs. 

4 - The Efficiency of Greek Pension Fund Portfolios. An 

Empirical Approach 

Alexandros Koulis, BUSINESS ADMINISTRATION, 

TECHNOLOGICAL EDUCATIONAL INSTITUTE OF 

CHALKIDA, PSAXNA, 34400, CHALKIDA, Greece, 

koulisa@otenet.gr, Christina Beneki, Maria Adam, Charalampos 

Botsaris 

Pension Funds in Greece can invest 23% in risk products in equities and mutual 

funds and the remaining 77% in fixed-income assets. In recent years, the 

investment in mutual funds is around 5% of their assets. This study is an empirical 

assessment of the performance of Greek fund managers based on Treynor- 

Mazuy model and performance measures, between the years 2000 and 2008. 

The results revealed evidence that the investment of Greek Pension Funds in 

mutual funds is not efficient due to the lack of the fund managers’ ability to 

time the market correctly or select undervalued securities. 

� TF-30 


8.2.13 

MCDM 2 




Chair: Nadia Sami, Computer Science, Cadi Ayyad University, 

Faculty of the Sciences Semlalia, Marrakech, Morocco, 

n.sami@ucam.ac.ma 

1 - Towards the integration of The Multi Criteria Decision 

Aid in the OLAP Process 

Nadia Sami, Computer Science, Cadi Ayyad University, Faculty 

of the Sciences Semlalia, Marrakech, Morocco, 

n.sami@ucam.ac.ma, Abdessadek Tikniouine 

Currently, most of the decision support systems (DSS) are based on models that 

take into account just one objective function (one criterion) that represents the 

preferences of the decision maker. The MCDA suggests methods that allow the 

aggregation of many criteria with the objective to select one or more solutions. 

In this paper, we suggest an approach which is based on the coupling of MCDA 

with the OLAP system (On Line Analytical Processing) which is an essential 

element of DSS analysis. 

2 - Multi-criteria decision analysis and optimization methods 

in supporting sustainable use of natural resources 

Tanja Myllyviita, Consumption and Production 

Centre/Environmental Performance Division, Finnish 

environment institute, Suksimestarintie 1 A 5, 80140, Joensuu, 

Finland, tanja.myllyviita@joensuu.fi, Teppo Hujala, Annika 

Kangas, Pekka Leskinen 

Various decision aid methods are used to support sustainability in natural resources 

management (NRM). To assess methods’ suitability to foster sustainability, 

24 peer-reviewed NRM case-studies were evaluated. In the evaluated 

case-studies, optimization methods rarely involved stakeholders and social sustainability 

was less processed. Multi-criteria decision analysis (MCDA) methods 

included participation of stakeholders, and social sustainability was inherent. 

It is concluded that mixed-methods MCDA provide advantages in ensuring 

sustainability of NRM. 


3 - Lessons from Two Recent Major MCDA Projects 

David Collier, Decision Science, Golder Associates, 113 York 

Road, Montpelier, BS6 5QG, Bristol, United Kingdom, 

dcollier@golder.com 

The paper contrasts two recent projects involving substantial specialist teams. 

One was based on a conventional scored and weighted multi-attribute decision 

analysis. In the other, the MADA framework was used largely without quantification 

to structure the problem and understand the main decision drivers. The 

aim was to make the best decision, transparently and with stakeholder involvement. 

We will discuss the lessons for the use of MCDA frameworks and stakeholder 

involvement. We will draw conclusions about the appropriate use of quantification, 

the structured integration of MCDA outputs with other strands of 

information and management insight, and the role of decision specialists. 

4 - The effect of customer satisfaction on financial performance: 

The case of mobile companies in Greece 

Dimitrios Drosos, Department of Business Administration, 

Graduate Technological Education Institute of Pireaus, 250, 

Thivon ave., 12244, Aigaleo, Greece, drososd@in.teipir.gr, 

Nikos Tsotsolas, Panagiotis Manolitzas, Denis Yannacopoulos 

In this paper the relationship between the customer satisfaction from the three 

mobile companies in Greece and the financial performance of these companies 

is examined. Specifically, data regarding customer satisfaction were collected 

from customers of the three service providers using a web survey which were 

analysed using MUSA methodology. Then the correlation between MUSA 

indices for each company and their corresponding financial performance ratios 

was examined by using multivariate statistics and multicriteria analysis 

approaches. 

� TF-31 


8.2.15 

Communication Network Design 



Chair: Carlo Mannino, Informatica e Sistemistica, Universita’ La 

Sapienza, Via Buonarroti 12„ 00185, Rome, Italy, 

mannino@dis.uniroma1.it 

1 - Models for FTTx network planning 

Axel Werner, Optimization, Zuse Institut Berlin (ZIB), Takustr 7, 

D-14195, Berlin-Dahlem, Germany, werner@zib.de 

Telecommunication access networks of the future are based on fiber technology. 

The rollout of the infrastructure comprises large investments and therefore 

requires thorough planning, taking economic, technical and strategic considerations 

into account. 

We present integer programming models trying to incorporate as many aspects 

of this problem as possible and show how they work together to plan costefficient 

rollout of FTTx networks. Real-world application of the approaches 

(in joint work with atesio GmbH) are discussed as well as the mathematical 

context for solving the problems. 

2 - Solving A Robust Network Loading Problem - Models, 

Inequalities and Computations 

Christian Raack, Optimization, ZIB, Takustr. 7, D-14195, Berlin, 

Germany, raack@zib.de 

In this talk we consider network loading under demand uncertainty, adapting 

the robust optimization approach of Bertsimas and Sim [2004]. The corresponding 

polyhedral demand uncertainty set provides a reasonable alternative 

to the well-known hose model. We present different ways to solve the corresponding 

robust counterpart. To enhance the performance of the MIP solver 

we use cutset inequalities and arc-residual capacity inequalities that generalize 

their deterministic counterparts. Comprehensive computational studies are 

provided based on realistic networks and live traffic measurements. 

3 - Self-Organization within LTE networks: Soft integration 

of new base 

Andreas Eisenblätter, Konrad-Zuse-Zentrum für 

Informationstechnik Berlin (ZIB), D-14195 , Berlin, 

eisenblaetter@atesio.de

Self-organization is a key requirement in the design of the 4th generation of mobile 

communication networks. Self-configuration, self-optimization, and selfhealing 

shall reduce the need for manual intervention. LTE is being standardized 

as an evolution of UMTS, and first networks are already operational. The 

European FP7 project SOCRATES develops methods for self-organization in 

the radio part of LTE networks. We present the novel concept of soft-integration 

for new base stations, which combines classical off-line planning with on-line 

adjustments once live measurements become available. 

4 - The Robust Network Loading Problem with Dynamic 

Routing 

Sara Mattia, DIS, Università di Roma, via Buonarroti 12, 00185, 

Roma, Italy, mattia@dis.uniroma1.it 

Given a graph and a set of traffic matrices, the Robust Network Loading Problem 

(RNL) consists of choosing minimum cost integer capacities for the edges, 

such that all the matrices can be routed non-simultaneously on the network. 

The routing scheme is dynamic if we can choose a (possibly) different routing 

for every matrix, it is static if the routing must be the same for all the matrices. 

The flows are unsplittable if each commodity can use only one path and 

splittable otherwise. We present the first exact branch-and-cut appoach for the 

RNL problem with dynamic routing and splittable flows. 

� TF-32 


8.2.17 

OR & Water Management 



Chair: Janne Helin, Economics, Agrifood Research Finland, 

Luutnantintie 13, 00410, Helsinki, Finland, janne.helin@iki.fi 

1 - Multicriteria Decision Support System for Surface Irrigation 

Design. Application to Ras-El-Ain Irrigation District, 

Syria 

Hanaa Darouich, Administration of Natural Resources Research, 

General Commission for Scientific Agriculture Research, Quatli 

St - Duma, Damascus, Damascus, Syrian Arab Republic, 

m_hanaa20@yahoo.com, José Manuel Gonçalves, André Muga, 

Paula Paredes, Luis S. Pereira 

The DSS SADREG for design and selection of farm surface irrigation systems 

and its application to Ras-El-Ain area, Syria, are described. It focuses 

on water savings, environmental and socio-economic issues adopting a multicriteria 

decision aid framework. The DSS allows designing and selecting an 

alternative set of solutions considering the criteria value functions, satisfaction 

pre-analysis and distance-based or ELECTRE II methods. The decision-maker 

interactively participates in the process, inputting field data, knowledge, and 

options. The advisable solutions considering several priority scenarios, and its 

trade-off between economic and environment aspects are included. 

2 - GIS-based Water Distribution Schedule for Agriculture 

in Madeira Island 

João Carlos Sousa, Intergraph Portugal, Rua Ivone Silva, 6, Piso 

4, 1050-124, Lisboa, Portugal, joao.sousa@intergraph.com, Joao 

Telhada, José Paixão 

Madeira Island has approx. 54k small individual owned parcels for which adequate 

amounts of water should be delivered, totalizing 60k hours of watering. A 

network of canals, ca. 2800 Km, is used on which different types of infrastructures 

are installed. On the other hand, a large field team is available to operate 

some of the infrastructure elements. This workforce has to be managed such 

that meets water distribution schedule. A constructive intelligence approach is 

used to determine the deployed schedule, both for owners and the field team. 

A GIS is used to handle data in an interactive mode. 

3 - Multiobjective Sustainable Water Management in Alentejo 

Region 

Vladimir Bushenkov, Department of Mathematics, University of 

Evora, 7000, Evora, Portugal, bushen@uevora.pt, Rui Fragoso, 

Carlos Marques 


The Feasible Goal Method/Interactive Decision Maps (FGM/IDM) approach 

is applied to solve a multiobjective problem of sustainable water management 

of the Odivelas irrigation system in the Alentejo region, Portugal, where water 

scarcity is real. This technique explores all water allocation combinations 

taking into account different water users interests. The Pareto frontier in the 

criteria space is visualized in the form of the tradeoff curves. The decision 

maker identifies here a preferred combination of the objective values (the goal) 

for which the computer calculates the model solution. 

4 - Manure transport model for water protection of pig 

farms 

Janne Helin, Economics, Agrifood Research Finland, 

Luutnantintie 13, 00410, Helsinki, Finland, janne.helin@iki.fi 

Scale economies give raise to a larger unit size in animal production. On the 

flip side are increased transport costs and environmental problems. This study 

attempts to quantify how much costs the environmental regulation inflicts upon 

the Finnish farmers in form of increased manure transport distances based on 

economic farm management model and GIS. We find out that partly due the 

exceptions given in the environmental subsidy requirements for manure, the 

system does not force the modeled farms to export their manure outside their 

own farm land, but the allocation within the farm is shifted. 

� TF-33 


8.2.19 

Optimality of alternative policy instruments 

for climate and energy policies 



Chair: Pekka Pirilä, Dept. of Energy Technology, Aalto University, 

POB 14100, 00076, Aalto, Espoo, Finland, pekka@pirila.fi 

1 - Adaptation and mitigation strategies in the European 

power sector 

Anne Held, Energy Policy and Energy Systems, Fraunhofer 

Institute for Systems and Innovations research, Breslauer Str. 48, 

76139, Karlsruhe, Germany, anne.held@isi.fraunhofer.de, Ulrich 

Reiter, Mario Ragwitz 

This work addresses climate change adaptation and mitigation requirements of 

the European power sector. A hybrid modelling approach is applied in order to 

investigate the future role of renewable and conventional thermal power generation 

technologies until 2050. A mitigation pathway characterised by a global 

temperature rise of 2 C by 2100 is compared to a scenario, where the power 

sector adapts to climate change assuming a temperature increase by 4 C. Finally, 

cost estimates are provided and recommendations for an optimal design 

of adaptation and mitigation policies are derived. 

2 - Choosing policy instruments under influence of uncertain 

technological change and adaptation 

Pekka Pirilä, Dept. of Energy Technology, Aalto University, 

POB 14100, 00076, Aalto, Espoo, Finland, pekka@pirila.fi 

Optimal selection of policy instruments is dependent on the dynamic behaviour 

of technological change and other adaptation mechanisms. The problem is 

analysed using small models. Further results are obtained based on a case study 

on future heating choices for the city of Helsinki. The situation is complex due 

to greatly variable and uncertain alternatives: district heating based on nuclear 

power, coal with CCS, natural gas, biomass directly or gasified, as well as individual 

heat sources like heat pumps. The policy instruments may either support 

or counteract finding the optimal solution. 

3 - Long term climate mitigation and the impact of the 

treatment of uncertainty 

Ilkka Keppo, Policy Studies, Energy research Center of the 

Netherlands, Radarweg 60, 1043 NT, Amsterdam, Netherlands, 

keppo@ecn.nl 

We tackle the issue of large uncertainty concerning the future environmental 

requirements using an energy system model and studying a set of scenarios, 

covering a range of climate targets and technology futures, from three angles; 

1) assuming perfect foresight 2) using a myopic world view and 3) using a 

stochastic programming set-up. We find that if a very stringent target is a possibility, 

it dominates the solution. However, reaching the target comes at a high 

price, indicating that e.g. adaptation measures, or even climate damages, may 

be preferable to the high mitigation costs. 



� TF-34 


8.2.23 

Generalized Convexity and Related Topics 



Chair: Gabriela Cristescu, Department of Mathematics and Computer 

Science, Aurel Vlaicu University of Arad, Aurel Vlaicu University of 

Arad, Department of Mathematics and Computer Science„ Str. 

Revolutiei, No. 77„ 310130, Arad, Arad, Romania, 

gcristescu@inext.ro 

1 - Separation theorem for nonlinear inverse images of 

convex sets 

Zsolt Pales, Institute of Mathematics, University of Debrecen, 

Egyetem tér 1, 4032, Debrecen, Hungary, pales@math.klte.hu, 

Szabolcs Bajak 

First- and higher-order necessary conditions for the local disjointness of a finite 

system of sets that are nonlinear inverse images of convex sets are presented. 

The proof is based on the characterizations of alpha-admissible and 

alpha-tangent variations to nonlinear inverse images of convex sets and a necessary 

condition for the local disjointness in terms of these variations. As an 

application, the results are used to obtain first- and higher-order necessary conditions 

of optimality in constrained optimization problems. 

2 - On Wright convexity of higher order 

Gyula Maksa, Department of Analysis, University of Debrecen, 

Institute of Mathematics, University of Debrecen, 4010, 

Debrecen, Hungary, maksa@math.klte.hu, Zsolt Pales 

In the talk, we define the concept of n-Wright convex functions and show that 

these functions can be represented as a sum of a continuous n-convex function 

and a generalized polynomial of degree at most n. In the proof, a decomposition 

result on functions having Riemann integrable higher order difference 

functions plays an important role. 

3 - Subquadratic functions 

Attila Gilanyi, Faculty of Informatics, University of Debrecen, 

Pf. 12, 4010, Debrecen, Hungary, gilanyi@math.klte.hu, Csaba 

Kezi, Katarzyna Troczka-Pawelec 

Related to the theory of convex and subadditive functions, we investigate subquadratic 

mappings. Especially, we study the lower and upper hulls of such 

functions, we prove Bernstein–Doetsch-type theorems for them and we describe 

some connections between two different notions of subquadraticity. 

4 - Abstract convexity for Convex Along Lines functions 

Giovanni Paolo Crespi, Economis and Business Management, 

University of Valle d’Aosta, Loc. Grand Chemin 73/75, 11020, 

Saint Christophe, Aosta, Italy, g.crespi@univda.it, Ivan Ginchev, 

Matteo Rocca, Alexander Rubinov 

Abstract convexity arises from monographs by Pallaschke and Rolewicz, 

Singer, Rubinov. Several studies have been recently proposed within this topic, 

mainly due to closed relation and applications to global optimization. Still, 

many problems are open. Among them the characterization of functions defined 

as the upper envelope of min-type functions. It has already been proved 

that these functions are closely related to convex along rays functions. However, 

some refinements allow us to prove also Convex Along Lines functions 

(CAL) are abstract convex. 

� TF-35 


6.2.46 

Recent advances in mixed-integer nonlinear 

and global optimization 

Stream: Mixed-Integer Non Linear Programming 


Chair: Chris Floudas, Chemical Engineering, Princeton University, 

Olden St., 8540, Princeton, NJ, United States, 

floudas@titan.princeton.edu 


1 - Synthesis of chromatographic protein purification processes 

using optimisation techniques 

Eleftheria Polykarpou, Biochemical Engineering, UCL, United 

Kingdom, e.polykarpou@ucl.ac.uk, Paul Dalby, Lazaros 

Papageorgiou 

This paper presents mixed integer optimisation techniques for the synthesis 

of downstream purification processes by determining the minimum number of 

chromatographic steps in the optimum sequence for given purity and recovery 

levels. Alternative mixed integer linear and non-linear programming (MILP/ 

MINLP) models will be presented and discussed. The applicability of these 

models is demonstrated by examples that rely on experimental data together 

with comparative results. 

2 - Global Optimization of Signomial Geometric Programming 

Problems 

Pedro Castro, UMOSE, LNEG, 1649-038, Lisbon, Portugal, 

pedro.castro@ineti.pt, João Teles, Henrique Matos 

This paper discusses some power-based transformation techniques that are especially 

useful when solving signomial optimization problems (SOP). A signomial 

MINLP problem is first transformed into a new reformulated problem containing 

artificial non-negative variables. By applying a discretization scheme 

with multi-parametric elements an upper bounding convex MILP problem can 

be derived and solved to global optimization with standard methods to overestimate 

the original SOP within a tolerable error. Numerical examples are 

presented to demonstrate the effectiveness of the proposed method. 

3 - New formulations for the mixture design problem 

Claire Adjiman, Chemical Engineering, Imperial College 

London, Department of Chemical Engineering, Centre for 

Process Systems Engineering, SW7 2AZ, London, United 

Kingdom, c.adjiman@imperial.ac.uk 

The design of mixtures, e.g. optimal product formulations in the pharmaceutical 

and personal care industries, is an important and challenging problem. 

Even when the components of the mixture are known a priori, the problem is 

highly combinatorial and nonlinear, due to the complex relationships between 

composition and physical properties. This has hampered the formulation and 

solution of the design problem. In this paper, we present novel formulations 

of the mixture design problem based on Generalized Disjunctive Programming 

and demonstrate the effectiveness of this approach on case studies. 

4 - A Mixed-integer Quadratic Approximation Algorithm for 

the Solution of Multiparametric Mixed-integer Nonlinear 

Programming Problems 

Efstratios Pistikopoulos, Chemical Engineering and Chemical 

Technology, Imperial College London, SW72AZ, London, 

United Kingdom, e.pistikopoulos@imperial.ac.uk, Luis 

Dominguez 

In this work we present a novel algorithm for the solution of multiparamet ric 

mixed-integer nonlinear programming (mp-MINLP) problems. Similar to the 

algorithm introduced earlier by Dua and Pistikopoulos (1999), the algorithm 

presented here is based on a decomposition strategy where a sequence of multiparametric 

nonlinear programming problems (primal mp-NLPs) and deterministic 

mixed-integer nonlinear programming problems (master MINLPs) are 

solved. In this work, we alleviate some of the limitations of the algorithm presented 

in Dua and Pistikopoulos (1999) by addressing the primal subproblems 

via novel multiparametric nonlinear programming techniques. The proposed 

algorithm solves the sequence of primal subproblems via a multi-parametric 

quadratic approximation (mp-QA) algorithm and the corresponding MINLP 

subproblems via a MINLP solver. Finally, we present numerical examples 

which demonstrate the computational advantages of the mp-MIQA algorithm 

versus previous ones. 

� TF-36 


3.1.05 

Robust Optimization 

Stream: OR and Real Implementations 


Chair: Daniel Kuhn, Department of Computing, Imperial College 

London, United Kingdom, dkuhn@doc.ic.ac.uk

1 - Multistage stochastic portfolio optimization in deregulated 

electricity markets using linear decision rules 

Paula Rocha, Computing, Imperial College London, SE7 2AZ, 

London, United Kingdom, 

paula.martins-da-silva-rocha08@imperial.ac.uk, Daniel Kuhn 

We present a multistage stochastic mean-variance optimization model for the 

management of electricity portfolios from the viewpoint of a price-taking retailer. 

To reduce computational complexity, we perform two approximations: 

stage-aggregation and linear decision rules. The latter consists of restricting the 

set of decision rules to those affine in the history of the risk factors. When applied 

to mean-variance optimization models, it leads to convex quadratic programs. 

Since their size grows only polynomially with the number of stages, 

problems with many stages can be efficiently solved. 

2 - Linearly Adjustable International Portfolio Optimization 

Raquel Fonseca, Department of Computing, Imperial College, 

180 Queen’s Gate, SW7 2AZ, London, 

r.fonseca@imperial.ac.uk, Daniel Kuhn, Berc Rustem 

We present an approach to multiperiod international portfolio optimization 

based on the imposition of a linear structure on the recourse decisions. Multiperiod 

decision problems have traditionally been formulated as stochastic programs. 

These however can become severely intractable as the number of stages 

increases. By restricting the space of decision policies to linear rules, we obtain 

a conservative tractable approximation to the original problem. Local asset and 

currency returns are modelled separately, which allows for hedging policies 

regarding the currency risk to be implemented. 

3 - Interdicting a project to develop nuclear weapons 

Wolfram Wiesemann, Department of Computing, Imperial 

College of Science, Technology & Medicine, 180 Queen’s Gate, 

SW7 2BZ, London, United Kingdom, 

wolfram.wiesemann@gmail.com, Daniel Kuhn, Berc Rustem 

We study a two-player game in which one player (the "proliferator") aims to develop 

a batch of nuclear weapons as quickly as possible, while the other player 

(the "interdictor") seeks to delay the completion of this project. The game is an 

instance of an "interdiction game", a problem class with applications in managerial 

and military decision-making. To date, interdiction games are solved 

as generalized semi-infinite optimization problems, which are difficult to solve 

in practice. We develop a novel reformulation based on robust optimization 

principles. 

� TF-37 


3.1.09 

’MCDA & uncertainty 

Stream: MCDA I: New Approaches and Applications 


Chair: Sarah Ben Amor, Telfer School of Management, University of 

Ottawa, 55 Laurier E (7123), K1N6N5, Ottawa, Ontario, Canada, 

benamor@telfer.uottawa.ca 

1 - A general framework for integrating imperfect information 

in the PROMETHEE methods 

Sarah Ben Amor, Telfer School of Management, University of 


benamor@telfer.uottawa.ca, Bertrand Mareschal 

The PROMETHEE methods are well-known in the field of MCDA. One current 

limit of these methods is that they usually assume that perfect information 

is available. In practice however the evaluation of actions on some criteria may 

be imprecise or uncertain. In this paper we propose a general framework for 

integrating such imperfect information. Different models are considered including 

probability, possibility, fuzzy logic and evidence theory, as imperfect 

information can have different causes (uncertainty, imprecision, ...) Missing 

data is also included as a special case. 

2 - Pre-posterior analysis for additionnal information in a 

multiple criteria context with information imperfections 

Sarah Ben Amor, Telfer School of Management, University of 


benamor@telfer.uottawa.ca, Kazimierz Zaras, Jean-Marc Martel 


Multiple criteria decision aid situations are usually facing different types of 

information imperfections (uncertainty, imprecision . . . ) A unified procedure 

aiming at reducing these imperfections was designed to allow for processing 

additional information in such a context. It is based on the Bayesian decision 

model where prior and posterior analyses have been achieved leading to prior 

and posterior global preference relational systems. Pre-posterior analysis will 

be addressed in this paper for a pre-assessment of the resources that can be 

allocated to such additional information. 

3 - Ranking medical equipment for protection against 

earthquakes: a multicriteria approach 

Rui Oliveira, CESUR/IST, Technical University of Lisbon, Av. 

Rovisco Pais, 1049-001, Lisbon, Portugal, roliv@ist.utl.pt, 

Miguel Snow 

It is vital that hospitals and the medical equipments they accommodate remain 

operational in case of earthquake occurrence so that the affected population 

receives proper medical care in that catastrophic scenario. In order to 

define appropriate strategies for medical equipment protection against earthquakes, 

a multicriteria model was developed, allocating a priority index to each 

piece of equipment. The model considers as fundamental points of view the 

equipment’s price, seismic vulnerability and clinical relevance in that catastrophic 

scenario. The process associated with the development of the model 

is described, for which the Macbeth approach was extensively used to elicit 

value functions and weighting constants. Results and conclusions are presented 

herein. 

� TF-40 


6.2.52 

Operation Planning and Control in 

Container Terminals 



Chair: Kap Hwan Kim, Industrial Engineering, Pusan National 

University, Jangjeon-dong, Gunjeong-gu, 609-735, Busan, Korea, 

Republic Of, kapkim@pusan.ac.kr 

Chair: Loo Hay Lee, Industrial and Systems Engineering, National 

University of Singapore, Singapore, iseleelh@nus.edu.sg 

1 - Dynamic Yard Storage Strategy for Container Terminal 

Loo Hay Lee, Industrial & Systems Engineering, National 

University of Singapore, 10 Kent Ridge Crescent, 119260, 

Singapore, Singapore, iseleelh@nus.edu.sg, Ek Peng Chew, 

Xinjia Jiang 

In this talk, we will present a storage yard management problem in a transshipment 

hub where the loading and unloading activities are both concentrated and 

need to be considered at the same time. A dynamic yard template concept is 

proposed to reuse the storage space for different vessels during different shifts 

in order to make full use of the storage space. In addition, the consignment 

strategy is used to reduce the number of reshuffles and the high-low workload 

balancing protocol is used to reduce the traffic congestion of prime movers. 

2 - Dispatching Vehicles Supporting Multi-lift Operations 

of Quay Cranes 

Kap Hwan Kim, Industrial Engineering, Pusan National 

University, Jangjeon-dong, Gunjeong-gu, 609-735, Busan, 

Korea, Republic Of, kapkim@pusan.ac.kr, Vu Duc Nguyen 

To improve the ship operation in container terminals, it is important to schedule 

different types of handling equipment to operate synchronously. For example, 

a vehicle with container receiving and lifting capabilities is used to transport 

containers from a storage yard to a vessel and vice versa, while a multi-lift 

quay crane (QC) can handle up to three 40-ft containers simultaneously. This 

paper discusses a method in which vehicles are assigned to containers to support 

such multi-lifts of QCs by using information about the locations and times 

of deliveries. A mixed-integer programming model is introduced to optimally 

assign delivery tasks to vehicles. This model considers the constraint imposed 

by the limited buffer space under each QC. A procedure for converting bufferspace 

constraints into time window constraints and a heuristic algorithm for 

overcoming the excessive computational time required for solving the mathematical 

model are suggested. A numerical experiment is conducted to compare 

the objective values and computational times of the heuristic algorithm with 

those of the optimizing method to evaluate the performance of the heuristic 

algorithm. 



3 - Online Rules for Container Stacking at an import Terminal 

Eelco van Asperen, Center for Maritime Economics & Logistics, 

Erasmus University Rotterdam, Room H10-10, PObox 1738, 

3000 DR , Rotterdam, Netherlands, vanasperen@ese.eur.nl, 

Rommert Dekker 

Container stacking rules are an important factor in container terminal efficiency. 

In this paper, we investigate two concepts to increase efficiency and 

compare them to several benchmark algorithms, using a discrete-event simulation 

tool. The first concept is to use knowledge about container departure times, 

in order to limit the number of reshuffles. We stack containers leaving shortly 

before each other on top of each other. The second concept is the tradeoff between 

stacking further away in the terminal vs. stacking close to the exit points 

and accepting more reshuffles. It is concluded that even the use of imperfect or 

imprecise departure time information leads to significant improvements in efficiency. 

Minimizing the difference in departure times proved to be important. 

It was also found that the tradeoff between stacking further away in the terminal 

vs. stacking close by the exit points and accepting more reshuffles leads to 

improvements over the benchmark. 

4 - Minimizing the makespan of container storages and retrievals 

Amir Hossein Gharehgozli, Rotterdam School of Management, 

Erasmus University Rotterdam, Postbus 1738, 3000 DR, 

Rotterdam, Netherlands, agharehgozli@rsm.nl, Yugang Yu, René 

de Koster, Jan Tijmen Udding 

We study in which sequence to finish a list of container storage and retrieval 

jobs for a container stack to minimize the total makespan. The container stack 

consisting of multiple rows, bays and tiers is managed by an automated yard 

gantry. The storage positions of the containers located at different Input/output 

(I/O) points are given. The I/O points are located at the both seaside and landside 

of each row. The retrieved containers can be delivered to any of the I/O 

points at the seaside or landside. We aim to develop an algorithm for efficiently 

and near-optimally sequencing the jobs. 

� TF-42 


3.1.07 

Decision Making 2 



Chair: Jessika Grunwald, FWW / Entreprenuership, University 

Magdeburg, Universitätsplatz 2, G23 / R200, 39106, Magdeburg, 

Germany, jessika.grunwald@ovgu.de 

1 - Opportunities for decision support in telematic acute 

stroke care 

Jessika Grunwald, FWW / Entreprenuership, University 

Magdeburg, Universitätsplatz 2, G23 / R200, 39106, Magdeburg, 

Germany, jessika.grunwald@ovgu.de, Peter Knüppel 

Telemedical solutions are increasingly used to provide neurological expertise in 

networks of small primary care hospitals and stroke centers around the world. 

Electronic clinical decision support can hereby support physicians and optimize 

resource allocation. By using process analysis we identified ways of how Decision 

Support Systems can facilitate the acute stroke care process: 1) workflowbased 

control 2) support for differential diagnosis 3) outcome prognosis. On 

this basis we introduce initial modeling approaches as graphs, Bayesian networks 

and logistic regression equations. 

2 - The Analysis of the Innovation Performance of the 

Countries through Bayesian Causal Map 


Fusun Ulengin, Department of Industrial Engineering, Dogus 

University, Zeamet Sk„ Acıbadem, Kadıköy, 34722, Istanbul, 

Turkey, fulengin@dogus.edu.tr, Sule Onsel, Gündüz Ulusoy, 

Emel Aktas, Özgür Kabak 

In this study, the innovation performance of the countries is analyzed using 

a Bayesian Causal Map (BCM). Initially, a workshop is conducted to revise 

the innovation attributes used by the European Innovation Scoreboard (EIS). 

Similar to the EIS, these revised attributes are categorized as "innovation enhancers’, 

"knowledge creation’, "innovation and entrepreneurship’, "applications’, 

and "intellectual property’. Causal map is derived based on a second 

workshop. Finally, integrating the probabilities, BCM is developed, thus, a 

road map is provided for the policy makers to develop strategies for improving 

the innovation level of their country. 

3 - Tactical mission planning 

Nils-Hassan Quttineh, Department of Mathematics, Linköping 

University, SE-581 83, Linköping, Sweden, niqut@mai.liu.se, 

Kristian Lundberg, Kaj Holmberg 

We present a model for tactical mission planning, where the objective is to define 

a flight plan for aircrafts toward a target, maximizing the probability of mission 

success. Multiple units might need to approach the target simultaneously, 

adding the complexity of coordination to our problem. The surroundings are 

an important input to such problems, where features of nature (flat areas, high 

mountains) have significant impact on solutions. A decomposition approach is 

discussed, and by discretization of the environment we derive a network model 

to describe possible routes towards the target. 

4 - Expanding the Process of Knowledge Discovery in 

Databases to advance the quality of prediction 

Claudia Koschtial, Professur für ABWL, TU Bergakademie 

Freiberg, Lessingstr. 45, 09599, Freiberg, Germany, 

claudia.koschtial@web.de, Carsten Felden 

With the goal of providing a reliable basis for decisions, it can be necessary 

to predict future developments. Knowledge Discovery in Databases (KDD) 

is a method by which patterns discovered in historical data are used to predict 

future values. This implicates the assumption, that there are no fundamental 

differences in basic parameters. But this can not supposed. For example 

there are fundamental changes in demographics effecting societies and of 

course economies. The article provides an extension of the established model 

of KDD by integrating future data in the process of KDD and shows by an 

intense literature review that this was not done before. 

� TF-43 


8.2.02 

Simplicial methods in Global Optimization 



Chair: Eligius M.T. Hendrix, Computer Architecture, Universidad de 

Málaga, Campus de Teatinos, ETSI 2.2.28, 29017, Malaga, Spain, 

eligius.hendrix@wur.nl 

1 - Global optimization with simplicial partitions and Lipschitz 

bounds 

Julius Zilinskas, Institute of Mathematics and Informatics, LT 

08663, Vilnius, Lithuania, julius.zilinskas@ktl.mii.lt, Antanas 

Zilinskas 

Various versions of global optimization algorithms have been proposed using 

Lipschitz model of objective functions. The efficiency of such methods crucially 

depends on estimation of Lipschitz constant. In the present work an 

algorithm is proposed based on branch and bound approach with simplicial 

partitioning of the feasible region and adaptive estimate of Lipschitz constant 

using radial basis function interpolation. 

2 - The minimum volume simplex problem applied to spectral 

unmixing 

Eligius M.T. Hendrix, Computer Architecture, Universidad de 

Málaga, Campus de Teatinos, ETSI 2.2.28, 29017, Malaga, 

Spain, eligius.hendrix@wur.nl, I. Garcia 

We describe the minimum volume enclosing simplex problem, known to be a 

Global Optimization problem and illustrate its multimodality. The problem has 

been used as a basis to estimate so-called end-members and abundance fractions 

in unmixing spectral data from remotely sensed hyperspectral sensors. 

We explore the possibility of a new estimation algorithm using the minimum 

volume enclosing simplex problem. We investigate its behaviour numerically 

on designed instances, comparing its outcomes with a maximum volume enclosed 

simplex approach which is used frequently in spectral unmixing.

3 - Global optimization of Lipschitz differentiable functions 

Dmitri Kvasov, Department of Electronics, Computer Science 

and Systems, University of Calabria, DEIS, Via P. Bucci, Cubo 

42C, I-87036, Rende (CS), Italy, kvadim@si.deis.unical.it, 

Yaroslav Sergeyev 

Global optimization problems are considered where the objective functions 

are multidimensional, black-box, and their first derivatives satisfy the Lipschitz 

condition with unknown Lipschitz constants over a hyperinterval. A new 

method for solving such problems is presented and discussed from both the 

theoretical and numerical viewpoints. The proposed algorithm is based on efficient 

diagonal partitions and uses smooth auxiliary functions to approximate 

the objective function behavior at different parts of the search domain. 

4 - Solving the Minimum Sum-of-Squares Clustering Problem 

by Hyperbolic Smoothing and Partition into Boundary 

and Gravitational Regions 

Adilson Elias Xavier, Graduate School of Systems Engineering 

and Computer Scinces, Federal University of Rio de Janeiro, 

P.O. Box 68511, Ilha do Fundão - Centro Tecnologia - H319, 

21941-972, Rio de Janeiro, RJ, Brazil, adilson@cos.ufrj.br, 

Vinicius Layter Xavier 

The Hyperbolic Smoothing Clustering Method adopts a smoothing strategy that 

solves a sequence of differentiable unconstrained optimization problems. The 

paper presents a new idea: the partition of the set of observations into two non 

overlapping parts. The first set, named boundary band zone, corresponds to 

the observation points relatively close to two or more centroids. The second 

set, named gravitational points, corresponds to observation points significantly 

closer to a single centroid. The combination of the two methodologies drastically 

simplify the computational tasks. 

� TF-44 


8.2.03 

Vector and Set-Valued Optimization I 

Stream: Vector and Set-Valued Optimization 


Chair: Elena Molho, Dipartimento di Economia Politica e Metodi 

Quantitativi, Università di Pavia, Via San Felice 5, 27100, Pavia, 

Italy, molhoe@eco.unipv.it 

1 - Optimality conditions in scalar and vector optimization 

using a new second order directional derivative 

Bienvenido Jiménez, Departamento de Matemática Aplicada, 

UNED, Calle Juan del Rosal, 12, 28040, Madrid, Spain, 

bjimenez@ind.uned.es, César Gutiérrez, Vicente Novo 

We introduce a new second order directional derivative and study some properties. 

Using this derivative and the parabolic second order derivative we provide 

second order necessary and sufficient optimality conditions for a general scalar 

optimization problem through the asymptotic and parabolic second order tangent 

sets to the feasible set. By means of a suitable scalarization, these results 

are applied to a general vector optimization problem obtaining second order 

optimality conditions that generalize the differentiable case. 

2 - Stopping criteria for a general model of genetic algorithm 

with applications to multiobjective optimization 

Marcin Studniarski, Faculty of Mathematics and Computer 

Science, University of Lodz, ul. S. Banacha 22, 90-238, Lodz, 

marstud@math.uni.lodz.pl 

We consider a general Markov chain model of genetic algorithm; see [1]. We 

establish an upper bound for the number of iterations which must be executed 

in order to find an optimal solution with a prescribed probability. By combining 

this upper bound with the results of [2], we obtain some stopping criteria 

for multiobjective evolutionary algorithms. [1] C.R. Reeves, J.E. Rowe, 

Genetic Algorithms - Principles and Perspectives: A Guide to GA Theory, 

Kluwer, 2003. [2] G. Rudolph, A. Agapie, Convergence properties of some 

multi-objective evolutionary algorithms, Proc. CEC 2000 (2), 1010-1016. 

3 - Efficiency in V-KT- pseudoinvex control problems 

Manuel Arana-Jiménez, Estadistica e Invesitigacion Operativa, 

University of Cadiz, C/Chile, 1, 11002, Jerez de la Frontera, 


Cadiz, Spain, manuel.arana@uca.es, Gabriel Ruiz-Garzón, 

Antonio Rufián-Lizana, Rafaela Osuna-Gómez 

Control problems are often applied to engineering problems. In order to study 

their optimal solutions, KT-invexity and FJ-invexity have been introduced. Our 

aim is to generalize these properties and classes of functions and functionals to 

multiobjective control problems, for which we comment a V-KT- pseudoinvex 

control problem. This vector control problem is characterized so that a Kuhn- 

Tucker point is an efficient solution. Furthermore, this generalizes recently 

obtained optimality results of multiobjective mathematical programming problems. 

4 - An interior point method for linearly constrained multiobjective 

optimization without apriori scalarization 

Elena Molho, Dipartimento di Economia Politica e Metodi 

Quantitativi, Università di Pavia, Via San Felice 5, 27100, Pavia, 

Italy, molhoe@eco.unipv.it, Enrico Miglierina, Maria Cristina 

Recchioni 

An algorithm to find the critical points of a linearly constrained multiobjective 

optimization problem is developed without any "a priori’ scalarization. It is an 

interior point method based on a dynamical system defined by a vector field 

of descent directions that play the role of the projected gradient-like directions 

in the feasible region. The limit points of the solutions of this system satisfy 

the Karush-Kuhn-Tucker (KKT) first order necessary optimality conditions for 

the linearly constrained multiobjective optimization problem. Some numerical 

results on test problems are provided. 

� TF-46 


8.2.14 

Probabilistic Constrained Stochastic 

Programming 



Chair: Andras Prekopa, RUTCOR, Rutgers University, 640 

Barholomew Road, 08854-8003, Piscataway, New Jersey, United 

States, prekopa@rutcor.rutgers.edu 

1 - Stochastic Network Design with Normally Distributed 

Random Demands and Arc Capacities. 

Olga Myndyuk, Rutcor, Rutgers Center for Operations Research, 

Rutgers, The State University of New Jersey, 640 Bartholomew 

Rd, 08854, Piscataway, NJ, United States, 

olgamyn@eden.rutgers.edu, Andras Prekopa 

Networks are considered where some of the demands at the nodes and some 

of the arc capacities are random variables that have joint normal distribution. 

The Gale-Hoffman and Prekopa-Boros feasibilities provide us with a system 

of inequalities. A network design problem is formulated where a probabilistic 

constraint assures the existence of a feasible flow by a large probability. The 

numerical solution uses the method of Prekopa that combines a cutting plane 

method with supporting hyperplane method. 

2 - Uniform Quasi-Concavity in Probabilistic Constrained 

Programming 

Kunikazu Yoda, Rutcor, Rutgers Center for Operations Research, 

Rutgers University, 640 Bartholomew Rd, 08854, Piscataway, 

NJ, United States, kyoda@rutcor.rutgers.edu, Andras Prekopa 

A probabilistic constrained stochastic programming problem is considered, 

where the underlying problem has linear constraints with random technology 

matrix, the rows of which are assumed to be independent and normally distributed.The 

constraining function is a product of as many functins as the number 

of rows in the matrix. It is shown that the product is quasi-concave iff the 

factors are uniformly quasi-concave which implies that the covariance matrices 

are constant multiples of each other. Application to portfolio construction will 

be presented. 

3 - A new method for the valuation of Bermuda options using 

univariate numerical integration and bounding. 

Mariya Naumova, Rutcor, Rutgers Center for Operations 

Research, Rutgers University, 640 Bartholomew Road, 08854, 

Piscataway, NJ, United States, mnaumova@rci.rutgers.edu 



Prékopa and Szántai (2010) presented a dynamic programming formulation to 

compute the value of the Bermuda option with dividend. The numerical solution 

of the equation is, however, computationally intensive because it requires 

multivariate normal probability calculation. In this paper we replace the multivariate 

integration by univariate to obtain sharp lower and upper bounds for the 

values at each step. The accuracy of the new method is as good as that of the 

former one as we show it by numerical examples. 

4 - Solution of a Stochastic Network Design Problem with 

Probabilistic Constraint and Discrete Random Variables 

Andras Prekopa, RUTCOR, Rutgers University, 640 

Barholomew Road, 08854-8003, Piscataway, New Jersey, United 

States, prekopa@rutcor.rutgers.edu, Merve Unuvar 

Stochastic single commodity network design problem is formulated and solved, 

where a probabilistic constraint takes care for reliability, i.e., the probability 

that all demands are met. Demands at the nodes are discretized and approximated 

by bi-variate and multivariate distributions.The application of the Gale- 

Hoffman, Prekopa-Boros and further, more recent theorems allow fast enumeration 

of the p-efficient points. Two different algorithms are used to solve the 

problem; the generation of the p-efficient points is part of the algorithm. Numerical 

examples will be presented. 

� TF-47 


8.2.16 

Optimization in Water Systems II 

Stream: OR in Water Management 


Chair: Halil Önder, Civil Engineering, Middle East Technical 

University, Middle East Technical University, Department of Civil 

Engineering, 06531, Ankara, Turkey, onde@metu.edu.tr 

Chair: Maria Cunha, Civil Engineering, University of Coimbra, Polo 

2, 3030-290, Coimbra, Portugal, mccunha@dec.uc.pt 

1 - Optism — a decision model for the optimal operation of 

multisource water supply systems 

João Vieira, Civil Engineering, University of Coimbra, 

LHRHA-Departamento de Engenharia Civil, Pólo 

II-Universidade de Coimbra, 3030-290, Coimbra, Portugal, 

jvieira@dec.uc.pt, Maria Cunha, Luís Nunes, José Monteiro, 

Luís Ribeiro, Tibor Stigter, João Nascimento, Helena Lucas 

The OPTISM model was designed for determining the optimal operation of 

multisource supply systems dependent on surface water and groundwater. The 

planning objectives defined for the interventions are the reduction of operating 

costs, the satisfaction of the demand in exclusiveness by the water utility, and 

the supply of water with the appropriate quality. The model is highly nonlinear 

with discontinuous derivatives and is solved with GAMS/MINOS. The application 

of the model shows its usefulness in handling complex systems and in 

improving the conjunctive use of the different water sources. 

2 - Agricultural adaptation in a changing climate: optimising 

multiple objectives 

Olivier Crespo, CSAG, Dept. Environmental & Geographical 

Science, University of Cape Town, Private Bag X3, 7701, 

Rondebosch, Western Cape, South Africa, 

olivier.crespo@csag.uct.ac.za, Mark Tadross, Peter Thorburn 

Income, sustainability or resources are among multiple objectives to take into 

account for adaptation to our changing climate. We present a multiobjective 

optimization approach for assessing agricultural adaptations to climate change 

in two study cases; the introduction of a climate-resilient crop in Australia and 

adapting water supply in South Africa. We use the APSIM crop model to simulate, 

in each case, the agricultural response under future climate conditions. 

The expected outcomes and usefulness of the multiobjective approach for exploring 

agricultural adaptation will be presented. 

3 - Decision Support for Wastewater Systems Planning at 

Regional Level 


João Zeferino, DEC, FCTUC, Coimbra, Portugal, 

zeferino@dec.uc.pt, Antonio Antunes, Maria Cunha 

A decision support system for regional wastewater systems planning is presented. 

DSS is based on an optimization model that determines the best configuration 

for the system needed to drain the wastewater generated by the population 

centers of a region. Quality standards defined for the receiving water bodies 

is taking into account. A hybrid simulated annealing—local improvement 

algorithm is used for solving the model. This algorithm has been investigated 

with regard to computation effort and solution quality. The model is applied to 

a case study designed to mimic a real-world problem. 

� TF-48 


8.2.04 

Nonlinear Optimization and Applications 2 

Stream: Nonlinear Programming [c] 


Chair: Edite M.G.P. Fernandes, Production and Systems, University 

of Minho, School of Engineering, Campus de Gualtar, 4710-057, 

Braga, Portugal, emgpf@dps.uminho.pt 

1 - Measure theory approach in sliding mode control for 

nonlinear systems 

Mohammad Reza Zarrabi, applied mathematics, ferdowsi, 

mashhad-ferdowsi univesity-applied mathematic departmant, 

098, mashhad, Iran, Islamic Republic Of, mo.za870@gmail.com, 

Mohammad Hadi Farahi 

A new sliding mode control (SMC) design approach using measure theory and 

Lyapunov functional candidate is presented for nonlinear control problems. A 

Lyapunov function is supposed for designing a sliding surface (SS). In fact the 

problem that is considered is as follows. A state trajectory from a given initial 

point reaches into a given point on a sliding surface in the minimum time, and 

then tends to the origin (equilibrium point) along the sliding surface. A measure 

theory approach with embedding process is used to solve such a problem in 

two phases. In the first phase, after designing an appropriate SS by a suggested 

Lyapunove function, and using measure theory, an embedding is constructed to 

solve a time optimal control problem such that the system trajectory reaches a 

SS in minimum time, then in the second phase, using SS, a control is designed 

such that the system trajectory tends to the origin along the SS. A numerical 

example is presented to illustrate the effectiveness of the proposed method. 

2 - A new aproach using interior point method with relaxation 

based upon modified logarithmic barrier 

Mayk Coelho, UNICAMP, Brazil, mayk@cose.fee.unicamp.br, 

Aurelio Oliveira, Anesio Santos 

An modified relaxed logarithmic interior point method for the optimal DC 

power flow solution of a hydrothermal power system. The dual nonnegative 

variables are updated via complementarity with relaxed logarithmic barrier, 

while the free dual variables are updated via Newton’s method. The matrix 

structure results in a constant Jacobian therefore leading very fast iterations in 

comparison with traditional approaches. A case study using the IEEE30 test 

system is performed.

Wednesday, 9:00-10:20 

� WA-02 


3.2.14 







1 - A Combinatorial Approach to Secondary Spectrum 

Auctions 

Berthold Vöcking, Department of Computer Science, RWTH 

Aachen University, Templergraben 55, 52056, Aachen, Germany, 

voecking@cs.rwth-aachen.de 

On the secondary spectrum market, licenses are valid only for local regions. 

Spectrum allocations have to take into account interferences between users in 

neighboring regions. We show that problem formulations for secondary spectrum 

auctions in well established, but rather technical models for wireless communication 

like, e.g., the protocol or the physical model, can be represented in 

terms of a plain combinatorial model in which interference conditions are described 

by a conflict graph. The conflict graphs obtained from the wireless 

models have an interesting combinatorial property: The so-called inductive independence 

can be bounded by a slowly growing function. We investigate how 

this property can be exploited for the design of approximation algorithms for 

efficient spectrum allocations. *Joint work with Martin Hoefer and Thomas 

Kesselheim 

� WA-04 


3.2.13 

Resource sizing and allocation 






Chair: Simge Yelkenci, Department of Industrial Engineering, Dokuz 

Eylul University - The Graduate School of Natural and Applied 

Sciences, Buca Tinaztepe Campus / IZMIR - TURKEY, 35160 , 

Izmir, Turkey, simge.yelkenci@deu.edu.tr 

1 - Buffer allocation in unreliable production lines using 

adaptive tabu search approach 

Leyla Demir, Industrial Engineering, Dokuz Eylul University, 

Tınaztepe Campus - Buca, 35160, Izmir, Turkey, 

leyla.demir@deu.edu.tr, Semra Tunali 

The buffer allocation problem, i.e. how much buffer storage to allow and where 

to place it within the line, is a nonlinear integer programming problem. In this 

study, buffer allocation problem is solved for unreliable production lines. The 

objective is to maximize the throughput of the line, i.e. production rate. To 

evaluate the throughput of the line the decomposition method is employed. An 

adaptive tabu search approach is proposed to optimize the buffer sizes for each 

location. The performance of the proposed approach is demonstrated using 

existing benchmark problems. 

2 - Investigating Multi-Objective Buffer Allocation Problem 

in Unreliable Production Systems Using a Modified Artificial 

Immune System Algorithm 

Yamani Massim, Institut Charles Delaunay (CNRS FRE 2848), 

LOSI„ University of technology of Troyes, University of 

Technology of Troyes, 12 rue Marie Curie„ 10000, Troyes, 

France, yamanimassim@yahoo.fr, Farouk Yalaoui, Lionel 

Amodeo, Alice Yalaoui, Abdelkader Zeblah 

EURO 24 - Lisbon 2010 WA-05 

In this paper we investigate the multiobjective buffer allocation problem 

(MOBAP) in production systems under space constraint. The aim of the work 

is to determine the Pareto fronts related to the conflicting throughput and work 

in process inventory objectives which represent important measures of the production 

systems economic efficiency. A modified immune algorithm incorporating 

a solution upgrading heuristic has been developed to generate Pareto 

fronts for transfer lines and systems incorporating parallel machines. Outline 

of the algorithm as well as computational results are presented. 

3 - A hybrid method combining genetic algorithm and simulation 

for buffer size allocation in a real manufacturing 

system 

Simge Yelkenci, Department of Industrial Engineering, Dokuz 

Eylul University - The Graduate School of Natural and Applied 

Sciences, Buca Tinaztepe Campus / IZMIR - TURKEY, 35160 , 

Izmir, Turkey, simge.yelkenci@deu.edu.tr, Semra Tunali 

This study presents a simulation-based genetic algorithm approach to find optimal 

buffer sizes.The objective is to improve the capacity of production line 

by implementing proposed hybrid approach.First,a detailed stochastic and dynamic 

simulation model of the line is developed to identify bottleneck machines.After 

that GA-based simulation optimization approach is employed to 

decide how to allocate buffers to the bottleneck machines identified so that 

throughput of the line can be maximized.The proposed method is tested in the 

real manufacturing environment.Empirical results show promise for the practical 

application of the proposed methodology. 

� WA-05 


3.2.16 

Networks 



Chair: Roberto Battiti, DIT - Dipartimento di Informatica e 

Telecomunicazioni, Universita’ di Trento, Via Sommarive, 14, 38100, 

Trento, Italy, battiti@dit.unitn.it 

Chair: João Claro, INESC Porto, Faculty of Engineering, University 

of Porto, Portugal, jclaro@fe.up.pt 

1 - Genetic Algorithm approach for network intrusion detection 

using mobile agent 

Khaled Sellami, Computer science, University of Bejaia, Route 

de targua ouzemour , Université de Bejaia, Algérie, 06000, 

Bejaia, Bejaia, Algeria, skhaled36@yahoo.fr, Mohamed 

Ahmed-nacer, Lynda Sellami 

Due to the increase in access of malicious data over the internet resources, intrusions 

Detection Systems (IDSs) have become the necessary component of 

the computer and information security framework. Although the field of IDSs 

is still developing, they are not able to detect all types of intrusions. This work 

discuss about the ways of implementing Genetic Algorithm (GA) to detect intrusions. 

We first use a mobile agent technology for collecting data properties. 

These data are evaluated by the genetic algorithm to improve the quality of 

basic rules by trying to find the best solution. 

2 - A Multiobjective Approach for Rapid Transit Network 

Design 

Erdem Gundogdu, Industrial Engineering, Istanbul Kultur 

University, IKU Atakoy Campus Room:210 Bakirkoy, 34156, 

Istanbul, gundogduerdem@gmail.com, Orhan Feyzioglu 

Urban rapid transit network design consists of the location of train alignments 

and stations in an urban traffic context. In this study, we investigate the design 

of a single transit line with no predetermined origin and destination while 

considering the existing road network. The problem is formulated as a bilevel 

multiobjective optimization problem where line investment and car emissions 

are considered at the upper level, and user traffic behavior is considered at the 

lower level. A genetic algorithm is developed to obtain non-dominated solutions 

and it is applied on an illustrative example. 

3 - Simultaneous Design of Urban Road and Public Transit 

Networks Using Hybrid Evolutionary Metaheuristics 

Elnaz Miandoabchi, Industrial Engineering, AmirKabir 

University of Technology, 123456, Tehran, Tehran, Iran, Islamic 

Republic Of, el.miandoabchi@aut.ac.ir, Reza Zanjirani Farahani 

211

WA-06 EURO 24 - Lisbon 2010 

In this paper, a multi-modal urban road network design problem with auto and 

bus modes is considered. The problem is to concurrently design the road and 

bus networks. The road network decisions are, adding new streets, and new 

lanes to the existing streets, orienting one-way streets, and two-way street lane 

allocations. The bus network decisions are redesigning of routes of the existing 

bus lines. The problem is modeled as a multi-objective bi-level model with a 

combined modal-split/assignment problem. A genetic algorithm and a clonal 

selection algorithm are developed as solution methods. 

4 - Mean-risk multistage transmission network expansion 

planning using multiobjective local search 

João Claro, INESC Porto, Faculty of Engineering, University of 

Porto, Portugal, jclaro@fe.up.pt, A. Miguel Gomes 

We present a multiobjective local search approach for a mean-risk multistage 

formulation of the Transmission Network Expansion Planning problem. Uncertainty 

in generation costs and demand loads is captured through a scenario 

tree. In each period, discrete decisions concerning the investment in the network, 

and continuous decisions concerning the utilization of the network, are 

considered. The approach integrates liner programming to address the utilization 

problem. We present experimental results with a set of instances derived 

from the literature. 

� WA-06 


8.2.30 

DEA Application V — Industry and natural 

resources 



Chair: Vania Sena, Aston University, B4 7ET, Birmingham, United 

Kingdom, v.sena@aston.ac.uk 

1 - Data Envelopment Analysis for Privatizing Decisions of 

Sugar Factories 

Ezgi Aktar Demirtas, Industrial Engineering, Eskisehir 

Osmangazi University, Eskisehir Osmangazi University 

Department of IE, Meselik ESKISEHIR/TURKIYE, 26480, 

ESKISEHIR, Turkey, eaktar@ogu.edu.tr, Abdullah Korkut 

Üstün 

Data envelopment analysis (DEA) is a nonparametric method for the estimation 

of production frontiers. It is used to empirically measure efficiency scores of 

decision making units. In this study, the efficiency scores of 26 sugar factories 

are evaluated by both input and output oriented DEA models. These efficiency 

scores and input/output target values can be used for privatizing decisions of 

public factories in Turkey. 

2 - Technical and economic efficiency analysis applied to 

artisanal fisheries 

Manuela Maria Oliveira, U-REMS, IPIMAR, Portugal, 

moliveira@ipimar.pt, Ana Camanho, Miguel Gaspar 

The research sought to determine the efficiency of vessels using DEA models.Using 

data on the prices of each species in wholesale market,revenue efficiency 

was also estimated to complement the technical efficiency analysis.An 

advantage of this approach resides in the ability to separate technical from allocative 

aspects in the efficiency assessment,enabling a graphical representation 

of the performance of vessels in two dimensions.This approach enables the 

specification of targets for inefficient vessels that corresponds to the amount of 

captures by species that enable maximizing the revenue 

3 - Efficiency of Public Forestry Firms in Switzerland 

Alexander Mack, Engineering and Information Technology, Bern 

University of Applied Sciences, Quellgasse 21, 2501, Biel, 

Switzerland, alexander.mack@bfh.ch, Bernur Acikgoz 

This paper is an empirical study of the productive efficiency of public forestry 

firms in Switzerland, using original unpublished data. By comparing various 

forestry firms among themselves, one can identify the most efficient units and 

their distinctive features. In order to determine the productive efficiency, DEA 

analysis is used. In-depth analysis suggests that management style, environmental 

influences (e.g., hurricane Lothar), and the forest region affect the efficiency 

of the firms. In addition, Pedroni panel cointegration technique is applied 

by using the DEA scores. 

212 

4 - Productivity change of textile and clothing firms in 

world regions with an application of Malmquist index 

with bootstrap 

Magdalena Kapelko, Department of Business Administration, 

Universidad Carlos III de Madrid, Calle Madrid 126, Office 

7.0.56, 28903, Getafe (Madrid), Spain, 

magdalena.kapelko@gmail.com 

The purpose of this paper is to study the productivity change of 4985 observations 

of firms in the textile and clothing industry for 1995-2004 time-period. 

To differentiate from previous research we analyze the firms in different world 

regions. We use the Malmquist index with bootstrap and decompose it into 

three sources of productivity change. Our results indicate a relatively small deterioration 

in productivity, which is driven by the technical and scale efficiency 

decline, in spite of the technological progress. However, those tendencies are 

not a common feature for the regions considered. 

� WA-07 


8.2.47 

Project scheduling 



Chair: Willy Herroelen, Decision Sciences and Information 

Management, Katholieke Universiteit Leuven, Naamsestraat 69, 

3000, Leuven, Belgium, willy.herroelen@econ.kuleuven.be 

1 - Project scheduling in the financial management of supply 

chains 

Bajis Dodin, Anderson Graduate Scool of Management, 

University of California, Anderson Graduate Scool of 

Management, University of California, 92521, Riverside, 

California, United States, bajis.dodin@ucr.edu, Abdelghani 

Elimam 

Most research on managing Supply Chains has focused on the physical aspects, 

very little on the Financial Management of SC (FMSC). However, return 

on investment can be a complicated task in large SCs. Long cycle times affect 

the value of material, and the cost of borrowing/investing the necessary 

funds. In this paper we first determine the issues affecting the FMSC, and then 

use project scheduling methods to analyze the relationship between these issues 

and the SC cycle time. This analysis led to establishing optimal and near 

optimal cash management policies for all participants of the SC. 

2 - Tactical risk management in the construction industry 

Stefan Creemers, K.U.Leuven, 3000, Leuven, Belgium, 

stefan.creemers@econ.kuleuven.be 

Construction projects may be divided into several phases: design, build, finance, 

operate and/or maintain. Typically, contractors are only involved in the 

build phase. In PPPs (public private partnerships) however, contractors assume 

additional responsibilities as to improve the time to market, to cut costs, etc. 

Additional responsibilities give rise to additional risks; risks that contractors 

traditionally did not face. We devise a practical tool that helps contractors to 

make the right decisions. 

3 - Applying Critical Chain Scheduling and Buffer Management 

on the Discrete Time/Resource Trade-off Problem 

Erik Demeulemeester, Decision Sciences and Information 


3000, Leuven, Belgium, 

erik.demeulemeester@econ.kuleuven.be, Wendi Tian 

Goldratt’s Critical Chain scheduling and Buffer Management methodology, 

which relies on deterministic scheduling techniques to build a baseline schedule 

that is made robust by inserting various types of buffers, has attracted attention 

for scheduling the resource-constrained project scheduling problem (RCPSP) 

in a robust way. We extend CC/BM to the discrete time/resource trade-off 

problem in which the duration of an activity is assumed to be a discrete, nonincreasing 

function of the amount of a single renewable resource committed to 

it. Extensive computational results will be presented.

4 - Robust optimization for resource-constrained project 

scheduling with uncertain activity durations 

Christian Artigues, LAAS, CNRS, 7 avenue du Colonel Roche, 

31077, Toulouse Cedex 4, artigues@laas.fr, Roel Leus, Fabrice 

Talla Nobibon 

We propose models and methods for resource-constrained project scheduling 

with uncertain activity durations. Our modeling techniques stem from robust 

optimization and the objective is to find a scheduling policy that minimizes 

the maximum absolute regret over all scenarios. We propose lower and upper 

bounds for the maximal regret, integer linear programming formulations and a 

scenario-relaxation-based solution method. We provide computational results 

that illustrate the compromise between solution quality and robustness. 

� WA-08 


6.1.36 

Shop Scheduling 



Chair: Pierre Lopez, Groupe MOGISA, LAAS-CNRS, 7 avenue du 

Colonel Roche, 31077, Toulouse, France, pierre.lopez@laas.fr 

1 - Job-Shop with several transport robots with non unary 

capacities 

Mohand Larabi, LIMOS (ISIMA), Blaise Pascal University, 

LIMOS (ISIMA), Campus des Cezeaux, 63177, 

Clermont-Ferraand, Puy-de-Dôme, France, larabi@isima.fr, 

Philippe Lacomme, Tchernev Nikolai 

This paper concerns the Job-Shop scheduling problems with several transport 

robots having non unary capacities. The problem is a generalisation of the jobshop 

including transport operations achieved by a fleet of homogenous robots 

with multi load capacity. To model the problem we extend the classical disjunctive 

graph dedicated to job-shob. We propose a framework based on a 

powerful local search which uses new specific problem properties. Experiment 

shows that the method compete with the well known methods on the case of 

the job-shop with several robots of unary capacities. 

2 - Generalized resource constraint propagation for job 

shop scheduling with time lags 

Pierre Lopez, Groupe MOGISA, LAAS-CNRS, 7 avenue du 

Colonel Roche, 31077, Toulouse, France, pierre.lopez@laas.fr, 

Marie-José Huguet, Christian Artigues 

The aim of this work is to illustrate the efficiency of generalized resource constraint 

propagations for solving job shop scheduling problems with time lags. 

We propose a heuristic based on job insertion and a branch-and-bound benefiting 

from generalized resource constraint propagation techniques. This generalization 

comes from the search for longest paths on a complete constraint 

network associated with the scheduling problem. We evaluate the impact of our 

insertion heuristic and we show the interest of generalized propagation through 

experimental results and comparisons with other methods. 

� WA-09 


6.2.53 

Recent Developments and Applications in 







Chair: Leena Suhl, Int. Graduate School of Dynamic Intelligent 

Systems, University of Paderborn, Decision Support & OR Lab, 

Warburger Str. 100, 33098, Paderborn, Germany, suhl@upb.de 

EURO 24 - Lisbon 2010 WA-10 

1 - On the Conjecture of Aouchiche and Hansen about the 

Randic Index 

Ljiljana Pavlovic, Department of Mathematics, Faculty of 

Natural Sciences nd Mathematics, Radoja Domanovica 14, 

34000, Kragujevac, Serbia, Serbia, pavlovic@kg.ac.rs, Marina 

Stojanovic 

Let G(k,n) be the set of connected simple graphs which have n vertices and the 

minimum degree of vertices is k. The Randic index of a graph G is defined as 

sum of d(u)d(v) raised to the power of -1/2 , where d(u) is the degree of vertex 

u and the summation extends over all edges uv of G. In this paper we prove the 

conjecture on the graphs for which the Randic index attains its minimum value 

when k is greater or equal to n/2. We show that the extremal graphs have only 

degree k and degree n-1, and the number of vertices of degree k is as close to 

n/2 as possible. 

2 - Integrating production, staff size and product price in 

aggregate planning 

Amaia Lusa, IOC Research Institute / Management Department, 

Universitat Politècnica de Catalunya, Avda. Diagonal 647, p11, 

08028, Barcelona, Spain, amaia.lusa@upc.edu, Carme Martinez, 

Marta Mas 

Integrating decisions of different areas is a current trend in operations management 

that it is possible thanks to improvements both in hardware and software 

capacity. In this work, we discuss an aggregate planning problem that includes 

production, staff size, cash management and marketing decisions. The demand 

is considered to be a nonlinear function of the product price. The problem, 

which is modelled as a mixed integer linear program, can be solved using standard 

optimization software. The results of a computational experiment show 

the efficiency and convenience of the proposed model. 

3 - Optimizing a job-shop plant using queuing networks 

techniques 

Artur Barreiros, Mechanical Engineering Department, Instituto 

Superior Técnico, Av. Rovisco Pais, 1049-001, Lisbon, Portugal, 

barreiros@ist.utl.pt, Dulce Silva 

This paper applies open queuing networks to evaluate and optimize the performance 

of a job-shop plant. The performance diagnostics involves analysis of 

the work-in-process and the production lead-times. We assume that the plant 

can be modelled by a collection of stations. The optimization procedure involves 

the probabilistic characterization of the demand and characterization of 

the probability of transference of customers. A waiting-cost function is evaluated 

using the queuing parameters, which is used to compare alternative solutions. 

4 - Computational Experiments with Multi-Row Cuts 

Franz Wesselmann, Decision Support & OR Lab, University of 

Paderborn, Warburger Str. 100, 33098, Paderborn, Germany, 

wesselmann@dsor.de, Leena Suhl, Uwe Suhl 

Recently, there has been a renewed interest in cutting planes from multiple rows 

of the simplex tableau. In particular, a beautiful correspondence between minimal 

valid inequalities for the infinite group relaxation and maximal lattice-free 

convex polyhedra was discovered. In this talk, we discuss our implementation 

of multi-row cut generators. We detail the construction of the multi-row relaxation 

and the selection of a maximal lattice-free convex set. We also report on 

computational experience with multi-row cutting planes and empirically compare 

them with Gomory mixed-integer cuts. 

� WA-10 


6.2.56 

OR in Supply Chain Management I 



Chair: Eren Ozceylan, Industrial Engineering, Natural and Applied 

Sciences, Selcuk University Industrial Engineering Department, 

Campus, 42031, Konya, Turkey, eozceylan@selcuk.edu.tr 




Chair: Edite M.G.P. Fernandes, Production and Systems, University 

of Minho, School of Engineering, Campus de Gualtar, 4710-057, 

Braga, Portugal, emgpf@dps.uminho.pt 

213


1 - Characterizing and modeling supply chain system dissipations 

Iskander Zouaghi, CERAG, Grenoble University, 150 rue de la 

chimie, 38400, Saint Martin d’Hères, France, 

iskander.zouaghi@upmf-grenoble.fr, Tarik Saikouk, Alain 

Spalanzani 

The complexity of firms’ environment is characterized by open markets, globalization, 

information technologies, decreasing of products lifecycles, increased 

demand, but also by consumers who are becoming increasingly demanding. 

Demand has become too volatile. We use a multi-objective optimization model 

to conceptualize dissipations in supply chain systems, by considering two main 

conflicting objectives. These objective functions are optimized under dissipation 

constraints. 

2 - The Robust Model of Multiple Sourcing: Alchemy of 

Risk Reduction in Supplier Selection 

Adel Azar, Tarbiat Modarres, 12345678, Tehran, 

azara@modarres.ac.ir, Masood Rabieh, Mohammad Javad 

Feyzollahi 

Selecting appropriate suppliers can significantly increase the competitiveness 

of organizations. Challenges of new complex environments cause increased 

uncertainty, so that we use robust optimization. A robust nonlinear optimization 

model is developed for supplier selection of raw material at Isfahan Steel 

Company. Then the robust model is solved with different risk levels. The quality 

of solutions is analyzed by a simulation technique. 

3 - Multiple Sourcing Strategies: Trade-off between Costoriented 

and flexibility-oriented Supplier 

Yuan-Du Hsiao, Business Administration, Chungyu Institute of 

Technology, No40,Yi7th Rd,Keelung20103,Taiwan,R.O.C„ 

6f,No46,Dong-Guang Rd,Keelung20103,Taiwan,R.O.C„ 

20103Taiwa, Keelung,City, Keelung,Taiwan,R.O.C, Taiwan, 

yuduhsiao4471@yahoo.com.tw, Fan-Yun Pai, Tsu-Ming Yeh 

Within the environment of SCM, a multiple sourcing strategy can be adopted 

to take advantages of different suppliers with their different capabilities. We 

construct a tailored sourcing model to find out the retailer’s orders allocation 

mechanism between emergency vendor and discount vendor. From the analysis 

results, the existence of an emergency vendor will reduce order quantity from 

discount vendor. The profit depends on the difference between the order quantity 

from a discount vendor and from an emergency vendor and the expected 

value of demand function. 

4 - DEA Analysis with Data Which are not Crisp 

Soheila Ebrahimkhany Ghazy, basic Science college, Science 

and Research Branch, IAU, TEHRAN -POUNAK 

SQUARE-HESARAK- SCIENCE AND RESEARCH 

UNIVERSITY, Ave sheikhbahaee . St west kashfiyan Nom:43 

Building star, 1477893855, tehran, Iran, Islamic Republic Of, 

soheila7764@yahoo.com, Mohsen Rostamy-malkhalifeh 

Earlier methods in classic DEA investigate congestion for DMUs with crisp 

data. Although in real world we face with data that aren’t crisp necessarily, 

available methode in this situation are not enough. Congestion occurs when 

the reduction of selected inputs causes some, rather than all, outputs to increase, 

without a worsening of others. As a result in this paper for investigation 

of congestion we promote a new method with Fuzzy data. Then explain our 

method with a numerical example. 

214 

� WA-11 


8.2.38 

Various New OR Tools and Technologies II: 

Quality Management Emphasized 



Chair: Gulser Koksal, Industrial Engineering, Middle East Technical 

University, Inonu Blvd., 06531, Ankara, Turkey, 

koksal@ie.metu.edu.tr 







1 - A Feature Extraction Method for Fault Detection of 

Cyclic Signals 

Jonghyuck Park, Graduate School of Information Management 

and Security, Korea University, Anamdong, Sungbukgu, 

136-713, Seoul, Korea, Republic Of, nrevival@gmail.com, 

Jun-Geol Baek, Sung-Shick Kim 

This paper presents a feature extraction method based on Wavelet Multi- 

Resolution Analysis to detect faults in cyclic signals. The proposed method 

is expected to reduce the size of data and well reflect the characteristic of original 

signal. However, extracted features vary by the signals due to the variations 

of the process; common features should be chosen to build up a fault detection 

model. Therefore, feature scoring algorithm for the selection of common 

feature is also presented. Performance evaluation of the proposed method is 

provided in comparison to well-known feature selection rules. 

2 - Beta-Geometric Model for Estimating the Product Quality 

After Multiple Inspections 

Young Chun, Infor. Sys. & Dec. Sci., Louisiana State University, 

College of Business, 70803-6315, Baton Rouge, LA, United 

States, chun@lsu.edu 

A complex product such as a software document is inspected more than once 

to further improve its software quality. Authors have proposed various estimation 

methods for the number of faults still remaining in the software document. 

For each fault, the probability that it will be detected during one review 

cycle is an unknown constant to be estimated. We propose a beta-geometric 

inspection model. In a numerical study, we compare the effectiveness of our 

beta-geometric model with those of traditional inspection models in which the 

detection probability is assumed to be a constant. 

3 - Competition in Remanufacturing 

Serra Caner, Operations, University of Groningen, Nettelbosje 2, 

9747 AE, Groningen, Netherlands, s.caner@rug.nl, Ruud 

Teunter, Xiang Zhu 

We study remanufacturing competition between an Original Equipment Manufacturer 

and an Independent Operator. Differently from literature, the OEM 

and IO compete for cores with their acquisition prices. We consider a 2-period 

model with manufacturing by the OEM in the first period, and manufacturing 

as well as remanufacturing in the second. We determine optimal policies for 

both players by establishing a Nash Equilibrium in the second period. A sensitivity 

analysis leads to a number of managerial insights. Further insights are 

obtained from a numerical investigation. 

4 - Controlling product returns from sales demonstrations 

Luc Muyldermans, Business School, Nottingham University, 


luc.muyldermans@nottingham.ac.uk, Luk Van Wassenhove, 

Daniel Guide 

Some manufacturers demonstrate their products so that customers can gain experience 

before making a purchase. We analyse a case where the product returns 

from sales demonstrations are substantial, and derive optimal demo policies 

analytically. The key trade-off in our model is either to reuse returned 

demo products repeatedly, or to sell ex-demo products on a secondary market 

and use new products to fulfil demo requests. The optimal policies are straightforward 

in the case of constant product prices, but become more intricate when 

the resale price on the secondary market erodes over time.

� WA-12 


8.2.39 

ANP 06 


Processes 


Chair: Ziya Ulukan, industrial engineering, galatasaray university, 

ciragan cad. no:36, ortakoy, 34357, istanbul, Turkey, 

ziyaulukan@hotmail.com 

1 - An integrated AHP - ANP Quality Function Deployment 

framework for the competitive position improvement 

George Paltayian, Department of Business Administration, 

University of Macedonia, 156 Egnatia Street, 54006, 

Thessaloniki, Thessaloniki, Greece, gpaltag@uom.gr, Andreas 

Georgiou, Katerina Gotzamani, Andreas Andronikidis 

An integrated AHP — ANP - Quality Function Deployment (QFD) framework 

is presented in order to distinguish and translate customers’ needs into goals. 

The applicability of the proposed framework is demonstrated in a case which 

summarizes the interventions and suggestions of the model’s application for 

a bank. It is worth investigating methods and techniques that could improve 

the competitive position and as a result, increase market share and profitability 

and the proposed framework can be employed in this direction to translate the 

customers’ needs into various effective strategies. 

2 - Selection of a suitable WEEE management system using 

fuzzy ANP 

Ziya Ulukan, industrial engineering, galatasaray university, 

ciragan cad. no:36, ortakoy, 34357, istanbul, Turkey, 

ziyaulukan@hotmail.com, Sigrid de Mendonca Andersen 

Socially acceptable, economically affordable, technologically reliable and environmentally 

friendly scenario’s for WEEE are the major concerns. The aim 

of this paper is to provide a multi-criteria decision making tool to select the best 

scenario. Firsttly, fuzzy AHP is used to determine the relative importance of 

the selection criteria. And then, with a fuzzy version of ANP, different WEEE 

management systems are evaluated. An illustrative application is also given 

to demonstrate the effectiveness of the methodology. The alternatives and the 

criteria are determined from the literature 

3 - Determining effective criteria on investment priorities 

of wood and paper industries, Case of the study: Iran 

Qom province 

Majid Azizi, wood and paper sciences and technology, Faculty of 

natural resources, University of Tehran, karaj, tehran, Iran, 

Islamic Republic Of, mazizi@ut.ac.ir, Mehdi Faezipour, Reza 

Roknedin Eftekhari, Rafat Dehghan Krooki 

Since the province has special capability to invest in wood and paper industries, 

determining effective criteria is vital. AHP has been applied and a decision tree 

planned with five major criteria: infrastructure, economic & financial, material 

& product, man force & technical, social & cultural and 37 their sub criteria. 

The questionnaires have been distributed and gathered from the experts. Expert 

Choice used to synthesize the results. Among criteria and sub criteria, 

economic & financial criterion and investment attraction sub criterion have the 

highest priorities respectively. 

� WA-13 


2.2.21 

Applications of Location 



Chair: Mercedes Landete, Departamento de Estadística y Matemática 

Aplicada, University Miguel Hernández of Elche, Avda. del 

Ferrocarril s/n, 03202 , Elche, Alicante, Spain, landete@umh.es 


1 - Optimal location of safety cameras at urban intersections 

for road accidents control: a pilot study in the city 

of Rome 

Antonino Sgalambro, Dipartimento di Statistica, Probabilità e 

Statistiche Applicate, Sapienza Università di Roma, Piazzale 

Aldo Moro 5, 00185, Roma, Italy, 

antonino.sgalambro@uniroma1.it, Paolo Dell’Olmo, Nicoletta 

Ricciardi, Riccardo Colicchia, Katia Chiusolo 

We consider the optimal location of safety cameras at urban road intersections 

to maximize road control and reduce accident effects. Integrated use of geographical 

information systems and statistical analyses on historical data is proposed 

to individuate black-spots, i.e. areas with high risk of accidents. The 

optimal location of safety cameras is then formulated and solved as a maximal 

covering location problem. The results of a pilot study on the application of 

this methodology to real data on road accidents that occurred in Rome between 

the years 2006 and 2009 is presented and discussed. 

2 - A Mathematical Model for Urban Concentrations with 

Respect to the Advance of Japanese Railway Networks 

Yudai Honma, Waseda Institute for Advanced Study, Waseda 

University, 60-02-05A, Okubo 3-4-1„ 169-8555, Shinjuku-ku, 

Tokyo, Japan, yudai@aoni.waseda.jp 

The locational dynamics of commercial activities in cities has been explored 

by the Harris-Wilson’s balancing-mechanism model. In this study, we take the 

balancing-mechanism in the general urban activities, and analyze the locational 

dynamics of urban activity distribution with respect to the advance of Japanese 

railway networks. In particular, focusing on the high-speed transit system, we 

examine how the constructions of Shinkansen bullet train and Maglev train affect 

the developments of cities in Japan. As a result, we clarified that the opening 

of high-speed transit system promotes the further concentrations to large 

cities, and miserable declinations of small cities. 

3 - The retrofit of a closed-loop distribution network: the 

case of lead batteries 

Maria Isabel Gomes, CMA - FCT - Universidade Nova de 

Lisboa, Monte da Caparica, 2829-516, Caparica, Portugal, 

mirg@fct.unl.pt, Ana Fernandes, Ana Paula Barbósa-Póvoa 

Nowadays many companies face new challenges regarding the management of 

end-of-life products. Their supply chains, once designed to efficiently satisfy 

customers’ demands are now facing a new product flow, the reserve flow. The 

simultaneous optimization of both networks is now an emerging challenge. In 

this work, the closed-loop supply chain of a company that produces, distributes 

and collects lead batteries is optimized by means of a multi-product, multiperiod, 

location-allocation model. The results obtained are compared with the 

existing network and important conclusions are drawn. 

4 - A multi-level approach to sitting EOL treatment resources 

Branislava Ratkovic, Logistics, Faculty of Traffic and Transport 

Engineering, University of Belgrade, Vojvode Stepe 305, 11000, 

Belgrade, Serbia, b.ratkovic@sf.bg.ac.rs, Milorad Vidovic 

This paper presents the modeling approach for establishing reverse logistics 

network, through defining optimal locations of collection points, consolidation 

points and treatment facilities. The active participation of consumers is essential 

for achieving any recovery objective, and in order to model the influence 

of distance between consumers and collection points on the optimal locations 

of three types of facilities to be located on this network, we introduce the collection 

point’s catchment area. Proposed modeling approach was tested on the 

numerical example. 

� WA-15 


2.2.12 

Vehicle Routing Applications II 



Chair: Min Wen, DTU Transport, DTU Transport, Building 115, 

Technical University of Denmark, 2800 Lyngby, Denmark, 2800, 

lyngby, Copenhagen, Denmark, mw@transport.dtu.dk 

215


1 - Utilization and Positioning of Transshipment Points in 

Pickup and Delivery 

Michael Schneider, BISOR, University of Kaiserslautern, 

Erwin-Schrödinger-Straße, Geb. 42-420, Kaiserslautern, 67653 , 

Kaiserslautern, Germany, schneider@bisor.de, Oliver Wendt 

The pickup and delivery problem with transshipment (PDPT) allows the transfer 

of goods between vehicles and thus the fulfilment of one pickup and delivery 

request using several vehicles. We tailor an established solution method 

for the classical PDP to be able to handle transshipment points. Our numerical 

studies investigate the structural properties of problem instances for which 

the utilization of transshipment points is beneficial. Moreover, we study the 

effectiveness of different schemes for determining the number and position of 

effective transshipment points. 

2 - Vehicle Routing Problem with Availability Constraints 

Farhana Johar, Mathematics, University of Southampton, School 

of Mathematics, Highfield Campus, SO17 1BJ, 

SOUTHAMPTON, United Kingdom, farhana1279@yahoo.com 

This research is classifies as non-classical Vehicle Routing Problem (VRP) 

where the maximum release date of customer’s demand of the route determine 

the vehicle departure time. Thus, there could be lateness on the delivery process 

from awaiting all customers’ demand of the route to be released. A mathematical 

formulation is developed to represent the problem studied. Insertion method 

based on the cheapest cost is used to generate an initial solution. Then, Local 

Search technique is applied to improve the solution in term of minimization of 

total traveling and tardiness cost. 

3 - Combined Vehicle Routing and Foldable Container 

Scheduling 

Jan Zazgornik, Institute of Production and Logistics, University 

of Natural Resources and Applied Life Sciences, Vienna, 

Feistmantelstrasse 4, 1180, Vienna, Austria, 

jan.zazgornik@boku.ac.at, Patrick Hirsch, Manfred Gronalt 

This work deals with a combined vehicle routing and container scheduling 

problem in forest industry. The problem is modeled as a vehicle routing problem 

with time windows and additional constraints for using foldable containers. 

These constraints ensure that containers are available at the pick-up locations 

when needed. Therefore, the problem is state-dependent during execution. Two 

Tabu Search variants with different neighborhood structures have been developed 

and tested in extensive numerical studies. The results show that the methods 

generate good solution in reasonable computing time. 

4 - The dynamic multi-period vehicle routing problem 

Min Wen, Department of Management Engineering, Technical 

University of Denmark, DTU - Bygning 426, room 043, 2800, 

Lyngby, Denmark, mw@imm.dtu.dk, Jean-François Cordeau, 

Gilbert Laporte, Jesper Larsen 

This presentation addresses the Dynamic Multi-Period Vehicle Routing Problem 

which deals with the distribution of orders from a depot to a set of customers 

over a multi-period time horizon. Customer orders and their feasible 

service periods are dynamically revealed over time. The objectives are to minimize 

total travel costs and customer waiting, and to balance the daily workload 

over the planning horizon. This problem is modeled as a mixed integer linear 

program, and solved by means of a three-phase heuristic that works over 

a rolling planning horizon. Computational results show that the proposed approach 

can yield high quality solutions within reasonable running times. 

� WA-16 


2.2.14 

Case studies in planning and operations 



Chair: Konstantinos Gkoumas, DITS (Dipartimento di Idraulica 

Trasporti e Strade), Sapienza Universita’ di Roma, Via Eudossiana 

18, 00184 , Rome, Italy, konstantinos.gkoumas@uniroma1.it 

1 - Public transport service reliability in high- and low frequency 

systems 

216 

Konstantinos Gkoumas, DITS (Dipartimento di Idraulica 

Trasporti e Strade), Sapienza Universita’ di Roma, Via 

Eudossiana 18, 00184 , Rome, Italy, 

konstantinos.gkoumas@uniroma1.it, Michael Schachtebeck, 

Anita Schoebel 

The aim of this study is the comparative service reliability analysis of high frequency 

(e.g. metro) and low frequency (e.g. railway) public transport systems. 

We critically review known reliability indices with respect to this question before 

we introduce a novel index accounting for the average delay experienced 

by the passengers in the network. We present numerical tests on the performance 

of this index, using close to real world data from the German railway 

system and from the Athens metro. We also discuss the effects of delay management 

strategies in high and low frequency systems. 

2 - Equity and externalities issues in the bus network design 

Ernesto Cipriani, Dept. Civil Engineering, University of Roma 

TRE, Via Vito Volterra 62, 00146, Roma, Italy, 

eciprian@uniroma3.it, Marco Petrelli, Gaetano Fusco 

The present paper represents the development of an optimization method for 

the bus network design, developed by authors, to define the network configuration 

in terms of bus routes and service frequencies minimizing an objective 

function representing total costs, internal and external, involved by the multimodal 

transport system with elastic demand. The main novelties of the design 

methodology presented in this paper concern with: a) the introduction of the 

equity concept in terms of level of bus service supply among different urban 

districts; b) the application of an explicit constraint to line capacity; c) a new 

procedure for the externalities estimation based on analytical computation. 

3 - Analysis of Bus Rapid Transit Systems 

Marco Petrelli, Università Roma Tre, Italy, 

mpetrelli@uniroma3.it, Ernesto Cipriani, Stefano Gori 

The present paper deals with a procedure for the evaluation of the performance 

of Bus Rapid Transit (BRT) systems and other related issues about the capabilities 

of BRT to provide a network of high quality and high reliability services 

in medium-high demand corridors. The problem addressed involves the performance 

analysis from the point of view of both the operators and the users 

and it deals with the basic BRT system implementation problem: the definition 

of the variables that establish the performance of each different system and the 

optimal field of existence of these elements. 

4 - Transit frequency optimization through Neural Networks 

Daniele Tiddi, DITS, University of Rome "Sapienza", Via 

Eudossiana 18, 00184, Rome, Italy, daniele.tiddi@uniroma1.it, 

Guido Gentile 

In the proposed work, we aim to minimize total travel time of public transport 

users, subject to constraints of resources availability. User travel time is 

evaluated by a deterministic transit assignment based on hyperpaths. Whilst 

the constraints yield a convex feasible set, the objective function is not convex 

due to the nature of user path choice (deterministic and based on hyperpaths). 

A neural network is then trained from the assignment results. Properties of 

the resulting function between line frequencies and total travel time are then 

investigated. 

� WA-17 


1.3.14 

Network Design for Road Transportation 



Chair: Mariano Gallo, Dipartimento di Ingegneria, Università degli 

Studi del Sannio, Piazza Roma 21, 82100, Benevento, Italy, 

gallo@unisannio.it 

1 - A Scatter Search algorithm for solving the road network 

design problem 

Mariano Gallo, Dipartimento di Ingegneria, Università degli 

Studi del Sannio, Piazza Roma 21, 82100, Benevento, Italy, 

gallo@unisannio.it, Luca D’Acierno, Bruno Montella

In this paper we focus on the road network design problem in regional contexts; 

in this case a planner may have financial resources to be invested for improving 

performances of an existing network. In the literature many optimisation models 

for solving this problem exist; we propose a model that considers also the 

evaluation of environmental external costs in the objective function. Then, the 

paper will focus on the solution algorithm: we will propose a Scatter Search 

algorithm that will be able to solve the problem for real-scale networks in acceptable 

computing times. 

2 - A Multiobjective Metaheuristic Approach to Support the 

Rural Network Planning 

Pablo Maya, Environment, technology and technology 

management, University of Antwerp, Stadscampus, S.B.513, 

Prinsstraat 13, 2000, Antwerp, Antwerp, Belgium, 

pmayaduque@gmail.com 

The departmental government of Antioquia (Colombia) has defined a methodology 

to build a ten year investment plan for the rural road network in the 

different subregions of the department. The aim of this work is to complement 

and improve this methodology by proposing strategies to support the decision 

making processes involved. The strategy we propose is based on multiobjective 

optimization model that integrates the different decisions and provides the decision 

makers with a tool to evaluate plans with different level of compromise 

of the multiple criteria considered. 

� WA-18 


1.3.15 

Graph Theory and Combinatorial 

Optimization 











1 - On the cubical dimension of four new classes of trees 

Kamal Kabyl, Laboratory of Modeling and Optimization of 

Systems LAMOS, Commercial Sciences Department, University 

of Bejaia, Algeria, 06000, Bejaia, Algeria, 

k_kabyle2000@yahoo.fr, Abdelhafid Berrachedi 

The hypercube of dimension n is the graph in which the set of vertices are binary 

n-tuples, and two vertices are adjacent if and only if they differ in only one 

component. Many authors have studied the embedding of trees in hypercube 

which allowed to characterize certain classes. The problem consist of giving 

the smallest dimension of a hypercube in which a given tree G is embeddable. 

We talk then about optimal hypercube and cubical dimension of a hypercube of 

the tree. In this paper, we give four new classes of trees for which the cubical 

dimension is established. 

2 - Dynamically Maintaining Chordal and Weakly Chordal 

Graphs 

Mohamed Amine Boutiche, Operations research, University of 

science and technology, BP 32, El Alia, 16111, Bab Ezzouar, 

Algiers, Algeria, boutichemedamine@yahoo.fr 

We present an algorithm that support operations for modifying a tree decomposition 

representation of chordal and weakly chordal graphs, by adding and 

deleting edges or vertices, such that after each modification the tree decomposition 

representation of both graphs is repaired in a minimal way. In particular, 

if the graph is not chordal (resp. weakly chordal) after the modification, the 

algorithm computes a valid tree decomposition for the modified graph. Moreover, 

we update the two parameters tree-length and tree-width of both graph 

classes after each modification. 

3 - On defining a combinatorial space 

Sergii Sirenko, V.M. Glushkov Institute of Cybernetics NAS 

Ukraine, Kyiv, Ukraine, ssirenko@acm.org, Leonid Hulianytskyi 

EURO 24 - Lisbon 2010 WA-20 

We introduce a formal approach to defining the important notions in combinatorial 

optimization: a combinatorial optimization problem (COP), a combinatorial 

space, a combinatorial object, a neighbourhood, a move operator, a path 

and a segment. The key distinction from currently dominating view is a possibility 

of the combinatorial space to be countable. It is defined using the notion 

of locally finite space. The combinatorial object generalizes Berge’s combinatorial 

configuration. Suggested definition of the directed segment provides 

constructive mechanism for using them for COP solving. 

4 - Integer Linear Stochastic Programming with Multiple 

Objective 

Amrouche Salima, Mathématiques, université Saad Dahlab 

Blida, Université SAAD DAHLAB Blida Route De Soumaa BP 

270 BLIDA, amrouchesalima@gmail.com, Algeria, 

amrouchesalima@gmail.com 

Real life decision problems have three main properties: (i) conflicting objectives 

in the problem structure, (ii) stochasticity in the description of problem 

parameters in contexts where the probability distribution of random parameters 

is known, and (iii) involvement of integer decision variables which increase the 

dimension. The proposed modeling and solution methods are able to identify 

all the integer feasible solutions which are efficient and with convergence in a 

finite number of iterations. 

� WA-20 


1.3.33A 

Data Mining and Credit Risk 

Stream: Data Mining and Decision Making 


Chair: Christophe Mues, School of Management, University of 

Southampton, SO17 1BJ, Southampton, United Kingdom, 

C.Mues@soton.ac.uk 

1 - Tuning metaheuristics: A data mining based approach 

for particle swarm optimization 

Stefan Lessmann, Institute of Information Systems, University of 

Hamburg, Von-Melle-Park 5, 20146, Hamburg, Germany, 

lessmann@econ.uni-hamburg.de, Idel Montalvo Arango, Marco 

Caserta 

Metaheuristics are powerful procedures for solving complex optimization problems. 

Their performance will usually depend upon a fine tuning of algorithmic 

parameters. Since metaheuristics operate in an iterative manner, data concerning 

effective parameter settings is naturally produced during execution. We 

propose employing this data for building regression models that facilitate an 

automatic tuning of parameters within an online learning framework. The feasibility 

of this approach is explored for different types of regression models in 

a case study of particle swarm optimization. 

2 - Mixture cure models in consumer credit risk 

Edward Tong, School of Management, University of 

Southampton, School of Management, University of 

Southampton, SO17 1BJ , Southampton, Hampshire, United 

Kingdom, e.tong@soton.ac.uk, Christophe Mues, Lyn Thomas 

Mixture cure models were proposed in medical statistics to model long-term 

survival of cancer patients as two distinct subpopulations — those cured of the 

event of interest along with those that are uncured and susceptible to relapse. 

We examine the performance of the mixture cure model relative to Cox regression 

and logistic regression on a UK personal loan portfolio. We develop default 

prediction models and assess discrimination and calibration performance 

on a validation dataset. Results for credit scoring at an account level and prediction 

of defaults at a portfolio level are discussed. 

3 - When to rebuild and when to recalibrate credit scorecards 

Lyn Thomas, University of Southampton, United Kingdom, 

l.thomas@soton.ac.uk, Gimun Jung 

Credit scorecards "age" and need recalibrating and rebuilding from time to 

time. This is similar to the maintenance and replacement problem which has 

become standard for physical equipment. We identify a simple way of describing 

the state of the scorecard in terms of two parameters which are used to 

construct the log odds to score graph, and then model the dynamics of these 

parameters from real data sets. Thus we are able to build a dynamic programming 

model of when one should recalibrate and when one should rebuild a 

scorecard so as to minimise the expected total cost of running the system. 

217


4 - Support of managerial decision making processes by 

transductive learning 

Hubertus Brandner, Universität Hamburg, Germany, 

hubertus.brandner@studium.uni-hamburg.de, Stefan Lessmann 

This study analyses to which extent the promising findings of transductive approaches 

can be transfered to business problems of classification. Different 

variants of Support Vector Machines are examined to compare the established 

inductive learning and the transductive technique. To that end a hybrid metaheuristic 

is implemented to solve the mathematical programming formulations 

in the same way. Empirical results confirm the potential of transductive inference. 

Therefore it is advisable to utilize the information of unlabeled data in 

the context of managerial decision making and planning. 

� WA-21 


6.2.47 

Optimization Algorithms I 



Chair: Simone Garatti, Dept. of Electronics and Information, 


sgaratti@elet.polimi.it 

1 - Solving uncertain programs via the scenario approach: 

the FAST algorithm 

Simone Garatti, Dept. of Electronics and Information, 


sgaratti@elet.polimi.it, Algo Carè, Marco Campi 

Uncertainty is ubiquitous in decision problems, and this leads naturally to uncertain 

programs (UP). Robust and chance-constrained solutions to UP can be 

difficult to obtain in general. In this talk, we discuss the use of the scenario 

approach, a handy methodology based on random sampling of constraints, to 

solve UP with a guaranteed degree of approximation. In particular, we introduce 

FAST (Fast Algorithm for the Scenario Technology), a variant of the 

standard scenario algorithm with reduced complexity, which improves the applicability 

of the scenario methodology to a high extent. 

2 - Scheduling optimization in virtual enterprises based on 

the hybridization of a CSP with a genetic algorithm 

Rabah Kassa, mathematique, Universite Bejaia algerie, universite 

de bejaia 06000 bejaia algerie, 06000, bejaia, Algeria, 

rabah_kassa2002@yahoo.fr, Djamila Boukredera, Zaidi 

Sahnoun 

Production scheduling represents an important manufacturing function whose 

quality remains an essential stake for virtual enterprises. To optimize its 

scheduling, a virtual enterprise aims to improve its profitability while minimizing 

the customer’s service costs and respecting manufacturing constraints. 

This can be formulated as a CSP. We suggest an optimization method of the 

CSP based on the genetic algorithm. This hybridization aim at better taking 

over of this kind of problem defined by a large research space and a complex 

constraint set and finds solutions of good quality. 

3 - ParadisEO: a framework for metaheuristics 

El-ghazali Talbi, University of Lille - INRIA - CNRS, Lille, 

El-ghazali.Talbi@lifl.fr 

We present the ParadisEO white-box object-oriented framework dedicated to 

the reusable design of metaheuristics. It provides a broad range of features 

including population based metaheuristics and single-solution metaheuristics. 

It basedes on a conceptual separation of the solution methods from the problems 

they are intended to solve. The fine-grained nature of the classes allows a 

high flexibility. ParadisEO is of the rare frameworks providing most common 

parallel and distributed models; implementation is portable and models can be 

exploited transparently. 

218 

� WA-23 


6.2.49 

Model Selection in Regression Analysis 

Stream: Data Mining in the Financial Sector 








1 - Model generation and model selection in credit scoring 

Vadim Strijov, Computing Center of the Russian Academy of 



The credit scorecard is the logistic regression model; it maps the feature space 

to the probability of default of a banking client. A classical scorecard is constructed 

by an analyst, who manually selects informative features and creates 

combinations of them. We propose a new technique for the automatic scorecard 

construction. To develop a scorecard, one must assign a set of primitive functions 

and model generation rules. The result model is an admissible superposition 

of the primitive functions and features. The coherent Bayesian inference 

is used to select features and their superpositions. 

2 - Algorithms of feature selection for volatility estimation 

of European options 

Ekaterina Krymova, Control/Management and Applied 

Mathematics, Moscow Institute of Physics and Technology, 9,35 

b.3, Nagornaya st„Moscow, 141981, Moscow region, Dubna, 

Bogolubova 33-304, 117186, Moscow, Russian Federation, 

ekkrym@gmail.com 

The problem of multicollinearity is commonly encountered in regression analysis. 

This problem may lead to overfitting and result in unstable model parameters. 

New approach to the feature generation and feature selection was 

proposed. The feature generation technique is based on Kolmogorov-Gabor 

polynomial construction. The features are superpositions of primitive functions 

and free variables. The generated features require reduction of multicollinearity. 

For this purpose, the LARS modification is developed. Historical data of 

European options is used as practical example. 

3 - A topological approach to formulating conditions of the 

uniform convergence of frequencies to probabilities 

Michael Khachay, Ural Branch of RAS, Institute of Mathematics 

and Mechanics, S.Kovalevskoy, 16, 620990, Ekaterinburg, 

Russian Federation, mkhachay@imm.uran.ru 

Existence of the uniform convergence of frequencies to probabilities over an 

appropriate events class is a well known sufficient consistency condition of 

the empirical risk minimization (ERM) in machine learning. The traditional 

approach for proving such convergence is based on a sublinear growth of entropy 

of the event class in question and obtaining upper VCD bounds for this 

class. In this paper, existence of the uniform convergence of frequencies to 

probabilities over an event class is related to some topological properties of the 

sigma-algebra, induced by this class. 

4 - Benchmarking Framework for Financial Text Mining 

Caslav Bozic, Institute AIFB, IME Graduate School, Karlsruhe 

Institute of Technology (KIT), Institute AIFB - 05.20, KIT 

Campus South, 76128, Karlsruhe, BW, Germany, bozic@kit.edu 

Different data mining methods for financial text and various sentiment measures 

are described in the existing literature, without common benchmark for 

comparing these approaches. Implemented system (which is a part of FINDS 

Project) and proposed framework are based on theoretical data integration, and 

they facilitate combining more sources of financial data into comprehensive integral 

dataset. The dataset is then used to analyse the candidate measure by 

regressing it on different returns and other financial indicators that can be defined 

using the system’s novel data transformation approach.

� WA-24 


6.2.50 

Workforce Scheduling 2 



Chair: Shuangqing Liao, Laboratoire Genie Industriel, Ecole Centrale 

Paris, Grande Voie des Vignes, F-92 295 , CHÂTENAY-MALABRY 

Cedex, France, shuangqing.liao@ecp.fr 

1 - Call Center staffing with uncertain arrival rates 

Shuangqing Liao, Laboratoire Genie Industriel, Ecole Centrale 

Paris, Grande Voie des Vignes, F-92 295 , 

CHÂTENAY-MALABRY Cedex, France, 

shuangqing.liao@ecp.fr, Christian Van Delft, Ger Koole, Oualid 

Jouini 

We considers a multi-shift contract center staffing problem with two types of 

jobs: calls and emails.It is modeled as a newsboy-type model and is to minimize 

the salary cost and the unsatisfied service penalty. We suppose that all shifts 

are without breaks then the shift-period matrix is unimodularity. The pure call 

center integer staffing problem (no emails) can be relaxed to a linear problem 

with automatical integer solutions. We decompose our original problem by the 

benders’ decomposition. Thanks to the relaxation and decomposition, we can 

arise the shift and scenarios numbers. 

2 - A heuristic branch-and-bound approach for aircraft line 

maintenance staffing and scheduling 

Jeroen Beliën, Center for modelling and simulation, Hogeschool 

Universiteit Brussel, Stormstraat 2, 1000, Brussel, Belgium, 

jeroen.belien@hubrussel.be, Erik Demeulemeester, Brecht 

Cardoen 

In this paper, we investigate how a line maintenance service company should 

determine its staffing and scheduling decisions in order to minimize the resulting 

labor costs. We develop a heuristic branch-and-bound approach that 

was successfully applied to the problem setting occurred in a line maintenance 

service provider at Brussels National Airport. Next to this case study, the algorithm 

was tested on a large set of generated problem instances, based on 

real-life data. We present the computational results and managerial insights 

gained from this experiment. 

3 - Large-Scale Staff Scheduling in Airport Ground Handling 

Andreas Klinkert, Institute of Data Analysis and Process Design, 

Zurich University of Applied Sciences, Rosenstrasse 3, P.O. Box, 

CH-8401, Winterthur, ZH, Switzerland, 

andreas.klinkert@zhaw.ch 

An integer programming model is presented to solve a large-scale acyclic daysoff 

scheduling problem for multi-skill staff in airport ground handling. Special 

focus is given to a tractable formulation of the daily staffing level constraints 

in order to provide enough workers with appropriate skills for every combination 

of shifts. The developed model successfully solves the complex problem 

instances posed by the industrial project partner and CPLEX 11 generates high 

quality solutions within a few hours which clearly outperform the sophisticated 

solutions constructed manually by the planning experts of the ground handling 

company. 

4 - Appointment Scheduling with Discrete Random Durations 

and Applications 

Mehmet Begen, Richard Ivey School of Business, University of 

Western Ontario, 1151 Richmond St. N., Ivey, N6A3K7, 

London, ON, Canada, mbegen@ivey.uwo.ca, Maurice 

Queyranne 

We determine optimal appointment schedule in polynomial time for a given 

sequence of jobs (e.g., surgeries) on a single processor (e.g., operating room), 

to minimize the expected total underage (idle-time of the processor) and overage 

costs (waiting time of jobs and overtime of the processor) when each job 

has an integer random processing duration given by a joint discrete probability 

distribution. Besides surgeries, there are other applications such as project 

scheduling, container vessel and terminal operations, gate and runway scheduling 

of aircrafts in an airport. 


� WA-25 


6.2.48 

ROADEF/EURO challenge junior session 2 



Chair: Eric Bourreau, COCONUT, LIRMM, 161 Rue Ada, 34000, 

Montpellier, France, eric.bourreau@lirmm.fr 

1 - ROADEF/EURO Challenge 2010: A large-scale energy 

management problem with varied constraints 

Guillaume Dereu, EDF, France, guillaume.dereu@edf.fr, 

Christian Artigues, Eric Bourreau, H. Murat Afsar, Ender Ozcan 

We present the subject of the ROADEF/EURO 2010 challenge, an international 

optimization contest proposed jointly by EURO, the French OR society 

(ROADEF) and an industrial partner (EDF). The subject concerns an integrated 

nuclear power plant outage scheduling and electricity production planning 

problem, under several demand scenarios. The subject was announced in 

2009 at EURO Bonn. From 44 teams that initially registered, 20 teams are still 

competing for the final round. Results will be announced during the last session 

of the challenge stream. 

2 - A local search algorithm with a repair procedure for the 

Roadef 2010 challenge 

Henri Tokola, Department of Engineering Design and 

Production, Aalto University, School of Science and Technology, 

Puumiehenkuja 3, 02015, Espoo, Finland, henri.tokola@tkk.fi, 

Lauri Ahlroth, Andre Schumacher 

We present an algorithm for solving the energy management problem of the 

Roadef 2010 challenge. Our algorithm consists of multiple separate parts. 

First, it generates an initial solution using a backtrack method. Then we use 

a local search method to improve the initial solution. Due to tight constraints 

on the outage start dates, some candidates that are generated may be infeasible. 

In order to obtain a feasible solution, we modify the candidate using a repair 

procedure, which is based on the min conflict heuristic. 

3 - Constraint programming and local search for a largescale 

energy management problem with varied constraints 

Niels Kjeldsen, Dept. of Mathematics and Computer Science, 

University of Southern Denmark, Campusvej 55, 5000, Odense, 

Denmark, nhk@imada.sdu.dk, Steffen Godskesen, Rune Larsen, 

Thomas Sejr Jensen 

The ROADEF challenge 2010 asks to decide maintenance schedules, refuel 

amounts, and production levels of nuclear power plants. The combination of 

these decisions makes the problem difficult. We construct an initial maintenance 

schedule by constraint programming on a reduced model for one average 

scenario, and decide the refueling and energy production levels by a greedy 

heuristic. We improve the solution by local search on the maintenance schedule 

with feasibility of refueling and energy production ensured by the greedy 

heuristic. A final post-processing refines the solution for each scenario. 

4 - A constraint integer programming approach to solve 

large-scale energy management problem with varied 

constraint 

Stefan Heinz, Zuse Institute Berlin, Germany, heinz@zib.de, 

Thomas Schlechte 

In this talk, we present a constraint integer programming (CIP) model for the 

large-scale energy management problem of the ROADEF/EURO Challenge 

2010. Furthermore, we give a brief description of the main components and 

techniques used in the constraint integer programming solver SCIP, which has 

been submitted to the challenge. Finally, we introduce the extensions we used 

for the solver SCIP, to solve this class of problem efficiently. 

219


� WA-26 


3.1.11 

Special classes of cooperative games and 

allocation rules 



Chair: Marco Slikker, Department of Industrial Engineering, 

Eindhoven University of Technology, P.O. Box 513, 5600 MB, 

Eindhoven, Netherlands, m.slikker@tue.nl 

1 - A new bankruptcy rule emerged from an old cooperative 

value 

Mikel Álvarez-Mozos, Estadística e Investigación Operativa, 

Universidade de Santiago de Compostela, Rúa Lope Gómez de 

Marzoa, s/n, Campus sur, 15782, Santiago de Compostela, 

Afghanistan, mikel.alvarez@usc.es, Jose M a Alonso-Meijide, M a 

Gloria Fiestras-Janeiro 

Bankruptcy problems study situations in which a group of agents claim a proportion 

of a resource larger than the amount available. Different rules have been 

proposed to divide the amount available among the agents, and many properties 

have been discussed in the literature. Using axiomatic characterizations 

one can study the different features of each rule. In this work we propose a new 

bankruptcy rule and study its main features. Our rule corresponds to a widely 

known value of the cooperative game associated with the bankruptcy problem. 

2 - Reasonable costs allocation rules for ATM games 

Andrés Jiménez-Losada, University of Seville, 41092, Seville, 

Spain, hispan@esi.us.es, Casi Chacón, Esperanza Lebrón 

ATM games were introduced to define rules for allocating costs in a bank ATM 

(Automated Teller Machines) network. There are two different approachs from 

the cooperative game theory: Gow and Thomas (1998) [Naval Research Logistic 

45(4) 407-417], and Bjorndal et al. (2004) [Mathematical Methods of 

Operations Research 59(3) 405-418]. We propose a common point of view 

of these papers introducing them in a family of reasonable solutions for the 

problem. These new rules take in account the different behaviour of the banks 

according to the locations of their ATMs 

3 - About a new solution concept of cooperative TU-games 

Nadezhda Smirnova, Faculty of Applied Mathematics and 

Control Processes, Saint-Petersburg State University, Russian 

Federation, nadezhda.v.smirnova@gmail.com, Svetlana 

Tarashnina 

In our work we describe a new solution concept of cooperative TU-games, 

called a set of a-prenucleoli. A set of a-prenucleoli takes into account both constructive 

power and blocking power of coalition S with all possible ratios. We 

show that a-prenucleoli of an arbitrary n—person TU-game coincides with the 

prenucleolus of a certain n—person constant-sum game, which is constructed 

as the weighted sum of the game and its dual. Also we consider what the connections 

are between a set of a-prenucleoli and such solutions as prenucleolus, 

SM-nucleolus, the Shapley value. 

4 - Spare parts inventory pooling games 

Marco Slikker, Department of Industrial Engineering, Eindhoven 

University of Technology, P.O. Box 513, 5600 MB, Eindhoven, 

Netherlands, m.slikker@tue.nl, Frank Karsten, Geert-Jan van 

Houtum 

We study a situation where several independent companies separately stock 

spare parts of the same item for a technically advanced machine. They may 

reduce expected joint holding and downtime costs by pooling inventory. We 

analyze these situations by defining a cooperative cost game. For situations 

allowing companies to have non-identical demand rates and base stock levels 

and for situations allowing companies to have non-identical downtime costs, 

we show that the core of the associated game is non-empty. However, in general, 

the associated game may have an empty core. 


� WA-27 


8.2.06 

TRAFFIC AND ENVIRONMENT 



Chair: Ana Luísa Ramos, Economics, Management and Industrial 

Engineering, University of Aveiro, Campo de Santiago, 3810-193, 

Aveiro, Portugal, aramos@ua.pt 

1 - A dss solution for air pollution emergencies managing 

with traffic flows 

Valerio de Martinis, Depart.of Transportation Engineering, via 

Claudio21, 80125, Naples, Italy, vdemartinis@unina.it, Luigi 

Biggiero, Roberta Di Pace 

The main objective of this work is to analyze the correlation between the PM10 

and the climatic (wind, temperature, mm and days of rain) and traffic variables; 

moreover it focuses the attention on the effects of the adopted traffic regulation 

strategies by the municipality in the city of Naples. For this purpose, 

two approaches have been implemented (the ARIMA model - Autoregressive 

Integrated Moving Average - and the Artificial Neural Networks) in order to 

forecast the PM10 of the following days and the results have been discussed. 

Finally, a Decision Support System through Decision Maps approach has been 

implemented in order to evaluate the traffic strategies for avoiding PM10 emergencies. 

2 - Shortest Path Performance within Real Road Network 

in Malaysia Perspective 

Rohaizan Ramlan, Department of Technology Management, 

University of Tun Hussein Onn Malaysia, 86400, Batu Pahat, 

Johor, Malaysia, rohaizan@uthm.edu.my, Faiz Shamsuddin 

Computational performance of shortest path algorithm has been testing in many 

research (Cherkassky et al. 1993);(Zhan And Noon, 1998, etc. However most 

of the computational testing on shortest path algorithms has been based on randomly 

generated networks, which may not have the characteristics of real road 

networks (Zhan and Noon, 1998). The purpose of this study is to identify the 

performance of shortest path on Malaysia road network. The runtimes is collected 

to obtain the result and performance each of shortest path algorithm is 

known. 

3 - Optimal hierarchical system of a grid road network 

Masashi Miyagawa, Ecosocial System Engineering, University 

of Yamanashi, 4-3-11 Takeda, Kofu, 400-8511, Yamanashi, 

Japan, mmiyagawa@yamanashi.ac.jp 

This paper develops a simple analytical model for determining the hierarchical 

system of road networks. The model is based on a grid road network. We derive 

the optimal ratios of road areas that minimize the average and maximum 

travel time. Minimizing the average travel time provides an efficient solution, 

whereas minimizing the maximum travel time provides an equitable solution. 

Both of the solutions are expressed in terms of road widths and travel speeds. 

As an application of the grid network model, we evaluate the hierarchical system 

of the road network of Tokyo. 

4 - Model Based Systems Engineering for Traffic & Environment 

Ana Luísa Ramos, Economics, Management and Industrial 

Engineering, University of Aveiro, Campo de Santiago, 

3810-193, Aveiro, Portugal, aramos@ua.pt, José Vasconcelos 

Ferreira, Jaume Barcelo 

The immaturity of the Systems Engineering field argues for empirical research 

to drive knowledge evolution. In order to contribute to this development, it 

was decided to work out on a contemporary real world challenging problem. 

The application domain, "at the agenda’ of world leaders, national and local 

governors, and academia, is the system ’Intelligent Urban Traffic & Environment 

Operations’ which is characterized by significant dimension, complexity, 

interdisciplinarity, and relevant socio technical patterns. The emerging MBSE 

approach is being used to develop the proposed system.

� WA-28 



Scheduling with Transportation 



Chair: Gul Didem Batur, Industrial Engineering, Gazi University, 

Gazi University Engineering Faculty, Industrial Engineering 

Department, 06570, Ankara, Turkey, dbatur@gazi.edu.tr 

1 - Transporting Jobs Through a Processing Centre with 

Two parallel Machines 

Alan Soper, School of Computing and Mathematical Sciences, 

University of Greenwich, Old Royal Naval College, Park Row, 

SE10 9LS, Greenwich, London, United Kingdom, 

A.J.Soper@gre.ac.uk, Hans Kellerer, Vitaly Strusevich 

We consider the problem of scheduling jobs on two identical parallel machines, 

allowing preemption. The jobs are brought to the system by a single transporter 

and moved between the processing machines by the same transporter. The purpose 

is to split the jobs into batches and to find the sequence of moves of the 

transporter so that the time by which the completed jobs are collected together 

on board the transporter is minimal. We present an FPTAS for the problem. 

2 - A dynamic approach to hybrid flow shop scheduling 

with transportation requests 

Verena Gondek, Department of Mathematics, Universität 

Duisburg-Essen, Forsthausweg 2, LE 431, 47057, Duisburg, 

Germany, verena.gondek@uni-due.de 

This work is motivated by a real-life problem arising in steel producing industries. 

For monitoring the manufacture of steel, samples are taken at different 

stages of the production process and are sent to an automatic laboratory to 

check their quality. The efficient organization of the workflow in this laboratory 

can be classified as a hybrid flow shop scheduling problem with transportation 

requests and jobs arriving over time. Due to vast restrictions in computational 

time, we develop and evaluate a two phase heuristic approach to accomplish 

the required on-line optimization. 

3 - Multiple Part-Type, 3 Parallel Machine Scheduling in 

Robotic Cells 

Gul Didem Batur, Industrial Engineering, Gazi University, Gazi 

University Engineering Faculty, Industrial Engineering 

Department, 06570, Ankara, Turkey, dbatur@gazi.edu.tr, Serpil 

Erol 

We focus on the scheduling problem arising in 3-machine manufacturing cells 

which repeatedly produce a set of multiple part-types, where the parts are carried 

between the machines by a robot. Due to the flexibility property of CNC 

machines, robot may choose either to perform all the processing of a part completely 

on any one of the machines or to share them among the machines. Decisions 

to be made include finding the robot move cycle, the part sequence and 

the allocation that jointly minimize the production cycle time. The problem is 

modeled as a special travelling salesman problem (TSP). 

� WA-29 


8.2.11 

Simulation and Optimization Modeling in 

Finance 



Chair: Ronald Hochreiter, Finance, Accounting and Statistics, WU 

Vienna University of Economics and Business, Augasse 2-6, 1090, 

Vienna, Austria, ronald.hochreiter@wu.ac.at 

1 - No-Arbitrage Conditions, Scenario Trees, and Multi- 

Asset Financial Optimization 

Michael Hanke, Dept. of Banking and Finance, University of 

Innsbruck, Universitaetsstr. 15, 6020, Innsbruck, Austria, 

Michael.Hanke@uibk.ac.at, Alois Geyer, Alex Weissensteiner 


Many numerical optimization methods use scenario trees as a discrete approximation 

for the true (multi-dimensional) probability distributions of the problem’s 

random variables. Realistic specifications in financial optimization models 

can lead to tree sizes that quickly become computationally intractable. In 

this paper we focus on the two main approaches proposed in the literature to 

deal with this problem: scenario reduction and state aggregation. We show that 

neither of these methods is suitable to solve financial optimization models in 

asset-liability or portfolio management. 

2 - Multi-stage stochastic interest rate management 

Ronald Hochreiter, Finance, Accounting and Statistics, WU 

Vienna University of Economics and Business, Augasse 2-6, 

1090, Vienna, Austria, ronald.hochreiter@wu.ac.at 

A major part of a banks total interest rate risk is due to the position of nonmaturing 

deposits. In this talk, a multi-stage stochastic programming model 

for managing this risk factor will be shown. The uncertainty is given both in 

the interest rate development as well as the volume of the specific product. Different 

deposit products from an Austrian retail bank will be used to show the 

applicability of the model. 

3 - Empirical Analysis of two Bi-directional Online Trading 

Algorithms 

Esther Mohr, Information and Technology Management, 

Saarland University, P.O. Box 151150, 66041, Saarbrücken, 

Germany, em@itm.uni-sb.de, Günter Schmidt, Mike Kersch 

Trading decisions in financial markets can be supported by the use of online 

algorithms. We evaluate the empirical performance of two bi-directional trading 

algorithms and compare it to a moving average algorithm, a trading range 

breakout algorithm, an optimal algorithm, and to buy-and-hold. The algorithms 

are compared using geometric returns generated with historical DAX prices for 

the years 2000 to 2009. The performance of the algorithms found in the simulation 

runs is analyzed using a t-test and a bootstrap procedure. We also compare 

its performance to results from worst case analysis. 

4 - Accounting for defined benefit plans (IAS 19) — Information 

bias in case of a degenerating workforce 

Prof. Dr. Matthias Amen, Chair for Quantitative Accounting & 

Financial Reporting, University of Bielefeld, Universitaetsstrasse 

25, 33615, Bielefeld, Germany, Matthias.Amen@web.de 

Accounting for pension obligations is addressed in the international accounting 

standard (IAS) 19. The complexity in accounting arises from the difficulty of 

calculation, the stochastic nature, and the long forecast horizon. It is known 

that human expectations on system behavior generally fail under these circumstances 

(Tversky/Kahneman (1974)). Furthermore unforeseeable effects might 

happen, even if the assumptions are „best estimates" (Anderson (1992)). We 

present results of a simulation study and discuss the information bias in financial 

reporting that arises from the current IAS 19. 

� WA-30 


8.2.13 

MCDM 3 




Chair: Tihomir Hunjak, Faculty of organization and informatics, 

University of Zagreb, Pavlinska 2, 42000, Varazdin, Croatia, 

tihomir.hunjak@foi.hr 

1 - Computation of Non-Cooperative Equilibrium in Multicriteria 

Games 

Naouel Yousfi, Operational Research, Laboratory of Modelling 

and Optimization of Systems, Algeria, yousfi_na@hotmail.com, 

Mohammed Said Radjef 

In order to analyze several economic situations and model a lot of real world 

applications, finite multicriteria games can be used. 

This paper deals with the non cooperative multicriteria games. We propose to 

find a Non-cooperative equilibrium by transforming finite multicriteria games 

to an ordinal game using outranking methods. For the transformed game we developed 

an algorithm to find these equilibria if they exist. Finally, we illustrate 

the result given by the algorithm on a numerical example. 

221


2 - Multi-criteria semantic association ranking based on instance 

knowledgebase analysis for criminal organisation 

identification. 

Mariusz Chmielewski, Cybernetics, Military University of 

Technology, Poland, mchmielewski@wat.edu.pl, Rafal Kasprzyk 

The main method assumption shows that generation of semantic network based 

on registered atomic events, provides data, in which the proposed algorithms 

identify indirect links between the vertices of the network. Developed semantic 

model had been extended towards multi-criteria decision-making. Ranking 

semantic associations requires more than filtering for relevance but also must 

consider extended quantity analysis. Based on multicriteria optimization the 

evaluation function for semantic association is able to exploit weights for ranking 

criteria according to analyst preferences. This way ranking process can be 

parameterized distinguishing the relevance of the associations. 

3 - A Multi-Criteria Ranking Procedure for ERP Software 

Selection 

Tugba Efendigil, Yildiz Technical University, Turkey, 

tfdolas@yildiz.edu.tr 

Enterprise Resource Planning (ERP) software selection is one of the most important 

decision making issues covering both qualitative and quantitative factors 

for organizations. This study presents a beneficial structure to the managers 

for use in ERP software vendor selection process. We utilized grey relational 

analysis (GRA) to rank the ERP software vendors by making a multi-criteria 

weighted-average with respect to several criteria. In the end of this study, a 

numerical example is also presented to illustrate efficiency of the methodology 

and its applicability in practice. 

4 - Comparison of two models for tender evaluation in public 

procurement 

Tihomir Hunjak, Faculty of organization and informatics, 

University of Zagreb, Pavlinska 2, 42000, Varazdin, Croatia, 

tihomir.hunjak@foi.hr, Vjeran Strahonja, Dragutin Kermek 

In public sector procurement an open or restricted procedure are preferred if 

we are comparing them with a negotiation procedure. In this paper two models 

of public tendering of complex software will be presented. Both based on the 

economically most advantageous tender. First model is a model of open and 

restricted procedure of public procurement and second model is a model based 

on the negotiation procedure in public procurement. The models are different 

but focused on the same criteria. There are special conditions for selecting the 

negotiated procedure instead of usually used open or restricted procedure. The 

advantages and disadvantages of these approaches will be analyzed. 

� WA-31 


8.2.15 

Industrial applications of scheduling and 

routing I 



Chair: Geir Hasle, Applied Mathematics, SINTEF ICT, P. O. Box 

124 Blindern, NO-0314, Oslo, Norway, geir.hasle@sintef.no 

1 - Aspects of routing problems in media product distribution 

Geir Hasle, Applied Mathematics, SINTEF ICT, P.O. Box 124 

Blindern, NO-0314, Oslo, Norway, Geir.Hasle@sintef.no, 

Oddvar Kloster, Morten Smedsrud 

Efficient construction and revision of delivery routes for newspapers and other 

media products is critical. We illustrate important aspects of routing problems 

in this industry. Typically, the goal is to optimize on several criteria that are 

partly in conflict: number of routes, delivery costs, route balancing, and route 

segregation. We describe how these aspects may be modeled and solved. Results 

from computational experiments on industrial cases and standard benchmarks 

are presented. The importance of cloud computing and parallel algorithms 

for modern computer architectures is described. 

222 

2 - Optimization in Waste Management 

Jens Baudach, Lehrstuhl für Verkehrssysteme und -logistik, 

Technische Universität Dortmund, Leonhard-Euler-Str. 2, 44227, 

Dortmund, NRW, Germany, baudach@vsl.mb.tu-dortmund.de 

Waste management involves an efficient planning of the two major resources 

collection-vehicles and corresponding crews. As a result of an applied research 

project we present an integrated approach which first optimizes disposal districts 

(i.e., corresponding routes of the vehicles) and then finds optimal schedules 

for the crews. We focus on the crew scheduling phase which includes 

network based column generation, lagrangean relaxation models, and problem 

specific algorithms. Computational results based on practical data sets by public 

and private companies will be presented. 

3 - Solving a Solid Waste Collection Real Problem 

Nuno Lebreiro, Production and Systems Department, University 

of Minho, Portugal, Edf.Engenharia2, Campus de Gualtar, 

4710-057, Braga, Braga, Portugal, nflebreiro@gmail.com, José 

Oliveira, Manuel Figueiredo 

The large scale selective waste routing process congregates a lot of information 

and is subject to a huge number of constraints, which makes it a hard problem 

to solve by exact methods. This project is based on a real case scenario for one 

of the most important selective waste Portuguese operators. We focus on the 

Vehicle Routing Problem optimization using heuristics. The estimation of the 

solid waste generation is based on a model using demographic and economic 

data. Preliminary results will be presented and discussed. 

4 - Handling due-date Rendezvous in Vehicle Routing 

Problems 

Stanislas Francfort, CORE/M2V, Orange Labs, 38-40 rue du 

general leclerc, issy les moulineaux, France, 

stanislas.francfort@orange-ftgroup.com, Matthieu Chardy 

For Orange, the provision of services and the maintenance of networks involve 

dozens of millions technicians’ interventions a year. Thus the optimization of 

field interventions management is a key issue, which can be modeled as variants 

of the Vehicle Routing Problem (VRP). This work deals with the modeling of 

due-date Rendezvous (trainings, meetings, breaks or hours-off...) in VRP and 

their impact on the tractability of the subproblems when column generation 

based solution methods are in use. Results of extensive testing performed on 

both academic and real-life instances are presented. 

� WA-32 


8.2.17 

Meeting EURO-WG OR in Agriculture and 

Forest Management 



Chair: Lluis Miquel Pla, Mathematics, University of lleida, Fac. Dret 

i Economia, Jaume II, 73, 25001, Lleida, lmpla@matematica.udl.cat 

1 - Meeting EURO-WG OR in Agriculture and Forest Management 

Lluis Miquel Pla, Mathematics, University of lleida, Fac. Dret i 

Economia, Jaume II, 73, 25001, Lleida, 

lmpla@matematica.udl.cat 

Meeting EURO-WG OR in Agriculture and Forest Management 

� WA-33 


8.2.19 

Optimal policy in the energy markets 



Chair: Elisabetta Allevi, Quantitative Methods, University of Brescia, 

Contrada Santa Chiara, 50, 25122 , Brescia, Italy, allevi@eco.unibs.it

1 - A Quasi-Variational model of imperfect coordination: 

an illustration from electricity restructuring 

Giorgia Oggioni, Department of Quantitative Methods, 

University of Brescia, Italy, Contrada Santa Chiara, 50, 25122, 

Brescia, Italy, Italy, oggioni@eco.unibs.it, Yves Smeers, 

Elisabetta Allevi, Siegfried Schaible 

Recent papers propose methods for finding a set of solutions of quasivariational 

inequality problems. We apply these methods to a problem of imperfect 

coordination of operation encountered in the restructuring of the electricity 

system. We first discuss the economic interpretation of the variational 

inequality problem and some of its implications for algorithmic purposes. We 

then apply the methods to a set of case studies and report the results as well 

as the advantages and shortcoming encountered. The paper emphasises the 

numerical aspects. 

2 - Modelling Policies for the EU ETS allowances 

Paolo Falbo, Dipartimento Metodi Quantitativi, Universita’ degli 

Studi di Brescia, Contrada Santa Chiara, 50, 25122, BRESCIA, 

Italy, falbo@eco.unibs.it, Cristian Pelizzari 

In this work the price of allowances in the EU ETS is linked to the value of the 

expected penalty for the CO2 produced in excess by diverse economic players. 

Too severe environmental policies cause a rise in the price of allowances. 

The compliance of such objectives can increase the production costs causing 

inflation and loss of competitiveness. An equilibrium model is advanced where 

the regulatory authority can influence the choices of the players by tuning several 

parameters, such as the emission cap, the number of allowances and their 

allocation to the economic sectors, the penalty level. 

3 - Data Envelopment Analysis with undesirable and uncertain 

outputs 

Rossana Riccardi, Statistics and Applied Mathematics, 

University of Pisa, Via ridolfi, 10, 56124, Pisa, Italy, 

riccardi@ec.unipi.it, Roberta Toninelli 

In this paper a Data Envelopment Analysis (DEA) model with undesirable and 

uncertain outputs is presented. It is known that the production processes may 

generate undesirable externalities like emissions of greenhouse gases, sulfur 

oxides or waste generation. The aim of the model is to measure the performance 

of these processes highlighting, as a reference standard, those processes 

that combine greater amounts of desirable production with lower levels of undesirable 

outputs. The model is applied to a real case study related to the impact 

of CO2 emissions on industry sectors. 

4 - The "Invisible Hand’ for Risk Averse Investment in Electricity 

Generation 

Daniel Ralph, Judge Business School, Cambridge University, 

Trumpington St, CB2 1AG, Cambridge, United Kingdom, 

d.ralph@jbs.cam.ac.uk, Yves Smeers 

We consider a perfectly competitive situation consisting of an electricity market 

(2nd stage) preceded by investment in generating plant capacity (1st stage). 

The classical "Invisible Hand’ says that if generators and consumers act in their 

own best interests, the result will be to minimize the net cost (or max net welfare). 

Motivated by energy developments in the European Union, our interest 

is the case when electricity generators are risk averse, and the cost of future 

production is assessed via "coherent risk measures’ instead of expectations. 

� WA-34 


8.2.23 

Computational Methods 



Chair: Laura Martein, department of statistics and applied 

mathematics, University of Pisa, via Ridolfi, 10, 56124, Pisa, Italy, 

lmartein@ec.unipi.it 

1 - Semismooth Newton method for quadratic programs 

with bound constraints 

Anna Daryina, Department of Nonlinear Analysis and Safety 

Problem, Institution of Russian Academy of Sciences 


Dorodnicyn Computing Centre of RAS, Vavilova str, 40, 

Moscow, Russian Federation, daryina@ccas.ru, Alexey Izmailov 

We propose to solve convex quadratic programs with bound constraints by applying 

the semismooth Newton method (SNM) to the appropriate reformulation 

of the corresponding variational inequality. SNM was compared with the 

gradient projection method and active set methods. Numerical experiments 

demonstrate that for strongly convex problems, SNM can be considerably more 

efficient than the traditional approaches. 

2 - Solving large-scale nonnegative least squares using an 

adaptive non-monotonic method 

Suvrit Sra, AGBS, Max Planck Institute for Biological 

Cybernetics, Spemannstr. 38, 72076, Tuebingen, Germany, 

suvrit@gmail.com, Dongmin Kim, Inderjit Dhillon 

We present an efficient algorithm for large-scale non-negative least-squares 

(NNLS). We solve NNLS by extending the unconstrained quadratic optimization 

method of Barzilai and Borwein (BB) to handle nonnegativity constraints. 

Our approach is simple yet efficient. It differs from other constrained BB variants 

as: (i) it uses a specific subset of variables for computing BB steps; and 

(ii) it scales these steps adaptively to ensure convergence. We compare our 

method with both established convex solvers and specialized NNLS methods, 

and observe highly competitive empirical performance. 

3 - Block Hessian matrices with 0-1 quadratic convex reformulations 

Yasmin Rios-solis, graduate school of systems engineering, 

Universidad Autonoma de Nuevo Leon, facultad de Ingenieria 

Mecanica y Electrica, av. Unviersidad s/n, 65450, monterrey, 

nuevo Leon, Mexico, agueda.riossl@uanl.edu.mx 

Many combinatorial optimization problems can be formulated as a binary non 

convex quadratic function with linear constraints. When these problems involve 

parallel machines (scheduling) or zones (territory design), their Hessian 

have a block structure. To exactly solve these problems by a quadratic branch 

and bound, we propose convex reformulations that take advantage of the Hessian 

block structure. We aim a trade off between the quality (tightness of the 

continuous relaxation) and the computing time of these reformulations. Computational 

experimentation show the interest of this method. 

4 - Optimal Load Balancing Between Erlang-C Queueing 

Systems 

Jorge Sá Esteves, Dep. of Mathematics, University of Aveiro, 

Campus Santiago, 3810-193, AVEIRO, Portugal, 

saesteves@ua.pt 

In computer networking, load balancing is a technique to distribute workload 

evenly across two or more computers, network links, or other resources, 

in order to get optimal resource utilization using different criteria (maximize 

throughput, minimize response time, avoid overload, etc.). In this talk the overall 

system is modeled as set of independent Erlang-C queueing systems. The 

optimization problem solved may be seen as a multicriterion convex separable 

problem. Pareto Optimal solutions are computed and computational and 

graphical results are presented. 

� WA-35 


6.2.46 

MINLP: new developments and applications 



Chair: Sonia Cafieri, LOTA, Ecole Nationale d’Aviation Civile, 

31055, Toulouse, France, sonia.cafieri@recherche.enac.fr 

1 - Feasibility Based Bounds Tightening as a Fixed Point 

in the Interval Lattice 

Leo Liberti, LIX, Ecole Polytechnique, LIX, Ecole 

Polytechnique, 91128, Palaiseau, France, leoliberti@gmail.com, 

Pietro Belotti, Sonia Cafieri, Jon Lee 

223


Feasibility Based Bounds Tightening (FBBT) is a device used at every node 

of the spatial Branch-and-Bound algorithm for Mixed-Integer Nonlinear Programs 

in order to tighten the variable ranges. FBBT works by deriving implicit 

constraint bounds using the original variable ranges and interval arithmetic, 

and then propagating the given constraint bounds to hopefully tighter variable 

ranges using inverse interval arithmetic. This basic iteration is then applied 

until convergence. In order to address the problem of FBBT’s erratically slow 

convergence, we see FBBT as a monotone, deflationary operator on a lattice of 

interval vectors under set inclusion, intersection and interval union. We show 

that the sequence of intervals derived by the repeated application of FBBT converges 

to its greatest fixed point, which we compute by simply solving an auxiliary 

Linear Program. 

2 - Global Optimization of Large Scale Quadratically Constrained 

MINLP Models 

Christodoulos Floudas, Chemical Engineering, Princeton 

University, 08544, Princeton, NJ, United States, 

floudas@princeton.edu 

In this talk, we present novel theoretical and algorithmic results for the global 

optimization of large scale quadratically constrained MINLP models that arise 

in generalized pooling problems. Convex relaxations of bilinear terms via a 

variety of novel piecewise linear underestimations are presented and embedded 

within an effective branch and bound framework. Extensive computational 

results on case studies with distinct bilinear terms ranging from 48 to 1,260 

demonstrate the potential of the proposed deterministic global optimization approach 

to address large scale models to global optimality. 

3 - A cutting-plane framework for weakly-coupled 0/1 

SOCPs 

Sarah Drewes, Department of Mathematics, Technische 

Universität Darmstadt, Dolivostr. 15, 64293, Darmstadt, 

Germany, drewes@mathematik.tu-darmstadt.de, Sebastian 

Pokutta 

We devise a cutting-plane framework for a special class of mixed 0/1 second order 

cone programs, where the fractional and binary variables are solely coupled 

via the conic constraints. The derived cuts are based on an implicit Sherali- 

Adams reformulation and the generalized Benders cut. They are very effective 

as symmetric solutions are simultaneously cut off and each equivalence class 

of 0/1 solutions is visited at most once. We give preliminary computational 

results showing the effectiveness of our method. 

4 - Extensions of Interval Branch-and-Bound Algorithms 

for Mixed Global Optimization Problems with Real and 

Categorical Mixed Variables 

Frederic Messine, ENSEEIHT-IRIT, 2 rue Camichel, 31000, 

TOULOUSE, France, Frederic.Messine@n7.fr, Bernard Jeannet 

Real global optimization problems often imply to take into account different 

kinds of variables: discrete and continuous. In this presentation, we study 

mixed problems which combine real and categorical variables (discrete variables 

without ordering). Four methods are presented and discussed in order 

to compute bounds for the categorical variables. This yields some properties 

and permits some extensions of classical interval branch-and-bound global optimization 

algorithms. Numerical tests will validate our approaches and a real 

example of design is then considered. 

� WA-36 


3.1.05 

OR in the Public Sector 



Chair: Zilla Sinuany-Stern, Industrial Engineering and Management, 

Ben Gurion University, Beer-Sheva, Israel, 84105, Israel, 

zilla@bgu.ac.il 

1 - Efficiency evaluation of Greek engineering departments 

224 

Yorgos Goletsis, Dept. of Economics, University of Ioannina, 

University campus, GR 45110, Ioannina, Greece, 

goletsis@cc.uoi.gr, Vassiliki Brahou 

As Greek universities are public institutions there is a growing pressure for efficient 

use of public funds. Due to their non-profit nature financial measures of 

efficiency are not applicable nor a typical production function is available. We 

use Data Envelopment Analysis (DEA) for evaluating 30 Engineering departments. 

Additionally, the Malmquist Total Factor Productivity index is used for 

examining productivity change over time while the notion of super-efficiency 

is used for obtaining complete rankings. Quality considerations are also discussed. 

2 - Mixed integer least squares optimization for flight and 

maintenance planning of mission aircraft 

George Kozanidis, Mechanical Engineering, University of 

Thessaly, Leoforos Athinon, Pedion Areos, 38334, Volos, 

Magnisia, Greece, gkoz@mie.uth.gr, Eftychia Kostarelou, 

Andreas Gavranis 

We address the problem of generating a joint flight and maintenance plan for 

a unit of mission aircraft. We propose a mixed integer nonlinear programming 

formulation and an exact search algorithm for the solution of the problem. We 

analyze the computational complexity of this algorithm and we present results, 

which evaluate its computational performance. These results reveal that the total 

computational effort required for the solution of the problem depends mainly 

on the size of the unit and the space capacity of the maintenance station. 

3 - Assessing the Influence of Operating Theatres on the 

ICU Bed Occupancy 

Fermin Mallor, Statistics and O.R., Public University of Navarre, 

Campus Arrosadía, 31192, Pamplona, Spain, 

mallor@unavarra.es, Cristina Azcarate 

Usually, a big percentage of arrivals to Intensive Care Units (ICU) come from 

operating theatres. This high flow of patients influences both the ICU bed occupancy 

and the operating theater scheduling, leading sometimes to cancel operations 

or to reject emergency patients because of lack of beds. Here, a simulation 

model for the ICU and the operating theatres of the Hospital of Navarre is used 

to point out these influences and to assess the consequences of managerial rules. 

We use the information of all ICU patients from 2000-2009 to estimate patient 

arrivals and length of stays. 

4 - Coping with the Efficiency-Equity Dilemma in Organ 

Sharing 

Amir Alalouf, Industrial-Engineering and Management, 

Ben-Gurion University, PoBox 603, 84105, Beer-Sheva, Israel, 

aamir@bgu.ac.il, Israel David, Joseph Pliskin 

In order to compare organ allocation policies we developed the dynamic organsharing 

evaluation system (DOSES). The DOSES is simulation-based software 

designed to be a decision support system. The DOSES is modular, to enable 

easy updating of the various components and provides an easy methodic way to 

quantify allocation policies. In addition we will present a new allocation policy 

called the extended David and Yechiali (EDY), and we will uses the DOSES to 

compare EDY to other allocation policies, like UNOS (United Organ Sharing) 

and FCFT (first-come-first-transplanted). 

� WA-37 


3.1.09 

MCDA applications in agricultural and 

environmental management 



Chair: Antonio Boggia, DSEEA, University of Perugia, Borgo XX 

giugno, 74, 06121, Perugia, boggia@unipg.it 

1 - Development of a benefit transfer model using dominace 

based rough set approach 

Lucia Rocchi, DSEEA, University of Perugia, Borgo XX giugno 

74, 06121, Perugia, lucia.rocchi@unipg.it, Antonio Boggia, 

Salvatore Greco 

Benefit Transfer Methods (BTMs) are a cluster of methods for a second best 

evaluation of environmental goods. Although there are several studies regarding 

the theoretical basis of BTMs, their reliability and effectiveness are still not 

clear. However, BTMs are useful when saving time and money is important. 

Multi-criteria analysis, in particular the Dominance Based Rough set Approach 

(DRSA), provides a possible useful tool to improve the reliability of the methods. 

In particular, we used DRSA before utility function transfer to evaluate 

environmental and recreational services of a lake.

2 - Multiobjective optimization for farms using the 

dominance-based rough set approach 

Luisa Paolotti, DSEEA, University of Perugia, Borgo XX 

Giugno 74, 06121, Perugia, Italy, luisa.paolotti@gmail.com, 

Antonio Boggia, Salvatore Greco 

The objective of the paper is to underline the applicability of the MCDA method 

combining the Dominance-based Rough Set Approach with the Interactive 

Multiobjective Optimization (Greco et al., 2008) to the agricultural sector, in 

order to determine optimal planning strategies for farms. The method is a novelty 

in the multiobjective optimization sector. It has been applied within a case 

study, to determine an optimal strategy to integrate economic and environmental 

objectives. We developed a strategy able to conciliate high income with low 

nitrogen leaching, soil erosion and water consumption. 

3 - Assessing biodiversity and sustainability in natura 

2000 sites. a multicriteria approach 

Carla Cortina, DSEEA, University of Perugia, Borgo XX giugno 

74, 06121, Perugia, Italy, ccortina@unipg.it, Antonio Boggia 

This paper shows how Multicriteria Decision Analysis (MCDA) can help in 

a complex process such as the assessment of the level of biodiversity in the 

Nature 2000 sites. In order to assess biodiversity, a methodological approach 

based on multi-criteria analysis has been developed. The aim is to rank Nature 

2000 sites in order to understand the level of biodiversity and the environmental 

pressure of human activities which are subject. This is what decision-makers 

need for having support to their decisions and could be useful in order to promote 

sustainable development in protected areas. 

4 - Implementation of dominance based rough set approach 

module in a geographic information system 

Antonio Boggia, DSEEA, University of Perugia, Borgo XX 

giugno, 74, 06121, Perugia, boggia@unipg.it, Salvatore Greco, 

Gianluca Massei 

This paper presents the implementation of dominance based rough set approach 

module in an open source GIS system. Features, possible use and output of the 

implemented module are presented. To implement the algorithms GRASS 6.3 

has been used, adding a specific module written in C language. In addition, 

the new module has been included in the GRASS toolbox, in QGIS 0.10. We 

present the application of the implemented module to a case study. 

� WA-38 


6.2.44 

Issues in Inventory Management 

Applications I 

Stream: Inventory Management 


Chair: Thomas Wensing, Chair of Production and Operations 

Management, Catholic University of Eichstaett-Ingolstadt, Auf der 

Schanz 49, 84049, Ingolstadt, Germany, 

thomas.wensing@ku-eichstaett.de 

1 - Managing subassemblies and components for safety 

stock of multi products 

Emre Cakmak, International Logistics, Okan University, Okan 

University, Akfirat Kampusu, 34959, Istanbul, Turkey, 

emre.cakmak@okan.edu.tr, Hande Gulkac, Mehmet Tanyas 

The modern business environment is highly unpredictable. Safety stock plays 

an important role to avoid stock-out due to the uncertainty and variability of 

products demands and raw materials supply. High safety stock level for products 

and raw materials can cause an increment of holding and financial costs. 

Otherwise the condition of dissatisfaction of demands can result in the decline 

of the consumer service level which will lead consumer dissatisfaction in the 

long term. In Assembly to Order (ATO) production systems components and/or 

subassemblies are stocked. The advantage of stocking subassemblies rather 

than stocking components is the escalation of consumer reaction speed. The 

advantage of stocking components rather than subassemblies is having the least 

holding costs. The aim of this paper is to manage components and subassemblies 

for safety stock of multi products by evaluating stock out and inventory 

costs with the help of a simulation program. 


2 - Analysis of an inventory system when orders may 

crossover 

Thomas Wensing, Chair of Production and Operations 

Management, Catholic University of Eichstaett-Ingolstadt, Auf 

der Schanz 49, 84049, Ingolstadt, Germany, 

thomas.wensing@ku-eichstaett.de, Heinrich Kuhn 

We study a periodic-review order-up-to (r,S) inventory system with a stochastic 

lead time process that allows for orders to crossover. Customer orders arrive 

on a periodical basis. We show that performance indicators such as the ready 

rate, fill rate and mean inventory levels may exactly be calculated if customer 

order volumes are constant. For stochastic order volumes, the approach is approximate, 

however we observe fairly accurate results if the demand volatility 

is moderate. 

� WA-39 


6.2.45 

Real-time scheduling of logistic warehouse 

operations 

Stream: Scheduling under Resource Constraints 


Chair: Vicente Valls, Departamento de Estadística e Investigación 

Operativa, University of Valencia, Dr. Moliner,50, Burjasot, 46100, 

Valencia, Spain, Vicente.Valls@uv.es 

1 - Real-time scheduling of picking and storing operations 

in a warehouse 

M. Angeles Pérez, Mathematics for Ecomomy, University of 

Valencia, Avda Naranjos s/n, 46021, Valencia, Spain, 

angeles.perez@uv.es, Francisco Ballestin, M.Pilar Lino, 

Sacramento Quintanilla, Vicente Valls 

In this paper, we deal with the problem of scheduling picking and storing operations 

in a warehouse in which the utilisation of the two main resources — 

forklifts and storage locations available in a limited amount — has to be optimized 

for an efficient management of warehouse operations and where picking 

and storing orders may have due dates associated to them. We present several 

real-time dispatching policies for this problem. Computational experiments 

analyse the relative efficiency of the proposed policies and the benefits of using, 

or not, real-time information and/or dual routes. 

2 - Efficient utilisation of resources in the management of 

warehouse operations: storage location dispersion and 

forklift workload balance 

Sacramento Quintanilla, Economia Financiera y matematica, 

University of Valencia, Avda. Naranjos s/n, 46071, Valencia, 

Spain, Maria.Quintanilla@uv.es, Francisco Ballestin, M.Pilar 

Lino, M. Angeles Pérez, Vicente Valls 

The problem of scheduling picking and storing orders consists in: being given a 

set of orders, finding a set of storage locations, sequencing the fulfilment of the 

orders, and assigning a forklift and a route for each order minimizing the total 

flow time. This time is largely influenced by congestion. Congestion occurs 

when multiple orders are intended to be executed at the same time in the same 

area. We present an evolutive algorithm for assigning locations to orders which 

aims at maximising the dispersion of the selected locations and a minimum cost 

flow model for forklift workload balance. 

3 - An optimisation-based decision tool for the management 

of warehouse operations 

Jose Ignacio Llop, Matemáticas para la Economía y la Empresa, 

Universidad de Valencia, Avda. Tarongers s/n, 46022 , Valencia, 

Spain, j.ignacio.llop@uv.es, Francisco Ballestin, M.Pilar Lino, 

M. Angeles Pérez, Sacramento Quintanilla, Vicente Valls 

We present and describe a decision tool for the management of warehouse operations. 

The tool has been developed using commercial simulation software, 

models defined by different sets of parameters and optimisation algorithms. 

The tool visualises in 3D the structure and the logistic processes in a warehouse, 

statistically analyses the results obtained and uses several graphics to 

better understand the results of numerical optimization. The tool is intended 

for both practitioners training and research validation. 

225


4 - Modeling with simulation for the synchronization of 

manufacturing and warehousing activities in a paint 

production company 

Noyan Sebla GÜnay, Industrial Engineering, Okan University, 

Turkey, sebla.gunay@okan.edu.tr, Emre Cakmak, Mehmet 

Tanyas 

Nowadays, managing resources effectively and efficiently requires great importance. 

In this context, speed of products from production department and 

put away rate in warehousing must be equal to each other to avoid stock in the 

transition area between departments of production and storage. In this paper, 

a simulation model was developed for the synchronization of filling and warehousing 

departments of a paint manufacturing company which has Make To 

Stock policy. ABC rule is taken into consideration for shipping rate of products 

during the modeling. 

� WA-40 


6.2.52 

Berth Allocation in Maritime Container 

Terminals 



Chair: Luigi Moccia, Istituto di Calcolo e Reti ad Alte Prestazioni - 

ICAR-CNR, Consiglio Nazionale delle Ricerche, Via P. Bucci 41C, 

87036, Rende, Cosenza, Italy, moccia@icar.cnr.it 

Chair: Giovanni Giallombardo, Dept. of Electronics Informatics and 

Systems, University of Calabria, Via Pietro Bucci, Cubo 41C, 87036, 

Rende, Italy, giallo@deis.unical.it 

1 - Packing based formulations and local branching approaches 

to the solution of the berth allocation problem 

Pasquale Legato, Dipartimento di Elettronica, Informatica e 

Sistemistica, Università della Calabria, Via P. Bucci, 41C, 87036, 

Rende (CS), Italy, legato@deis.unical.it, Daniel Gulli’ 

The BAP focuses on the optimal matching between a set of space slots available 

along the quay and a set of incoming vessels to be berthed in these slots in 

a fixed time horizon. Starting from a so-called relative positioning formulation 

(packing based), we propose a new model where space slots are managed as 

discrete berthing points. It used to support the search process for an optimal 

solution to the original formulation. A local branching strategy, initialized by 

a greedy heuristics, is specialized for both models by suitable local branching 

cuts. Numerical results will be discussed. 

2 - The robust paradigm for the Berth Allocation Problem: 

a review and new perspectives 

Giovanni Giallombardo, Dept. of Electronics Informatics and 

Systems, University of Calabria, Via Pietro Bucci, Cubo 41C, 

87036, Rende, Italy, giallo@deis.unical.it, Giovanna Miglionico, 

Luigi Moccia 

It is well known that some critical sources of uncertainty may significantly 

affect an "optimal’ berth plan, leading, at runtime, to implement an adjusted 

solution which is actually far from optimality. This has motivated the recent 

interest in dealing with uncertainties in berth allocation models. In particular, 

the robust paradigm, based on the worst-case approach, seems to play a key 

role as it allows to construct berth plans that are less sensitive to uncertainties 

than minimal-cost ones. In such a context, we review the existing literature, 

and propose extensions and new perspectives. 

3 - Optimizing seaside operations at dedicated marine 

container terminals: The limited berth capacity case 

Georgios Saharidis, Civil Engineering, Rutgers University, 623 

Bowser Road, 08854, Piscataway, NJ, United States, 

saharidis@gmail.com, Mihalis Golias, Maria Boile, Sotirios 

Theofanis 

We present a mixed integer programming model to optimize the diverted demand 

from a dedicated to a multi-user terminal by minimizing the total handling 

and delayed departure costs for the vessels associated with the dedicated 

terminal operator. Handling costs for the diverted demand is a discrete function 

of the resources requested by the dedicated terminal operator. A combination 

of an evolutionary and exact algorithm is proposed to solve the resulting problem. 

The applicability of a game theoretic framework to model the behavior of 

the two players is also discussed. 

226 

� WA-41 


3.1.06 

Long Term Decisions in Forestry 

Stream: Long Term Financial Decisions 


Chair: Christian Clasen, Fachgebiet fuer Waldinventur und 

nachhaltige Nutzung, TU Muenchen, Am Hochanger 13, 85354, 

Freising, Germany, clasen@forst.wzw.tum.de 

1 - Learning to Formulate Forest Stand and Enterprise 

Planning Problems 

Ruth Dirsch, Bayerische Landesanstalt für Wald und 

Forstwirtschaft, 85354 , Freising, ruth@dirsch.com 

Formulations integrating stand treatment alternatives - meaning different thinning 

intensities - into optimisation approaches on forest enterprise level are difficult 

to find. Ways of solving the problem by dynamic programming as well as 

a linear programming for the overall enterprise optimisation will be discussed. 

The applied framework of different management alternatives on enterprise level 

yielded better results when incorporating thinning alternatives. 

2 - Real options, optimal forest rotation and long-term land 

conversion decisions: an afforestation case study for 

the Canadian Agriculture-Forestry interface 

Denys Yemshanov, Great Lakes Forestry Centre, Natural 

Resources Canada, Canadian Forest Service, 1219 Queen Street 

East, P6A 2E5 , Sault Ste. Marie, Ontario, Canada, 

dyemshan@nrcan.gc.ca, Geoff McCarney, Marty Luckert, Jim 

Unterschultz, Dan McKenny 

We have developed a spatial model that evaluates the decision to change land 

use between agriculture and forestry, inclusive of the option value to delay the 

land conversion decision. We used a modified Black-Scholes European spread 

option and illustrated the concept with an afforestation case study for Alberta, 

Canada. Compared with the scenarios based on static NPV criteria, the option 

value premium shortens the optimal rotation age if the NPV of alternative, 

non-forest land use is considerably higher. 

3 - On the use of the "forestry interest rate’ as approximation 

approach in an uncertain world 

Thomas Burkhardt, Campus Koblenz, IfM, Universitaet 

Koblenz-Landau, Universitätsstr. 1, 56070, Koblenz, Germany, 

tburkha@uni-koblenz.de 

In forestry it is widespread acknowledged that an interest rate below market 

rates, even below the risk free rate should be used in order to get valuations 

by the present value approach close to observed market prices. This so called 

"forestry interest rate’ has been subject to intensive debates. We show a real options 

approach to the valuation and demonstrate that the use of the traditional 

forestry interest rate in present values might be interpreted as a reasonable first 

approximation to the more demanding real options approach. 

4 - Long-term investment in forests under changing conditions 

— A risk/reward analysis 

Christian Clasen, Institute of Forest Management, Technische 

Universität München, Hans-Carl-von-Carlowitz-Platz 2, 85354, 

Freising, Bavaria, Germany, clasen@forst.wzw.tum.de 

This paper will focus on the financial consequences that today’s forest management 

decisions and the expected climatic change will have on the future generations 

of forest owners. These influencing factors and the long-term investment 

period of a forest exceedingly determine gains and losses. To illustrate the impact 

of changing conditions, a forest growth model was used to simulate the 

growth of different forest types within various districts of Bavaria, Germany. A 

Monte-Carlo simulation-based model has been constructed to consider timber 

price fluctuations and natural hazards.

� WA-42 


3.1.07 




Chair: Cheng-Ru Wu, Yuanpei University, Taiwan, 

alexru00@ms41.hinet.net 

1 - Modelling Airport Performance Trade-off Analysis 

Michael Madas, Department of Management Science and 


Evelpidon 47A and Lefkados 33, 11362, Athens, Greece, 

mmadas@aueb.gr, Konstantinos Zografos, Yiannis Salouras 

The airport decision making process necessitates the deployment of advanced 

models to support airport planning decisions for a total airport (airside and landside) 

by capturing the trade-off aspects among various measures of airport effectiveness 

(capacity, delays, level of service, safety, noise, cost-effectiveness). 

The objective of this paper is twofold: i) to describe the development process of 

a Decision Support System for total airport performance assessment and tradeoff 

modelling, and ii) to demonstrate the decision support capabilities and basic 

functionalities of the proposed system. 

2 - Considering Customers’ Preferences to Analyze the 

Development of the LEDTV Market in Taiwan 

Fang-Mei Tseng, International Business, Yuan Ze University, 

135 Yuan-Tung Rod., 320, Chung-Li, Taiwan, 

fmtseng@saturn.yzu.edu.tw, Yadi Lin 

Science and technology faces sharp variations in market surroundings and drastic 

competition, with innovative products being introduced into the market 

quickly; to satisfy the consumer’s needs and increase the firm’s revenue. When 

a firm wishes to develop a new product, understanding the desires of buyers 

in advance becomes particularly important. Furthermore, knowing about the 

substitution and diffusion of products is also very important, for it will affect 

companies as they make decisions regarding the quantity to ship.. The recent 

technological reports, indicate that light-emitting diode (LED) TVs will be the 

main new TV products in 2010. Therefore, taking the customer’s preferences 

into consideration in analyzing the market shares of different TVs in the future 

is becoming increasingly important to firms. In accordance with prior studies, 

some researchers have used the technological replacement model to forecast the 

shipments and market potential of TVs, but they do not consider the consumer’s 

real needs. Hence, we employ Conjoint Analysis to analyze the consumer’s 

preferences and Delphi to evaluate the variations in market share in Taiwan’s 

TV market for each different generation, i.e., CRT, CCFL-LCD, LED, OLED 

TVs. when 2010, 2015, and 2020. Furthermore, we combine the results of the 

Conjoint Analysis and Delphi into an Innovative Diffusion Model, to forecast 

the sales volume for each generation in Taiwan over the next 10 years. 

3 - Pareto-efficient sub-branches of the decision tree 

Toms Reizins, Ventspils University College, Latvia, 

toms.reizins@gmail.com, Andrejs Jaunzems 

Goal oriented planning means knowing potentially possible versions of action 

and choosing the most appropriate alternative. The graphical model of decision 

making problem is almost always shown as oriented graph or decision tree. In 

this article we look at decision tree with postulate stochastic alternative probabilities, 

different durations of processes and different time winnings. The article 

introduces Pareto (mu, lambda)-effective strategy concept, to improve determination 

methods of rational strategy. The summary of the proposed method is 

illustrated with an example. 

4 - Evaluate Outsourcing Commerce Semiconductor Industry 

Based on Grey Situation Decision-making 

Juo-Yi Sun, Yuanpei University, 306 Yuanpei St., Hsin Chu 300, 

Taiwan, R.O.C., Taiwan, rowena3056@mail.ypu.edu.tw, 

Chia-Chun Liao, Cheng-Ru Wu, Che-Wei Chang 

Grey situation decision-making (GSDM) for this purpose of this study focused 

on decision support tools, statistical process control (SPC) in high-tech semiconductor 

industry to collaborate Commerce Systems of the case studies. Grey 

situation decision-making (GSDM) is used to evaluate the synergy of a number 

of Commerce system characteristics the quality of performance. The main 

contribution of the outsourcing commerce system (OCS) can provide decisionmakers 

to do the quality control of internal referral decision. 


� WA-43 


8.2.02 

Operational Research in Sustainable Urban 

Development 

Stream: OR for Sustainable Development 


Chair: Marija Burinskien, Department of Urban Engineering, Vilnius 

Gediminas technical university, Saultekio ave. 11, LT-10223, Vilnius, 

Lithuania, marbur@ap.vgtu.lt 

1 - Application of operation research for sustainable urban 

development 

Marija Burinskien, Urban Engineering, Vilnius Gediminas 

Technical University, Saultekio av. 11, LT-10223, Vilnius, 

Lithuania, marija.burinskiene@vgtu.lt, Vitalija Rudzkien, Jrat 

Venckauskait 

The sustainable urban development covers the consensus of a wide range of 

activities and aimed at equivalent coordination of the impact of economic, social 

and environmental factors. The article integrates the theoretical attitudes 

related to future insights into the sustainable urban development. Conceptual 

modelling principles are applied to determine critical values of the indicators 

that characterise the sustainable development. A system that helps to define 

features of urban districts, development alternatives and achievable permitted 

marginal values of the living quality is created. 

2 - Multi-Actor Multi-Criteria Analysis: A case study on 

night time delivery for urban distribution 

Ellen Van Hoeck, MOSI-T, Vrije Universiteit Brussel, Pleinlaan 

2, 1050, Brussels, Belgium, Ellen.Van.Hoeck@vub.ac.be, Sara 

Verlinde, Annelies Heemeryck, Cathy Macharis, Frank Witlox 

This paper presents the multi-actor multi-criteria analysis (MAMCA) 

(Macharis, 2004) as the appropriate tool for measuring public support for nighttime 

delivery in urban surroundings as it enables to incorporate the views of 

different stakeholders. In this case we considered the receivers, the transport 

sector, society as a whole and the employees as the most important stakeholders. 

These stakeholders were interviewed on their attitude towards five different 

scenarios in which the time periods for deliveries and/or the accompanying 

measures differ. 

3 - Sustainability in supplier selection: application of the 

MACBETH methodology 

Luis Ferreira, DEGEI, Universidade de Aveiro, Campus 

Universitário Santiago, 3810-193, Aveiro, Portugal, 

lmferreira@ua.pt, Vitor André 

The present situation of the environmental degradation has caused a significant 

change in the business world. A critical process is the supplier selection process. 

It is a crucial activity in any company with a significant importance in 

the improvement of the environmental responsibility. In this research project 

we intend to integrate environmental criteria in the supplier selection process. 

Such task was elaborated using the MACBETH methodology. The above mentioned 

model was developed, validated and applied in the context of a company 

integrated in the automotive sector in Portugal. 

� WA-44 


8.2.03 

Vector and Set-Valued Optimization II 



Chair: Miguel Sama, Matemática Aplicada, Universidad Nacional de 

Educación a Distancia, CJuan del Rosal, 12, 28040, Madrid, Spain, 

msama@ind.uned.es 

1 - An Approximation Algorithm for Convex Multi-objective 

Programming Problems 

Matthias Ehrgott, Engineering Science, University of Auckland, 

Private Bag 92019, 1001, Auckland, New Zealand, 


227


We propose an algorithm for approximating the non-dominated set of a convex 

multi-objective nonlinear programme. The algorithm extends Benson’s outer 

approximation algorithm for multi-objective linear programmes. We prove that 

it computes a set of epsilon-nondominated points. For the case of differentiable 

objectives and constraints we describe an efficient way to perform the 

main step of the algorithm. We provide examples that show that this cannot 

always be done in the same way in the case of non-differentiable objectives or 

constraints. 

2 - Generalizing trade-off directions in multiobjective optimization 

Yury Nikulin, Department of Mathematics, University of Turku, 

Matematiikan laitos,Turun yliopisto, Vesilinnantie 5, 

Luonnontieteiden talo, 20014 , Turku, Finland, yurnik@utu.fi, 

Marko M. Mäkelä, József Mezei 

We consider a general multiobjective optimization problem with five basic optimality 

principles: efficiency, weak and proper Pareto optimality, strong efficiency 

and lexicographic optimality. We generalize the concept of trade-off 

directions defining them as some optimal surface of appropriate cones. We 

derive necessary and sufficient geometrical optimality conditions in terms of 

corresponding trade-off directions for both convex and nonconvex cases. 

3 - On epiconvergence for vector-valued functions 

Ruben Luis Lopez Montoya, Departamento de Matematica y 

Física Aplicadas, Universidad Catolica de la Santisima 

Concepcion, Alonso Ribera 2850, Campus San Andres, Facultad 

de Ingenieria, casilla 297, Cardenal Cisneros 448, Lomas de San 

Andres, 409-0541 , Concepcion, VIII Region, Chile, 

rlopez@ucsc.cl, César Gutiérrez 

The aim of this work is to study an epiconvergence for vector-valued functions 

where the image space is partially ordered by a (solid) pointed closed convex 

cone. We characterize this convergence and obtain some properties similar to 

those for the scalar case. Finally, we apply this epiconvergence for studying 

vector optimization problems. 

* This work has been supported by Proyecto FONDECYT 1100919 through 

CONICYT-Chile. 

4 - Graphical differentiation of the gradient 

Miguel Sama, Matemática Aplicada, Universidad Nacional de 

Educación a Distancia, CJuan del Rosal, 12, 28040, Madrid, 

Spain, msama@ind.uned.es, Bienvenido Jiménez, Vicente Novo 

A natural manner of defining a second order derivative notion for a nonsmooth 

map is by considering the derivative of the gradient in some sense. In this talk 

we follow this approach, in particular we study the properties of the contingent 

derivative of the Clarke gradient of a lipschitzian map at a point in order 

to define new notions of second order derivatives. We study the relationship 

of these derivatives with other notions of second order derivative and we give 

applications in optimization. An extension to the vector case is also explored. 

� WA-45 


8.2.12 

New technologies in facility Logistics 

Stream: Facility Logistics 


Chair: Kai Furmans, IFL, University of Karlsruhe, 

Gotthard-Franz-Str. 8, 76131, Karlsruhe, Germany, 

kai.furmans@ifl.uka.de 

1 - A Storage System with Virtual Aisles 

Kai Furmans, IFL, University of Karlsruhe, Gotthard-Franz-Str. 

8, 76131, Karlsruhe, Germany, kai.furmans@ifl.uka.de, Kevin 

Gue 

We describe an abstract model for a storage system with "virtual aisles." The 

system is based on a grid, which manages unoccupied cells to create aisle segments 

as needed in order to move desired items to an input-output boundary. 

Control of the system is completely decentralized; that is, each cell takes action 

based only on the states of cells in its neighborhood. The system has the 

curious property that higher loads lead to less congestion. 

228 

2 - Optimal zone boundaries in a Dense, Autonomous and 

Intelligent storage system 

Nima Zaerpour, Department of Management of Technology and 

Innovation, RSM Erasmus University, Rotterdam School of 

Management (RSM), Erasmus University Rotterdam Department 

of Management of Technology and Innovation, P.O. Box 1738, 

3000 DR, Rotterdam, Netherlands, nzaerpour@rsm.nl, Yugang 

Yu, René de Koster 

This paper studies a dense, autonomous and intelligent (DAI) storage system. 

Loads can move in x- and y-directions as long as an empty slot is available 

next to it. The system uses class-based storage; products with high turnover 

are stored in the zone closest to the output location. We derive the expected 

travel time of a random load from its storage location to the output. Using 

this expression we optimize zone boundaries, leading to the minimum travel 

time. Results show that, compared to random storage, class-based storage with 

optimal boundaries can significantly reduce the travel time. 

3 - The Internet of Things in Facility Logistics 

Moritz Roidl, Chair of Materials Handling and Warehousing, 

Technische Universität Dortmund, Emil-Figge-Str. 73, 44221, 

Dortmund, Germany, moro@flw.mb.tu-dortmund.de, Michael 

ten Hompel 

The Internet of Things (IoT) is a concept for the control of material flow 

systems in facility logistics that features decentralized control strategies and 

layouts. We have implemented a demonstrator that introduces a distributed 

Service-oriented Architecture (SOA) in combination with a multiagent-based 

control system. The real-time capabilities of the system can be demonstrated 

via an 3D visualization. We discuss the implications for analyzing such systems 

in respect to their real-time requirements and capabilities. In doing so, we 

focus on the interplay between material and information flow. 

4 - Analysis of the discrete-time GX|G[L,K]|1-queue 

Eda Özden, Institu für Fördertechnik und Logistiksysteme, 

Karslruher Institut für technologie, Germany, 

eda.oezden@ifl.uni-karlsruhe.de, Kai Furmans 

This work presents a discrete-time method to compute the waiting time distribution 

for the GX|G[L,K]|1-queue with batch arrivals and batch services under 

the minimum batch size policy. This is achieved based on a bivariate Markov 

chain. The quality of the analytical results are evaluated by means of simulation. 

Moreover, the waiting time distribution of a modified system with 

sufficiently many servers is derived. A number of applications of the introduced 

methods to transport systems are also discussed. Finally, the effect of 

the vehicle availibility on the throughput time is studied. 

� WA-46 


8.2.14 

Stochastic Programming Tools 



Chair: Tamas Szantai, Institute of Mathematics, Budapest University 

of Technology and Economics, Muegyetem rkp. 3., 1111, Budapest, 

Hungary, szantai@math.bme.hu 

1 - Ranking and Optimization PMS Algorithm with Traffic 

Forecasting 

Andras Bako, Information Science, Technical College of 

Budapest, 6, Doberdo str., udapest, Hungary, H1034, Budapest, 

bakoa@bmf.hu, Foldesi Peter, Szuts Istvan 

The Road Management Systems (and the PMS) usually do not take into consideration 

the future traffic change. The maintenance and rehabilitation actions 

and the development of the road network structure and the changing traffic 

structure modify the amount of the traffic on the road section. The deterioration 

process depends on mostly the volume of the traffic. That is why it is important 

to take into consideration the change of the traffic volume during the planning 

time horizon. Four different models are used for traffic forecasting and assignment. 

In the lecture some techniques are shown which handle this problem: in 

multiperiod, long time model at each planning period the traffic volume change 

is take into consideration. The objective is also changed, instead of the traditional 

minimization an other function is used, which can be transforming to a 

linear function. In ranking models the problem could be handled and solved. 

In the case of one period Markov stabile model there is nothing to do. In the 

multiperiod model the problem could be solved also.

2 - Modeling redial cellular mobile networks in random environment 

Mj Lopez-Herrero, Statistics and O.R. III, Universidad 

Complutense de Madrid, Escuela Univ. de Estadistica, 28040, 

Madrid, Spain, lherrero@estad.ucm.es, Jesus Artalejo 

Real systems often operate in varying environment conditions. In this talk, we 

show how the matrix-analytic formalism gives one the ability to construct and 

study versatile cellular mobile networks with user retrials operating in random 

environment. More specifically, we investigate two four-dimensional Markovian 

models which allow us to represent two different options for the use of the 

guard channel concept. We put emphasis on the numerical evaluation of the 

redial behavior and the environmental factors on the system performance. 

3 - Probabilistic constrained stochastic programming with 

copula generated multivariate probability distributions 

Tamas Szantai, Institute of Mathematics, Budapest University of 

Technology and Economics, Muegyetem rkp. 3., 1111, 

Budapest, Hungary, szantai@math.bme.hu, Edith Kovacs 

In this talk we will investigate the applicability of copula generated multivariate 

probability distributions in probabilistic constrained stochastic programming. 

In the first part of talk we will compare the bivariate Gumbel and Gaussian copulas. 

It is well known that the bivariate Gumbel family of copulas has upper tail 

dependence for any of its parameter values greater than one. In the same time 

the bivariate Gaussian family of copulas has upper tail independence for any of 

its parameter values (correlation coefficient) less than one. We will show how 

these facts influence the feasible domain of probabilistic constrained stochastic 

programming problems. In the second part of talk we will show how can be 

solved probabilistic constrained stochastic programming problems with different 

families of multivariate Archimedean copulas, like Gumbel, Clayton and 

Frank copula families. In fact, we do not need multivariate numerical integration 

at all when solving these types of problems. 

� WA-47 


8.2.16 

OR in Educational Problems and Systems 

Stream: Young People for System Theory, Optimization 

and Education 


Chair: Alexander Makarenko, Institute for Applied System Analysis, 

National Technical University of Ukraine "KPI", Prospect Pobedy 37, 

03056, Kiev, Ukraine, makalex@i.com.ua 

Chair: Alexis Pasichny, Students’ Science Association, National 

Technical University of Ukraine "Kyiv Politechnic Institute", 37, 

Peremohy av., 03056, Kiev, Ukraine, alexis.pasichny@gmail.com 

1 - Usage of OR methods for research of contemporary 

geopolitics 

Alexis Pasichny, Students’ Science Association, National 

Technical University of Ukraine "Kyiv Politechnic Institute", 37, 

Peremohy av., 03056, Kiev, Ukraine, 

alexis.pasichny@gmail.com 

Current report focuses on the numerical evaluation of the differences between 

global cultures. During the research different OR approaches are used: Analytic 

Hierarchy Process for development of the set of criterion, fuzzy logic 

methods for mathematical description of countries and cultures, expert evaluation 

methods for acquiring the quantified characteristics of culture differences. 

Measuring of culture differences allows to make conclusions about contemporary 

geopolitical picture of the world and check the hypothesis of "clash of 

civilizations’ made by Samuel Huntington. 

2 - Scenarios Development Methodology Based on SWOT 

and Morphological Analysis 

Kateryna Pereverza, Students Science Association, National 

Technical University of Ukraine, Kyiv, Revutskogo, 19/1, app. 

282, 02091, Kyiv, Ukraine, pereverza.kate@gmail.com 


In this paper methodology of scenarios development will be presented. Proposed 

methodology assume usage of sustainable development indicators at the 

stage of key variables’ identification, SWOT analysis at the stage of current situation 

description, morphological analysis at the stage of scenarios construction 

and analysis and expert evaluation methods for detecting of current trends 

and interests of external actors. Scenario approach allows to explore the possible 

future and get prepared for it. As an example the result of this methodology 

usage the scenarios for Ukraine will be described. 

3 - Social Indicators for Supply Chain Analysis 

Ana Carvalho, Management, CEG-IST, Av Prof Cavaco Silva„ 

2780-990, Porto Salvo„ Portugal, anacarvalho@ist.utl.pt, Ana 

Paula Barbósa-Póvoa 

The sustainability concept has been based on the triple bottom line (People, 

Planet, Profit). Work has been done in the economic and environmental areas. 

Methodologies and models have been proposed to address these two issues, 

however the social area has not received the necessary attention. For the full 

study of sustainability, it is urgent to include the social part. We propose to 

analyse the actual situation of the social analysis. Based on that analysis, a set 

of social indicators will be proposed to evaluate the supply chains performance 

in terms of social responsibility. 

� WA-48 


8.2.04 

Variational inequalities and applications to 

economic market models 

Stream: Iterative Methods for Economic Models 


Chair: Patrizia Daniele, Department of Mathematics and Computer 

Science, University of Catania, Viale A. Doria, 6, 95125, Catania, 

Italy, daniele@dmi.unict.it 

1 - An inverse variational inequality approach to the evolutionary 

spatial price equilibrium problem 

Laura Scrimali, DMI, Università di Catania, Viale Andrea Doria 

6, 95125, Catania, Italy, scrimali@dmi.unict.it 

It is well-known that the time-dependent spatial price equilibrium problem, can 

be transformed into and studied as an evolutionary variational inequality. However, 

in some situations, control policies may be imposed to the end of regulating 

the amounts of production and consumption. As a consequence, the problem 

becomes a time-dependent spatial price equilibrium control problem and 

is formalized as an evolutionary inverse variational inequality. The existence 

of solutions is then investigated and numerical examples are also provided. 

2 - Mitigating Global Supply Chain Risks through Corporate 

Social Responsibility 

Jose Cruz, Operations and Information Management, University 

of Connecticut, School of Business, 2100 Hillside Road, 

06269-1041, Storrs, CT, United States, 

jcruz@business.uconn.edu 

This paper presents a decision model that captures supply side disruption risks, 

social risks, and demand side uncertainty within an integrated global supply 

chain and corporate social responsibility (CSR) modeling and analysis framework. 

The global supply chain decision-makers must decide on the level of 

investment in CSR activities and the choice of trading partners (manufacturer 

or retailer) given their CSR consciousness and perceived riskiness in order to 

maximize profit and minimize their overall risk. The results show that CSR 

activities can be used to mitigate global supply chain risk. 

3 - Dynamic oligopolistic market equilibrium problem with 

long-term memory: Lipschitz continuity and infinite dimensional 

duality theory 

Annamaria Barbagallo, Department of Mathematics and 

Computer Science, University of Catania, Viale Andrea Doria, 6, 

95125, Catania, Italy, barbagallo@dmi.unict.it, Antonino 

Maugeri, Rosalba Di Vincenzo 

229

WB-02 EURO 24 - Lisbon 2010 

The aim of the talk is to provide a more realistic model for the dynamic 

oligopolistic market equilibrium problems. In particular, a long-term memory 

is introduced and the corresponding variational inequality model is discussed 

in order to study the problem in presence of delay. Moreover, existence and 

regularity results are proved, giving more relevance to a Lipschitz continuity 

theorem. Finally, applying the infinite-dimensional duality results, the existence 

of the Lagrange variables, which allow to describe the behaviour of the 

market, is provided. 

4 - Combined Methods for Dynamic Spatial Auction Market 

Models 

Adriana Gnudi, Mathematics, Statistics, Computer science and 

Applications, University of Bergamo, Via dei Caniana, 2, 24127 , 

Bergamo, Italy, adriana.gnudi@unibg.it, Igor Konnov, Elisabetta 

Allevi 

An equilibrium model for description of behavior of dynamic system of auction 

markets of a homogeneous commodity joined by transmission lines subject to 

joint balance and capacity flows constraints is suggested. The model involves 

the cost commodity storage within a given time period. An extended primaldual 

system of variational inequalities is given whose solutions yield an equilibrium 

trajectory of this system. Several splitting type methods are proposed 

to find its solution. 

230 

Wednesday, 10:40 - 12:00 

� WB-02 


3.2.14 

Keynote Talk 10 



Chair: José Paixão, Dept. Statistics and Operations Research, Faculty 

of Sciences - University of Lisbon, Bloco C6 - Campo Grande, 

1749-016, LISBOA, Portugal, jpaixao@fc.ul.pt 

1 - Exact and Heuristic Algorithms for the Vertex Coloring 

Problem 

Paolo Toth, DEIS, University of Bologna, Viale Risorgimento 2, 

40136, Bologna, Italy, paolo.toth@unibo.it 

Given an undirected sparse graph G = (V,E), where V is the vertex set and E the 

edge set, the Vertex Coloring Problem (VCP) requires to assign a color to each 

vertex in such a way that colors on adjacent vertices are different and the number 

of colors used is minimized. The VCP is a well known NP-hard problem, 

and has received a large attention in the literature, not only for its real world 

applications in many engineering fields (including scheduling, timetabling, register 

allocation, frequency assignment and communication networks), but also 

for its theoretical aspects and for its difficulty from the computational point of 

view. 

In this talk, we review the main Integer Linear Programming formulations, exact 

algorithms, lower bounding procedures, heuristic and Metaheuristic algorithms 

proposed for the VCP. Extensive computational experiments on benchmark 

instances from the literature are reported, comparing the performance of 

the most effective exact and heuristic algorithms. 

� WB-04 


3.2.13 

Assignment and clustering problems 



Chair: Daniel Tuyttens, Mathematics and Operations Research, 

University of Mons, FPMs-MATHRO, Rue de Houdain, 9, 6000, 

Mons, Europe, Belgium, daniel.tuyttens@umons.ac.be 

Chair: Skander Htiouech, SysCom-ENIT National Engineering 

School of Tunis, Tunisia, htiouechskander@yahoo.fr 

1 - Approximation algorithms for the multi-resource agent 

bottleneck generalized assignment problem 

Ozlem Karabulut, Industrial Engineering, Middle East Technical 

University, ODTU Endusrti Muh. Bolumu, Oda No:328, 06531, 

Ankara, Turkey, ozlem@ie.metu.edu.tr, Meral Azizoðlu 

In this study, we consider a Multi-Resource Agent Bottleneck Generalized Assignment 

Problem with equal-length periods. Our aim is to balance the workloads 

of the agents by minimizing the maximum workload over all agents. We 

first study the Linear Programming (LP) relaxation of the problem and introduce 

some valid cuts. We then develop a tabu search algorithm to find approximate 

solutions. Our computational results have revealed the satisfactory 

performance of our procedures. 

2 - A Preference Based Multiobjective Algorithm for the 

Clustering Problem 

Kerem Demirta¸s, Industrial Engineering, Middle East Technical 

University, ODTU Endustri Muhendisligi Bolumu, Oda: 326, 

06531, Ankara, Turkey, demirtas@ie.metu.edu.tr, Nur Evin 

Ozdemirel, Esra Karasakal

In this study, we develop a preference-based evolutionary algorithm for clustering 

problem. Clustering, which can be defined as unsupervised classification of 

data points into meaningful groups (clusters), is an NP-Hard problem. In addition 

to its computational complexity, the problem is also difficult in its concept 

since the definition of an optimal solution is not well defined. We include 

the preferences of the decision maker into a well known multiobjective evolutionary 

algorithm, namely SPEA2, using reference points and achievement 

scalarizing functions to find the best clusters. 

3 - Using surrogate information to solve the multidimensional 

multi-choice problem 

Skander Htiouech, SysCom-ENIT National Engineering School 

of Tunis, Tunisia, htiouechskander@yahoo.fr 

A new heuristic for solving MMKP is presented. We explore both sides of 

the feasibility border that consists in oscillating between constructive and destructive 

phases. Performance analysis of the method demonstrates the merits 

of using surrogate constraint information as choice rules for solving this problem 

class. A constraint normalization method is also used to strengthen the 

surrogate constraint information in order to improve the computational results. 

Numerical results show that the performance of this approach is superior to 

previously published ones and also much better efficiency. 

4 - A GRASP algorithm to solve the label printing problem 

Daniel Tuyttens, Mathematics and Operations Research, 

University of Mons, FPMs-MATHRO, Rue de Houdain, 9, 6000, 

Mons, Europe, Belgium, daniel.tuyttens@umons.ac.be, Arnaud 

Vandaele 

In this paper, the label printing problem, which consists in the assignement of 

a fixed number of labels on different templates in order to fulfill the order requirements 

and to minimize the waste excessive printed labels, is discussed. 

Allocating suitable labels to each template is a strategical task. The considered 

problem is hard and we propose a GRASP algorithm to solve it. The proposed 

algorithm is tested on some references instances, comparing the obtained results 

with those found in the literature. The results prove that the GRASP 

algorithm is partricularly well suited to this problem. 

� WB-05 


3.2.16 

Graph problems 



Chair: Maria Soto, Lab-Sticc, Universitée de Bretagne Sud, Rue 

Saint Maude, 56100, Lorient, Brittany (Bretagne), France, 

maria.soto@univ-ubs.fr 

1 - A metaheuristic approach to the graceful labeling problem 

of graphs 

Houra Mahmoudzadeh, Industrial Engineering, Sharif University 

of Technology, P.O.Box 4168685451, No 59, 96 Street, Golsar 

Avenue, Rasht, Gilan, Iran, +98131, Rasht, Gilan, 

h_mahmoudzadeh@alum.sharif.edu, Kourosh Eshghi 

In this paper, the well-known graceful labeling problem of graphs is represented 

as an optimization problem, and an algorithm based on Ant Colony Optimization 

metaheuristic is proposed for finding its solutions. Then, the proposed 

algorithm is applied to different classes of graphs, and the results are compared 

with the few existing methods in the literature. The computational results show 

that ACO metaheuristic is a powerful tool for finding solutions for the graceful 

labeling problem of graphs, and outperforms the other existing methods in 

certain classes of graphs. 

2 - New Methods to solve the Graph Partitioning Problem 

Mark Macedo, INESC Porto, Rua Dr. Roberto Frias, 378, ., 4200 

- 465, Porto, Porto, Portugal, mmacedo@inescporto.pt, João 

Pedro Pedroso, José Soeiro Ferreira 

The graph partitioning problem is a classical NP-Hard combinatorial optimization 

problem which consists of partitioning a graph into two subsets with the 

same cardinality, such that the number of edges whose endpoints are in different 

subsets is minimized. New methods are described and discussed, based 

on well-known heuristics like the min-max greedy, on tabu search, and on partial 

tree search. Computational results based on standard benchmark instances 

are reviewed, and an evaluation of the different methods, based on the overall 

computational experience is presented. 

EURO 24 - Lisbon 2010 WB-06 

3 - A metaheuristic for a graph clustering formulation 

Mariá C. V. Nascimento, Instituto de Ciências Matemáticas e de 

Computação, São Carlos, Universidade de São Paulo, Av. 

Trabalhador-sãocarlense, 400, 13560-970 , São Carlos, São 

Paulo, Brazil, mariah.cris@gmail.com, Franklina Toledo 

Graph clustering is a research topic which can be approached using several 

mathematical formulations, mostly based on graph partitioning models. One 

example is the maximization of the modularity of graph partitions. In this formulation, 

the partitions are evaluated according to the expected number of connections 

inside the clusters. The solutions found by it are characterized by their 

high quality for different types of graphs. Due to the combinatorial nature of 

this problem, this study presents a metaheuristic hybridized with the spectral 

relaxation for this formulation. 

4 - Hybrid Electromagnetic-like metaheuristic for the Generalized 

Minimum Spanning Tree Problem 

Jouhaina Chaouachi Siala, Laboratory of Mathematical 

Engineering, Ecole Polytechnique de Tunisie, 2078, La Marsa, 

Tunisia, chaouachijouhaina@yahoo.fr, Ichraf Zaidi 

In this paper we investigate the Generalized Minimum Spanning Tree Problem 

(GMST) which requires spanning at least one vertex out of every set of 

disjoint vertices in a graph. For this NP-Hard problem, we develop a hybrid 

electromagnetism-like meta-heuristic (HEM). We describe also a Hybrid 

Discrete Particle Swarm Optimisation meta-heuristic (HDPSO). Our computational 

study through benchmark instances reveals that these newly developed 

methods are a powerful device to solve combinatorial optimization problems. 

The proposed HEM outperforms state-of-art existing meta-heuristics. 

� WB-06 


8.2.30 

DEA Application VI — Utilities 



Chair: Rui Marques, Department of Civil Engineering and 

Architecture, CESUR-IST, Technical Univeristy of Lisbon, Av. 

Rovisco Pais„ 1049-001, Lisbon, rcmar@civil.ist.utl.pt 

1 - Economies of scope in the Portuguese water utilities 

Pedro Carvalho, Departamento de Engenharia Civil e 

Arquitectura, Instituto Superior Técnico - Universidade Técnica 

de Lisboa, CESUR, Av. Rovisco Pais, 1049-001, Lisboa, 

Portugal, pcarvalho@civil.ist.utl.pt, Rui Marques 

This paper applies different non-parametric techniques to analyze the existence 

of economies of scope in the water sector in Portugal. We employed the 

approach of Morita (2003), the methodology proposed by Daraio and Simar 

(2005) which analyzes the influence of exogenous variables in the production 

process and a non-parametric method based on the Data Envelopment Analysis 

(DEA) and bootstrap simulation techniques, proposed by Daraio and Simar 

(2007).We make some judgments about the different techniques available in 

the literature to compute the economies of scope. 

2 - Comparing public and private urban waste services in 

Portugal 

Pedro Simões, Center of Urban and Regional Systems (CESUR), 

Instituto Superior Técnico, Technical University of Lisbon, 

Instituto Superior Técnico, Technical University of Lisbon, Av. 

Rovisco Pais, 1049-001, Lisbon, Portugal, 

psimoes@civil.ist.utl.pt, Rui Marques 

The current paper presents a research that compares the performance of the 

private urban waste utilities with the public urban waste utilities in Portugal. 

It comprises private and public "wholesale’ and "retail’ companies. We apply 

the non-parametric frontier benchmarking technique of data envelopment analysis 

(DEA) to analyse the performance of more than 100 urban waste services 

which encompass more than 60% of the Portuguese population in the ’retail’ 

segment. For the wholesale segment the sample includes all the companies of 

the country. The outcomes pointed out a significant inefficiency and showed 

that private companies perform better than the public ones. 

231


3 - The application and utilization of data envelopment 

analysis for evaluating the cost efficiency of brazilian 

electricity distribution utilities 

Marcus Vinicius Pereira de Souza, Engenharia Industrial, 

PUC-Rio, Rua Marquês de São Vicente 225 - Gávea, 22451-041 

, Rio de Janeiro, RJ, Brazil, mvinic@engenharia.ufjf.br, 

Madiagne Diallo, Reinaldo Souza, Tara Keshar Nanda Baidya 

In this paper we describe the use of Data Envelopment Analysis (DEA) for 

measuring the efficiency scores of 60 Brazilian electricity distribution utilities. 

In this framework, we develop an alternative approach based on cluster analysis 

and Cone-Ratio Method. It is worth pointing out that these developments can 

reduce the information asymmetry and improve the regulator’s skill to compare 

the performance of the utilities, a fundamental procedure in incentive regulation 

squemes. Last, we examine the problem of detecting influential observations. 

4 - Evaluating police stations’ performance via DEA 

Zilla Sinuany-Stern, Industrial Engineering and Management, 

Ben Gurion University, Beer-Sheva, Israel, 84105, Israel, 

zilla@bgu.ac.il, Doron Alper, Vered Berdugo-Kushnir 

This study estimates the relative productivity of 60 police stations in Israel 

during 2006-2007, based on two inputs and 16 outputs, using Data Envelopment 

Analysis (DEA). We utilized several DEA versions: constant and variable 

return to scale, with and without bounds on the virtual variables. Moreover, 

benchmark analysis was done. The police stations were ranked via cross efficiency 

analysis, and the Maverick index was derived for each station. The validation 

of DEA was implied when the various models were highly correlated. 

Regression analysis did not depict external factors to explain the variability of 

the efficiencies of the police stations. 

� WB-07 


8.2.47 

Scheduling Approaches for Complex 

Manufacturing Systems 



Chair: Lars Moench, Chair of Enterprise-wide Software Systems, 


lars.moench@FernUni-Hagen.de 

1 - Lot-Streaming and Just-In-Time Scheduling on Identical 

Parallel Machines with a Common Due Date 

Oncu Hazir, Laboratoire d’Informatique de Paris 6, France, 

oncu.hazir@lip6.fr, Kedad-sdihoum Safia 

In this research, we examine an integrated lot-streaming and just-in-time 

scheduling problem with a common due date. We address scheduling customer 

orders in batches of limited size in production systems with single or identical 

parallel machines. The objective is to find a feasible schedule to meet the requirements, 

which satisfies the batch size constraints and minimizes the sum 

of tardiness and earliness penalties and setup costs, which involves the cost of 

creating a new batch. 

We present some structural properties of an optimal solution and analyze some 

polynomial cases and propose solution algorithms. 

2 - Batch scheduling problems with parallel machines and 

identical jobs 

Liji Shen, TU Dresden, Striesener Str. 14, 01307, Dresden, 

Germany, liji@liji.de, Lars Moench, Udo Buscher 

This study addresses a batch scheduling problem with parallel machines to 

minimize total weighted completion time. More precisely, we consider serial 

batching problems where sequence dependent family setup times are present. 

We focus on the special case of identical jobs in the same family, which is 

known to be relevant in many manufacturing industries. In our study, structural 

properties are first examined, based on which, we present both mixed integer 

programming formulations and heuristics. Computational experiments are also 

conducted to test the performance of the proposed approaches. 

232 

3 - A Column Generation Approach to Minimize Total 

Weighted Tardiness for Jobs on Parallel Machines 

Lars Moench, Chair of Enterprise-wide Software Systems, 


lars.moench@FernUni-Hagen.de, Timm Ziarnetzky 

A column generation (CG) approach is studied to minimize the performance 

measure Total Weighted Tardiness on parallel identical machines. We suggest 

an efficient method for the solution of the sub problems that is based on random 

key genetic algorithms. This method is appropriate for large scale instances 

and outperforms corresponding heuristics based on dynamic programming. We 

provide the results of computational experiments for randomly generated test 

instances that clearly show that the CG outperforms list scheduling approaches 

based on the Apparent Tardiness Cost dispatching rule. 

4 - A multistage mathematical programming based 

scheduling approach for lithography areas in complex 

semiconductor manufacturing systems 

Andreas Klemmt, Electronics Packaging Laboratory, 

Tu-Dresden, Helmholzstr. 18, 01062, Dresden, Germany, 

klemmt@avt.et.tu-dresden.de, Gerald Weigert 

Typically, a bottleneck of a wafer fab is the lithography area since of its expensive 

tools and complex resource constraints. In this research, a multistage 

mixed integer programming based optimization approach for planning such an 

area is presented. Thereby, existing process constraints like equipment qualification, 

resist availability, vertical dedication, mask availability are taken into 

account. Goals are the maximization of throughput, the minimization of setup 

costs and a balancing of machine utilization. Based on real manufacturing data 

the benefit of the proposed approach is examined. 

5 - Scheduling Jobs with Ready Times and Incompatible 

Families on Unrelated Parallel Batch Machines 

Lars Moench, Chair of Enterprise-wide Software Systems, 


lars.moench@FernUni-Hagen.de, Christian Almeder 

This research is motivated by a scheduling problem found in the diffusion areas 

of wafer fabs. We model the problem as an unrelated parallel batch machine 

problem with incompatible families and unequal ready times of the jobs. 

The objective is to minimize the total weighted tardiness (TWT). Given that 

the problem is NP-hard, we propose a Variable Neighbourhood Search (VNS) 

scheme and a greedy randomized adaptive search procedure (GRASP) that is 

based on the Apparent Tardiness Cost (ATC) dispatching rule. We compare the 

performance of the two heuristics by randomly generated test instances. 

� WB-08 


6.1.36 

Approximation and Competition 



Chair: Sergei Chubanov, University of Siegen, Hoelderlinstr. 3, 

Siegen, Deutschland, 57076, Siegen, Germany, 

sergei.chubanov@uni-siegen.de 

1 - Decomposed Software Pipelining for VLIW with precedence 

delays and resource constraints. 

Abir Benabid, Lip6, 104 av kennedy 75016 Paris, 104 av 

kennedy, 75016, Paris, France, abir.benabid@lip6.fr, Claire 

Hanen 

We consider the problem of scheduling loops on VLIW architectures used 

in embedded systems. We address the cyclic problem of finding periodic 

schedules with minimal period taking into account all constraints induced by 

uniform data dependencies and pipelined functional units. A guaranteed approach, 

called decomposed software pipelining (DSP), is extended to consider 

the above constraints. A theoretical worst case ratio is evaluated and the practical 

interest of DSP is established using real VLIW architecture (ST200 of 

STMicroelectronics) and a benchmark of graphs issued from ST compiler.

2 - Approximation results for the two machine crossdocking 

problem 

Christophe Rapine, G-SCOP, ENSGI-INPG, 38031, Grenoble, 

France, christophe.rapine@g-scop.inpg.fr, Damien Prot, Olivier 

Goldschmidt 

The two machine cross-docking problem is to schedule two sets of tasks, each 

set on a dedicated machine. Precedence constraints only exist from a task of the 

first set to task of the second. Minimizing the makespan is strongly NP-hard 

even in the UET case, but any greedy method yields a 2-approximation solution. 

We show that a r-approximation algorithm on a connected precedence 

graph leads to a (1+r/2)-approximation in the general case. Also we prove that, 

in the UET case, if the precedence graph has inner degree k, then there exists a 

linear time algorithm with guarantee (2-1/(k-1)). 

3 - Cooperation and Competition of Subcontracted Operations 

George Vairaktarakis, Operations, Case Western Reserve 

University, 44106, Cleveland, Ohio, United States, 

gxv5@case.edu 

Manufacturers capable to process their own jobs are willing to improve cost 

performance by subcontracting part of their workload to a single 3rd party (3P). 

Under competition, 3P announces a workload-based priority rule and the players 

decide the amount to subcontract so as to minimize their makespan. Alternatively, 

a central decision maker dictates the subcontracted amount and when 

it would be scheduled so as to maximize 3P utilization. We present polynomial 

equilibrium and centralized solutions, and transfer payment schemes that make 

all players better off to coordinate. 

� WB-09 


6.2.53 

New Frontiers in the Application of 










1 - Reverse Automatic Differentiation in Shape Optimisation: 

Non-Axisymmetric Vesicle Shape Parameterisation 

Case Study 

Redouane Boudjemaa, Department of electrical and electronic 

engineering, University Mhamed Bougara of Boumerdes, 

Avenue de l’independance, 35000, Boumerdes, Boumerdes, 

Algeria, rboudjemaa@umbb.dz 

This work investigates the implementation of Reverse Automatic Differentiation 

as a tool in shape optimisation. Non-axisymmetric shapes are accurately 

parameterised based on a numerical optimisation of a surface energy model. 

The RAD process evaluates derivatives by applying systematically the chain 

rules of differentiation to a function computer program. The smooth surfaces 

are produced as a solution to a bounded sixth order elliptic partial differential 

equation, controlling parameters are introduced through the boundary conditions. 

2 - The Layout Design of a Semiconductor Fab with Direct- 

Transport Guide Paths 

Ying-Chin Ho, Institute of Industrial Management, National 

Central University, No.300, Jhongda Rd, 320, Chung-Li, 

Taoyuan, Taiwan, ho@cc.ncu.edu.tw, Ta-Wei Liao 

One common way to arrange guide paths in a semiconductor fab is that intrabay 

guide paths of different bays are not directly connected and intra-bay guide 

paths are connected via a stocker. To eliminate negative effects of existing disadvantages, 

a new guide path arrangement has been used in some fabs. We 

study the layout problem with not only the intra-/inter-bay but also directtransport 

guide paths. Mathematical programming and heuristics are proposed. 

We minimize the total flow distance in the system. An example illustrates the 

effectiveness and applicability of our methods. 

EURO 24 - Lisbon 2010 WB-10 

3 - Column Generation based Primal Heuristics 

François Vanderbeck, Institut de Mathématiques de Bordeaux, 

Université Bordeaux1 & INRIA Bordeaux, 351 cours de la 

Libération, F-33405, Talence- CEDEX, France, 

fv@math.u-bordeaux1.fr, Cédric Joncour, Sophie Michel, Ruslan 

Sadykov 

Generic primal heuristics have made their way into commercial MIP solvers. 

Extensions to a column generation context are not straightforward: there are 

technicalities specific to the dynamic generation of variables. The column 

generation literature reports many application specific studies. There remains 

to extract generic methods that could be seen as black-box primal heuristics. 

We review generic classes of column generation based primal heuristics. We 

then focus on and test so-called “diving” methods and other forms of truncated 

Branch-and-Price. 

4 - New Mathematical Models for Balance Stability in U- 

Type Assembly Lines 

Banu Guner, Industrial Engineering, Anadolu University, Iki 

Eylul Campus, 26555, Eskisehir, Turkey, 

badogan@anadolu.edu.tr, Servet Hasgul 

In this study, some mathematical models are presented for U-type assembly 

line balancing problems. These models consider some ergonomic factors that 

may be encountered in work environments. It is seen that assembly line balancing 

will be more stable by taking care of these mentioned factors. The model 

parameters are examined as deterministic and fuzzy. Illustrative examples are 

presented to demonstrate the validity of the proposed models. 

� WB-10 


6.2.56 

Emerging Nonlinear Optimization 

Applications of OR 



Chair: A. Egemen Yilmaz, Electronics Engineering, Ankara 

University, Ankara Universitesi Tandogan Kampusu, Elektronik Muh. 

Bolumu, 06100, Ankara, -, Turkey, aeyilmaz@eng.ankara.edu.tr 




1 - Hierarchical Multi-player Game-solution Identification 

for Cooperative Advertising Supply Chain Using Multiobjective 

Particle Swarm Optimization 

Chie-bein Chen, International Business, National Dong Hwa 

University, 1, Sec. 2, Da-hsueh Rd. Shou-feng, 974, Hualien, 

Taiwan, cbchen@mail.ndhu.edu.tw, Ming-Hui Chen, 

Shiuan-Yiang Huang, Wei Ting Cho, Po-Liang Chao, 

Chung-Chang Lien 

We construct mathematic models in different market response functions associated 

with the hierarchical Stackelberg game structure, identify its equilibrium 

and explore some preference conditions for both supply chain players. An 

algorithm is proposed to optimize the vertical cooperative advertising multiobjective 

problems. We implement a real case and their numerical results are 

compared with different strategies which demonstrate the cooperative advertising 

and robustness of MOPSO-CD. 

2 - Robust R&D Project Management 

Ruken Duzgun, ISE Dept, Lehigh University, 200 W Packer Ave, 

18015, Bethlehem, PA, United States, rukenduzgun@gmail.com, 

Aurelie Thiele 

We consider robust optimization approaches to R&D project selection when 

investments are done in stages and cash flows are uncertain. We consider an 

approach with two ranges (high and low) at each time period and a parameter 

limiting the number of times the cash flow of the projects can be in the low 

range. The use of binary variables to represent project selection raises challenges 

to develop tractable robust counterparts. We discuss ways to address 

these issues and present theoretical insights as well as numerical results, and 

we provide extensions to the initial setup. 

233


3 - Novel Approaches in Particle Swarm Optimization and 

Potential Applications in OR 

Okkes Tolga Altinoz, Bala Vocational School of Industrial 

Electronics, Hacettepe University, 06100, Ankara, 

taltinoz@hacettepe.edu.tr, A. Egemen Yilmaz, Gerhard-Wilhelm 

Weber 

Particle Swarm Optimization is a simple but powerful heuristic technique for 

the solution of the complicated multidimensional optimization problems. Since 

its development many derivatives of PSO and its hybridizations with other 

methods have been proposed for different purposes. We will try to revise the 

most significant and interesting recent ideas in PSO; and we will try to identify 

and discuss the potential usages of these recent PSO formulations in the OR 

discipline. 

4 - A Stochastically Perturbed Particle Swarm Optimization 

for Idetical Parallel Machine Scheduling Problems 

Mehmet Sevkli, Industrial Engineering, Fatih University, Fatih 

University, Buyukcekmece, 34500, ISTANBUL, Turkey, 

msevkli@fatih.edu.tr 

A Stochastically Perturbed Particle Swarm optimization algorithm with a different 

search strategy is proposed for the identical parallel machine scheduling 

problems. The algorithm is applied to the non-preemptive parallel machine 

scheduling problem with the objective of minimizing makespan. SPPSO’s performance 

is compared with against two other recent PSO algorithms. The 

SPPSO algorithm performed better in terms of obtaining optimum solutions 

and consumes less time. 

� WB-11 


8.2.38 

Recent Advances in Quality Management 

and OR in Reliability Engineering 



Chair: Gulser Koksal, Industrial Engineering, Middle East Technical 

University, Inonu Blvd., 06531, Ankara, Turkey, 





1 - Improving the Quality of Plastic Injection Molding Using 

Taguchi Method and PCA 

Fong Jung Yu, Industrial Engineering and Technology 

Management, DaYeh University, 168, university road, Dacun, 

51591, Changhua, Taiwan, fischer@mail.dyu.edu.tw, Chingpou 

Chang, Kai-I Huang 

The combination of Taguchi method and principal component analysis (PCA) 

is used to promote product quality of conventional plastics injection molding 

process through the optimization of the parameters in this study. We found that 

PCA could be effectively used in multiple-quality-characteristics problems to 

obtain integral solution. Furthermore, holding pressure could be regarded as an 

adjustment factor to the process mean. From the result of analysis of variance, 

material temperature, holding pressure and injection location have significant 

impact on size of finished product. 

2 - Optimum process mean, specification limits, and manufacturing 

quantity based on quadratic loss function 

and rectifying inspection plan with perfect and imperfect 

reworks 

Muhammad Al-Salamah, Systems Engnieering, King Fahd 

University, Box 5067, KFUPM, 31261, Dhahran, Western, 

salamah@kfupm.edu.sa, Ashraf El-ga’aly 

In this paper, a modified economic manufacturing quantity (EMQ) model under 

the imperfect product quality is developed. Taguchi’s quadratic quality loss 

function is integrated into the model to evaluate the product quality. The imperfect 

reworks of product are considered in the modified EMQ model. By solving 

the modified model, we can obtain the optimum combination of the production 

quantity, process mean and the specification limits that lead to the maximized 

profit. A numerical example and sensitivity analysis of parameters of the model 

are provided for illustration. 

234 

3 - Ordering Policy For Spare Preventive Replacement 

Jih-An Chen, Department of Business Administration, Kao-Yuan 

University, NO.1821, Jhongshan Rd., Lujhu Township, 82151, 

Kaohsiung County, Taiwan, jachen@cc.kyu.edu.tw 

We develop a preventive replacement policy of spare ordering under cost effectiveness 

criterion. The spare unit for replacement is available only by order 

and the lead time for delivery the spare due to regular or expedited ordering 

follows general distributions. Introducing costs due to ordering, repairs, downtime 

and replacements, as well as the salvage value of system, we derive the 

expected cost effectiveness per unit time in the long run as a criterion of optimality. 

There exists a finite and unique optimum policy of ordering time which 

maximizes the expected cost effectiveness. 

4 - Redundancy Allocation for Serial Systems 

Kurtulus Baris Oner, Industrial Engineering, Eindhoven 

University of Technology, P.O. Box 513, 5600 MB, Eindhoven, 

Netherlands, k.b.oner@tue.nl, Geert-Jan van Houtum, Alan 

Scheller-Wolf 

We consider a situation in which a user buys a number of units of a serial system. 

The systems are supported by a single spare parts inventory stock point. A 

predetermined emergency procedure is performed on stock-out events. Three 

policies, which are different combinations of applying the emergency procedure 

and having a cold-standby unit, can be implemented per subsystem. We 

formulate a model to minimize the Total Cost of Ownership of the systems 

under a minimum availability constraint. Next, we conduct exact analysis and 

derive results on the optimality of the policies. 

� WB-12 


8.2.39 

AHP 06 


Processes 


Chair: Ahmet Kandakoglu, Department of Industrial Engineering, 

Istanbul Technical University, Macka, 34367, Istanbul, Turkey, 

kandakoglu@itu.edu.tr 

1 - Developing an Organizational Performance Model for 

Community Empowerment in Taiwan 

Pin-Ju Juan, Department of Tourism and Hospitality, Tamkang 

University, 180 Linwei Road., Chiao-hsi Shiang„ 26247, I-lan 

County, Taiwan, pj@ocu.edu.tw, Shin-Yi Lin, Benjamin J.C. 

Yuan 

This study intends to explore an evaluation model for effectiveness and satisfaction 

of the community empowerment. This study will also use focus groups 

approach to construct an assessment model. The purpose of this paper is to construct 

a performance measurement model combine balance scorecard (BSC) 

and Analytical Hierarchy Process (AHP). By applying the AHP to obtain factors 

and criteria weights, this model can assist decision makers or administrators 

of Community Development in assessing the performance and improving 

applicability for future use. 

2 - Development and Application of an Analytic Hierarchy 

Process Model for Assessing Airport Advanced Surface 

Movement Guidance and Control Systems 

Konstantinos Zografos, Department of Management Science and 


Evelpidon 47A & Lefkados 33, 11362, Athens, Greece, 

kostas.zografos@aueb.gr, Konstantinos Androutsopoulos 

This paper presents an AHP based methodology for the assessment of an Advanced 

Surface Movement Guidance and Control System (A-SMGCS). A- 

SMGCS provide ground traffic management services for the improvement of 

the efficiency and safety of airport operations. The proposed methodology involves 

an AHP model that takes into account multiple tangible and intangible 

criteria and indicators associated to the A-SMGCS assessment. The AHP 

model identifies the relative importance of the assessment criteria and indicators 

and provides a comparative assessment of the A-SMGCS against a baseline 

system.

3 - The Evaluation Analysis of RFID Adoption for Taiwan 

Logistic Industry 

Ling-Lang Tang, School of Management, Yuan Ze University, 

135 Yuan-tung road„ 320, Chung-Li, Taoyuan, Taiwan, 

balltang@saturn.yzu.edu.tw, Cheng-Chuang Hon 

This study adopts TOE as the framework that included four dimensions and 

33 criteria in assisting logistic managers evaluating RFID adoption. Firstly, 18 

evaluation criteria were selected by experts from 33 decision criteria of RFID 

implementation. The selected criteria include four dimensions: environment, 

technology, organization, and cost. Secondly, the FAHP weights evaluate and 

judge the key success factors of RFID adoption. Finally, it explores the importance 

factors affecting RFID implementation among three phases. 

� WB-13 


2.2.21 

Heuristics in location 



Chair: Said Salhi, Kent Business School, University of Kent, Centre 

for Heuristic Optimisation„ Canterbury, Kent, CT2 7PE, United 

Kingdom, s.salhi@kent.ac.uk 

1 - Evaluation of an Ant Based Hybrid Metaheuristic to 

solve the Capacitated Facility Location Problem 

Harry Venables, Business School, University of Sunderland, Reg 

Vardy Centre, St Peter’s Campus, SR6 0DD, Sunderland, United 

Kingdom, harry.venables@sunderland.ac.uk, Alfredo 

Moscardini 

Hybrid metaheuristics belong to an emerging area that makes use of integrating 

exact and/or approximate techniques. Although this approach has been attempted 

before for the capacitated facility location problem, previous attempts 

suffered from excessive run-times. A hybrid ant colony optimisation algorithm 

that integrates a pheromone based model with an exact solution technique is 

proposed. Also, a simple binary flip local search mechanism which makes use 

of the same exact solver is integrated within the design. Run-time distributions 

are generated and compared with other algorithms. 

2 - Goal Programming in Location Problems 

Juraj Pekár, Department of Operations Research and 

Econometrics, University of Economics, Dolnozemska 1, 85235, 

Bratislava, Slovakia, pekar@euba.sk, Ivan Brezina, Zuzana 

Cicková 

Solving the location problems often involves the preference of locality that is 

based on certain criteria (largeness of a node, population density, waste concentration 

etc.). For that reason it seems be effective to use the modification that 

is based on multi-criteria decision analysis, concrete the goal programming. 

Via goal programming it is possible to specify such variant that comes close 

to real needs. Provided that the considered problems are equally structured, 

constructed problem might be of their combination. 

3 - Formulation space search for Location-Allocation problem 

Nenad Mladenovic, Faculty of Organizational, University of 

Belgrade, 11000, Belgrade, Serbia, nenad.mladenovic@gerad.ca, 

Jack Brimberg 

For solving the multi-source Weber problem we suggest heuristic method based 

on the recent Formulation space search approach. In our new local search we 

explore the fact that continuous and discrete formulations of the problem (i.e., 

the p-median) tend to be the same if the number of possible location points 

in p-median tends to infinity. So, in each iteration we switch from continuous 

formulation to discrete, adding new (non-occupied) Weber facilities to the set 

of potential p-median sites. Encouraging computational results are reported. 


� WB-14 


2.2.15 

Pension funds 

Stream: Actuarial Sciences and Stochastic Calculus 


Chair: Elena Vigna, Dipartimento di Statistica e Matematica 

Applicata, Università di Torino and Collegio Carlo Alberto, corso 

Unione Sovietica 218 bis, 10135, Torino, Italy, 

elena.vigna@econ.unito.it 

1 - Optimal asset allocation and sharing rule for pension 

funds in DC and DB frameworks 

Francesco Menoncin, Economics, Brescia University, Via S. 

Faustino, 74/B, 25122, Brescia, Italy, menoncin@eco.unibs.it 

We study both the optimal asset allocation and the optimal share of performances 

for a pension fund which operates in a complete financial market where 

the prices of risky assets are driven by a set of stochastic variables. Furthermore, 

we take into account a stochastic force of mortality whose risk (i.e. both 

mortality and longevity risk) can be hedges through a demographic asset as a 

derivative on this force of mortality. The framework is useful for both a defined 

contribution and a defined benefit pension scheme. We demonstrate that it is 

optimal to share all the wealth exceeding the prospective mathematical reserve 

and that the optimal portfolio must contain a component which hedges against 

the changes in such a reserve due to the changes in the values of the stochastic 

state variables. 

2 - Demographic Risk Sharing in Defined Contribution 

Pension Funds 

Martino Grasselli, Dipartimento di Matematica Pura ed 

Applicata, Università degli Studi di Padova and ESILV, via 

Trieste 63, 35121, Padova, grassell@math.unipd.it, Daniel 

Gabay 

We consider the problem of a defined contribution fund manager looking for a 

fair remuneration rule for all the retirees by smoothing the fluctuations arising 

from the non stationarity of the fund population. We find that the optimal remuneration 

policy consists in delivering the guarantee plus a surplus, which is 

a (positive) random amount depending on market performance. In this case the 

benefits turn out to be volatile. 

3 - Constrained portfolio choices in the decumulation 

phase of a pension plan 

Fausto Gozzi, Dipartimento di Scienze Economiche e Aziendali, 

Luiss University - Roma - Italy, viale Romania 32, 00197, Roma, 

RM, Italy, fgozzi@luiss.it, Salvatore Federico, Marina Di 

Giacinto, Elena Vigna 

We deal with a constrained investment problem for a DC pension fund in the 

decumulation phase. We consider the basic model of GHV (2004) with fixed 

consumption and annuitization time. Firstly, we require no short-selling. Secondly, 

we impose a final capital requirement, implying no ruin. We approach 

the constrained stochastic control problem with dynamic programming. We 

give a general result of existence and uniqueness of regular solutions of HJB 

equation. In a special case we explicitly compute value function and optimal 

policy and show a numerical application. 

4 - Stochastic pension funding when benefits follow a 

jump diffusion process 

Juan Pablo Rincon-Zapatero, Economia, Universidad Carlos III 

de Madrid, c/ Madrid, 126, 28903, Getafe, Madrid, Spain, 

jrincon@eco.uc3m.es, Ricardo Josa-Fombellida 

We consider a defined benefit stochastic pension fund where benefits follow a 

jump diffusion process. The fund manager invests in a risky portfolio and selects 

an amortization rate to keep stable the fund evolution within prescribed 

targets. The problem is solved analytically by means of the dynamic programming 

approach and the technical interest rate is selected in order to attain a 

neutral risk valuation of the liabilities. 

235


� WB-15 


2.2.12 

Vehicle Routing Applications III 



Chair: Joaquín Pacheco, Applied Economy, University of Burgos, 

Plaza Infanta Elena s/n, 09001, Burgos, Spain, jpacheco@ubu.es 

1 - Solving a Stochastic Bi-Objective Covering Tour Problem 

Fabien Tricoire, Department of Business Administration, 

University of Vienna, Chair for Production and Operations 

Management, Brünner Straße 72, 1210, Vienna, 

fabien.tricoire@univie.ac.at, Walter Gutjahr 

In a disaster relief mission, the population of the affected region is to be provided 

with emergency supplies. These supplies are delivered to distribution 

centers (DCs), where people can pick them up. The percentage of inhabitants 

of a village going to the next DC depends on the distance between the village 

and the DC. We therefore have to solve an extended bi-objective covering tour 

problem. The decisions concern both the DCs to open and the tours to deliver 

the goods to them. Pareto fronts for the two objectives "cost" and "coverage" 

are determined under sampled stochastic demand. 

2 - Vehicle routing: A case study for laundry services 

Azmin Azliza Aziz, Warwick Business School, University of 

Warwick, Warwick Business School, The University of Warwick, 

CV4 7AL, Coventry, United Kingdom, 

A.A.Aziz@warwick.ac.uk 

This case study focuses on laundry service in Coventry which can be modelled 

as a vehicle routing problem. At present, the daily routing activities are done 

manually. Therefore improvement of current system is necessary by performing 

rescheduling exercise using different approaches whilst satisfying related 

constraints. Comparison amongst approaches is conducted to investigate potential 

cost savings to the problem. Experimental results reveal that the optimization 

software significantly improve the manual implementation. The next 

interest would be to perform robustness checks on the solutions. 

3 - Optimizing vehicle routes in a bakery company 

Joaquín Pacheco, Applied Economy, University of Burgos, Plaza 

Infanta Elena s/n, 09001, Burgos, Spain, jpacheco@ubu.es, Ada 

Alvarez, Irma García, Francisco R. Angel-Bello 

The work is motivated from a real problem of a bakery company in Northern 

Spain. The objective is to minimize the total distance traveled for the daily 

routes over the week. In order to reduce this total distance, some flexibility in 

the dates of delivery is introduced. A two-phase method based in GRASP and 

Path Relinking metaheuristics is proposed. Computational experiments show 

that the method performs very well, obtaining high quality solutions in short 

computational times. When it is applied to real-data-based instances the obtained 

solutions considerably reduce transportation costs 

� WB-16 


2.2.14 

OR models in Public Transport 

Stream: Public Transport [c] 


Chair: Leo Kroon, Rotterdam School of Management, Erasmus 

University Rotterdam, P.O. Box 1738, 3000 DR, Rotterdam, 

Netherlands, lkroon@rsm.nl 

1 - Route of New High-speed Line Which Minimizes Total 

Travel Time of Railway Users 

236 

Takamori Ukai, Nanzan university, 27, Seirei-cho, 4890863, 

Seto, Aichi, Japan, ukai@nanzan-u.ac.jp 

In this presentation, we discuss the route of new high-speed railway which minimizes 

total travel time of users. High-speed mass transit, like bullet train, decreases 

travel time drastically, and it affects the economy of the regions where 

the stations are located. By this reason, many alternatives are suggested to 

select the locations of stations. We formulate the problem as a mathematical 

programming problem and propose the route minimizing the total travel time 

under the assumption that individuals take the route of which their travel time 

is shortest. 

2 - Conflict Avoidance for Air Traffic Flow Management 

Problem, pure and MIP models 

Pablo Olaso, Department of Statistics and Operations Research, 

Universidad Rey Juan Carlos, c Tulipán s/n, 28933, Móstoles, 

pablo.olaso@urjc.es, Celeste Pizarro Romero, Laureano 

Fernando Escudero 

Two linear optimization models for conflict avoidance between any number of 

aircrafts in the airspace are proposed, the first one a pure 0—1 LP which avoid 

conflicts by means of altitude changes, and the second one a Mixed 0—1 LP 

whose strategy is based on velocity and altitude changes. Several objective 

functions are established and some extensions to improve both models are proposed. 

Due to the small computational time for solving both problems, this 

approach can be used in real time by using standard optimization software. 

3 - An Ant Colony Optimisation algorithm for simulating 

hyper-path choices on real-scale networks 

Luca D’Acierno, Dipartimento di Ingegneria dei Trasporti, 

Università degli Studi di Napoli "Federico II", Via Claudio, 21, 

80125, Napoli, Italy, dacierno@unina.it, Mariano Gallo, Bruno 

Montella 

In this paper, we propose an ACO-based algorithm to imitate the behaviour 

of public transport users. In particular, we show that the proposed algorithm 

allows transit systems to be simulated in less time but with the same accuracy 

compared with traditional assignment algorithms. Moreover, we state 

the equivalence in terms of hyperpath choice behaviour between artificial ants 

(simulated with the proposed algorithm) and transit users (simulated with traditional 

assignment algorithms). Finally, we apply the proposed algorithm on 

a real-scale network highlighting performances of the ACO approach. 

4 - A Column Generation Approach to the Vehicle Positioning 

Problem 

Carlos Cardonha, Optimization, Zuse Institute Berlin, 

Takustrasse 7, Berlin-Dahlem, D-14195, Berlin, Berlin, 

Germany, carlos.cardonha@gmail.com, Ralf Borndörfer 

The Vehicle Positioning Problem consists of assigning the vehicles of a public 

transport company to parking positions in a depot. We propose a set partitioning 

model approach for the basic and for the integrated versions of the problem 

and an associated column generation solution approach for them. The proposed 

models provide tight linear descriptions of the problems and the pricing 

routines for some versions can be solved in pseudo-polynomial time. The computational 

results of these methods show that they provide satisfactory solutions 

to large-scale instances of the problem." 

� WB-17 


1.3.14 

Financial Mathematics and Stochastic 

Modelling 

Stream: Computational Statistics 


Chair: Efsun Kürüm, Financial Mathematics, Institute of Applied 

Mathematics, METU, Institute of Applied Mathematics, 06531, 

Ankara, Turkey, efsun.kurum@gmail.com 

1 - Economical applications of Almost Stochastic Dominance 

Elena Almaraz Luengo, Estadística e Investigación Operativa, 

Facultad de Ciencias Matemásticas (UCM), c/Arroyo de la 

Media Legua 68, 2 o B, 28030, Madrid, Spain, 

ealmarazluengo@mat.ucm.es

In this work we study some classical methods to compare options. First, 

Stochastic Dominance (SD) and Mean Variance (MV) rules with their principal 

applications are commented. In the situations in which we cant rank the 

options using these rules, we need other criteria, in particular Almost Stochastic 

Dominance (ASD) rules. We also study examples in which ASD rules let us 

to make a choice between different options. 

2 - Optimal control of a finite-capacity stochastic inventory 

system with setup cost and lost sales 

Xiuli Chao, IOE, University of Michigan, 1205 Beal Ave, 48109, 

Ann Arbor, MI, United States, xchao@umich.edu, Yifan Xu 

One of the most fundamental results in inventory theory is the optimality of 

(s, S) policy for inventory systems with setup cost. This result is established 

based on a key assumption of infinite production/ordering capacity. Several 

studies have shown that, when there is a finite production/ordering capacity, 

the optimal policy for the inventory system is very complicated. We consider 

a continuous review inventory system with finite production/ordering capacity 

and setup cost, and show that the optimal control policy for this system has a 

very simple structure. 

3 - Modelling Predecisional Bias in a Speculative Financial 

Market 

David McDonald, School of Management, University of 


United Kingdom, d.mcdonald@soton.ac.uk, Ming-Chien Sung, 

Johnnie Johnson 

Predecisional bias, where individuals evaluate attributes to favour the leading 

option, has been observed in laboratory studies. We investigate this phenomenon 

in a real world environment, using logistic regression to model decisions 

made in a horserace betting market. We find that bettors do alter their 

probability judgements and prefer a leading option when it is sufficiently more 

favoured than the other choices available. In addition, we find that the bias is 

unaffected by alternative-based complexity but is stronger in situations involving 

high attribute-based complexity. 

4 - Testing for breaks in stochastic volatility model 

Carlos Rivero, Statistics and Operations Research, University 

Complutense of Madrid, Campus de Somosaguas, 28223, 

Madrid, Spain, crivero@estad.ucm.es 

Volatility plays an important role in economics and finance. Deciding whether 

volatility models have constant coefficients or exhibit breaks is an important 

issue in applied works. The statistical properties of the volatility models introduce 

some complexity to test for breaks. This paper proposes a method to test 

for constant parameters in Stochastic Volatility Models. The method can be applied 

to any SVM with autoregressive structure in the conditional variance. The 

proposed test adds a useful tool to financial econometrics and applied finance. 

Simulations confirm the good performance of the method. 

� WB-18 


1.3.15 

Stochastic Models and Queueing Systems 











1 - Equilibrium and socially optimal strategies in the M/G/1 

queue with vacations 

Antonio Gomez-Corral, Department of Statistics and OR, 

Complutense University of Madrid, Faculty of Mathematics, 

Plaza de Ciencias, 3, 28040, Madrid, Spain, 

antonio_gomez@mat.ucm.es 


We study the balking behavior of customers in the single-server queue with vacations. 

Arriving customers decide whether to enter the system or balk, based 

on a natural reward-cost structure that incorporates their desire for service, as 

well as their unwillingness to wait. We identify equilibrium strategies and socially 

optimal strategies under two information assumptions. In a first case, 

customers make individual decisions without knowing the system state. In a 

second case, they are informed about the server’s current status. 

2 - Nonparametric Estimation for a Numerical Evaluation 

of the Proximity of G/G/1 and G/M/1 Systems 

Aïcha Bareche, Laboratory of Modelisation and Optimization of 

Systems, University of Bejaia, Algeria, 06000, Bejaia, Algeria, 

aicha_bareche@yahoo.fr, Djamil Aïssani 

We study strong stability of the G/M/1 queueing system after perturbation of 

the service times. We are interested in the determination of the proximity error 

between the corresponding service time distributions of the G/G/1 and G/M/1 

systems and the approximation error on their stationary distributions in the stationary 

state, when the general distribution of service times G in the G/G/1 

system is unknown and must be estimated by the means of a nonparametric 

estimation method. The boundary effects are taken into consideration. Simulation 

studies are realized to support the results. 

3 - An M/G/1 retrial queue with exhaustive service and multiple 

vacation policy 

Mohamed Boualem, RO, University of Bejaia, LAMOS, 

University of Bejaia, 06000 Bejaia, Algeria, Bejaia, Algeria, 

robertt15dz@yahoo.fr, Djamil Aissani, Djellab Natalia 

In this paper, we consider an M/G/1 retrial queue with server vacations, when 

service times and vacation times are arbitrary distributed. The distribution of 

the number of customers in the system in stationary regime is obtained in terms 

of generating function. Next, we show that the system size can be decomposed 

into two random variables, one of which corresponds to the system size of the 

ordinary M/G/1 FIFO queue without vacation. Such a stochastic decomposition 

property is useful for the computation of performance measures of interest. 

4 - Stochastic inequalities for MX/G/1 retrial queue with impatient 

customers with high retrial rate 

Nawel Arrar, Mathematics, Annaba university, BP.12, 23000, 

Annaba, Algeria, nawel.arrar@univ-annaba.org 

We carry out a stochastic analysis of the M/G/1 retrial queue with batch arrivals 

and impatient customers. This includes steady state joint distribution of the 

server state and the number of customers in retrial group, embedded Markov 

chain and stochastic decomposition for the number of customers. Under high 

retrial intensity, the study state distribution of a retrial queue converges to a 

limit. In our case, it is intuitive that it is the classical M/G/1 queue with batch 

arrivals and impatience phenomenon. We prove this heuristic argument with 

stochastic decomposition property. 

� WB-19 



Nonsmooth Optimization and Its 

Applications 

Stream: Nonsmooth Optimization 


Chair: Antonio Fuduli, Department of Mathematics, Universita’ della 

Calabria, Via P. Bucci, CUBO 31B, 87036, Rende, Italy, 

antonio.fuduli@unical.it 

Chair: Annabella Astorino, ICAR, CNR, C/0 DEIS - UNICAL, 

CUBO 41 C, 87036, RENDE, Italy, astorino@icar.cnr.it 

1 - Generalized Bundle Methods for Decomposable Functions 

with "Easy" Components 

Enrico Gorgone, DEIS - Dipartimento di Elettronica Informatica 

e Sistemistica, Università della Calabria, Via Bucci, CUBO 41c, 

VI piano, 87036, Rende, Cosenza, Italy, 

egorgone@deis.unical.it, Antonio Frangioni 

237

WB-20 EURO 24 - Lisbon 2010 

Many large-scale optimization problems exhibit a block-structure that can be 

algorithmically exploited by means of decomposition approaches. We propose 

a modification to the (generalized) bundle scheme for minimization of a decomposable 

nonsmooth convex function, where some components are easy convex 

programs. We show how to construct, for this case, a suitably modified representation 

of the original convex subproblem, providing it with "exact" information 

about "easy" components of the function to be minimized. We report 

numerical results for Network Design problem. 

2 - Analysis of optimization algorithms for clustering 

Zorana Luzanin, Department of Mathematics and Inforormatics, 


Sad, Serbia, zorana@dmi.uns.ac.rs 

Two main issues in cluster analysis are considered in this paper. The first one 

is the weighting of attributes while the second is application of nonsmooth optimization. 

The effectiveness of algorithms is demonstrated on real data sets. 

Cluster analysis was used on the results of the poll organized by the Ministry 

of Health which pertains to the health of the population in Serbia. The basic 

features included in this cluster analysis are socio-economic status, hygienic 

habits, diet, leisure, knowledge of health topics, subjective life satisfaction, and 

mental health. 

3 - Heuristic algorithm for clustering large data sets based 

on nonsmooth optimization approach 

Adil Bagirov, School of Information Technology & 

Mathematical Sciences, University of Ballarat, University Drive, 

Mount Helen, P.O. Box 663, 3353, Ballarat, Victoria, Australia, 

a.bagirov@ballarat.edu.au 

The k-means algorithm is known to be fast clustering algorithm. However, 

it is sensitive to the choice of starting points and is inefficient for clustering 

large datasets. Recently, incremental approaches have been developed to resolve 

difficulties with the choice of starting points. The global and modified 

global k-means algorithms are based on such an approach. However, they are 

not suitable for clustering very large data sets. We propose a new version of the 

modified global k-means algorithm which is based on nonsmooth optimization 

approach and is suitable for clustering large datasets. 

4 - Nonsmooth Convex Optimization via Piecewise 

Quadratic Approximations 

Annabella Astorino, ICAR, CNR, C/0 DEIS - UNICAL, CUBO 

41 C, 87036, RENDE, Italy, astorino@icar.cnr.it, Antonio 

Frangioni, Manlio Gaudioso, Enrico Gorgone 

We present a numerical method, of bundle type, for minimization of a real 

convex function of several variables, not necessarily differentiable. 

Differently from standard methods, based on the cutting plane approach, in our 

method we approximate the objective function via a piecewise quadratic model 

of max type. It retains the property of interpolating the original objective function 

at the "bundle’ points. 

Termination at an approximate optimal solution is proved and numerical results 

are reported. 

� WB-20 


1.3.33A 

Data Mining and Decision Making II 

Stream: Data Mining and Decision Making 


Chair: Hsin-Vonn Seow, Business School, University of Nottingham- 

Malaysia Campus, Jalan Broga, Selangor Darul Ehsan, 43500, 

Semenyih, Selangor, Malaysia, 

Hsin-Vonn.Seow@nottingham.edu.my 

Chair: Lai-Soon Lee, UPM Serdang, 43400, Selangor Darul Ehsan, 

Malaysia, lslee@math.upm.edu.my 

1 - Rule Based Predictive Models, Decision Table and Tree: 

An Empirical Evaluation on Comprehensibility 

238 

Karel Dejaeger, Faculty of Business and economics, Katholieke 

Universiteit Leuven, Naamsestraat 69, 3000, Leuven, Belgium, 

Karel.dejaeger@econ.kuleuven.be, Wouter Verbeke, Johan 

Huysmans, Christophe Mues, Jan Vanthienen, Bart Baesens 

Little research has been performed to assess the comprehensibility to the enduser 

of predictive data mining models. In this paper, an empirical study is 

presented which investigates the comprehensibility of a number of alternative 

representation formats for classification. An end-user experiment is designed 

to test the accuracy, response time and answer confidence for a set of problemsolving 

tasks involving different representation formats. The formats under 

consideration are decision tables, decision trees, propositional and oblique 

rules. 

� WB-21 


6.2.47 

Optimization Algorithms II 



Chair: Tamas Kis, Computer and Automation Research Institute, 

Hungarian Academy of Sciences, P.O.Box 63, 1518, Budapest, 

Hungary, kistamas@sztaki.hu 

1 - Strengthening split cuts by lift-and-project 

Tamas Kis, Computer and Autmoation Research Institute, Kende 

str. 13-17, 1111, Budapest, Hungary, tamas.kis@sztaki.hu, Egon 

Balas 

We propose a procedure for improving arbitrary split cuts by pivoting in the LP 

tableau based on the lift-and-project reduced costs. We show how this procedure, 

successfully used in the last few years to improve mixed integer Gomory 

cuts, can be applied to other split cuts by creating a new source row specific to 

the family of cuts in question. The starting point of the new procedure is any 

split disjunction which can be obtained by combining integer variables using 

integer multipliers. Computational experience will be discussed. 

2 - A hybrid algorithm for global optimization problems 

Leticia Velazquez, Mathematical Sciences, The University of 

Texas at El Paso, 500 West University Avenue, 79968-0514, El 

Paso, Texas, United States, leti@utep.edu, Miguel Argaez, 

MIguel Hernandez, Carlos Ramirez, Reinaldo Sanchez 

We are developing a hybrid algorithm for solving global optimization problems 

that is based on the coupling of a stochastic global method (Simultaneous 

Perturbation Stochastic Approximation, Simulated Annealing, Genetic 

Algorithms) and a local method (Newton-Krylov Interior-Point) via a surrogate 

model. There exist verified algorithms for finding approximate global 

solutions, but our technique will further guarantee that such solutions satisfy 

physical bounds of the problem. First, the SPSA algorithm conjectures regions 

where a global solution may exist. Next, some data points from the regions 

are selected to generate a continuously differentiable surrogate model that approximates 

the original function. Finally, the interior-point Newton algorithm 

is applied to the surrogate model subject to bound constraints for obtaining a 

feasible approximate global solution. We present some encouraging numerical 

results of small to large scale parameter estimation problems. The authors acknowledge 

the support from the Department of the Army Grant No. W911NF- 

07-02-0027. 

3 - Some experimental results measuring the complexity 

of the B&B algorithm versus the geometry of the polyhedron 

Ivan Derpich, Industrial Engineering Department, University of 

Santiago of Chile, Av. Ecuador 3769, Estacion Central, Volcan 

Lanin 205, las Condes, 0000, Santiago, Region metropolitana, 

Chile, ivan.derpich@usach.cl 

In this presentation we show first a new bound for the width of the polyhedron 

based in the eigenvalue of the polyhedron. Then we use this bound for estimate 

the maximum number of nodes to branch in the B&B algorithm. Finally we 

compare this maximum number of nodes with the number of nodes obtained in 

a set of instances taken from a set of test problem.

� WB-22 



Maritime transportation 

Stream: Maritime Logistics 


Chair: Marielle Christiansen, Norwegian University of Science and 

Technology, N-7291, Trondheim, Norway, 

marielle.christiansen@iot.ntnu.no 

1 - Integrated cargo routing and ship scheduling in liner 

shipping 

Karina Kjeldsen, CORAL, Department of Business Studies, 

Aarhus School of Business, Fuglesangs Alle 4, 8210, Aarhus V, 

Denmark, kahk@asb.dk, Oguz Solyali 

The problem consists of creating routes and schedules for a heterogeneous fleet 

of ships while determining the cargo routing and the speed for all relevant port 

pair/ship combinations. Transshipment is allowed in ports with transshipment 

capabilities. The service frequency is fixed at one week. Since the speed of 

the ships is a decision variable, the developed model is nonlinear. The model 

is made linear by means of variable redefinition. Using decomposition the linearized 

model is split into a master problem and a sub problem per ship, and 

solved by a column generation algorithm. 

2 - A Branch-and-Price-and-Cut Method for a Maritime 

Pickup and Delivery Problem with Time Windows and 

Split Loads 

Magnus Stålhane, Industrial Economics and Technology 

Management, NTNU, Alfred Getz veg 3, 7491, Trondheim, 

Norway, stalham@iot.ntnu.no, Henrik Andersson, Marielle 

Christiansen, Jean-François Cordeau, Guy Desaulniers 

We present a branch-and-price-and-cut method for a maritime pickup and delivery 

problem with time windows and split loads. The fleet is heterogeneous 

with each ship having a different load capacity, speed and cost structure. There 

are both mandatory and optional cargoes to transport. Each cargo has a time 

window both at the pickup and the delivery ports, and earns revenue for transporting 

it. Each cargo may be transported by one or more ships, potentially 

giving substantial savings compared to solutions where splitting of cargoes is 

not allowed. Computational results will be presented. 

3 - A medium-term short sea fuel oil distribution problem 

Alexandrino Duarte Delgado, Matemática, Universidade de 

Aveiro, Campus Campus Universitário de Santiago, 3810 - 193, 

Aveiro, Portugal, alexandrino.delgado@unicv.edu.cv, Agostinho 

Agra, Marielle Christiansen 

We consider a medium-term short-sea fuel oil distribution problem occurring 

in the archipelago at Cape Verde. Here, an oil company is responsible for the 

inventory management and for the routing of ships between the islands such 

that the demand for various products is satisfied during the planning horizon of 

several weeks. Inventory capacities are considered and a constant demand rate 

at each port is assumed. 

We present a mathematical model of the problem and discuss different strategies 

to improve the proposed model, such as the use of an extended formulation 

and the inclusion of valid inequalities. Computational results based on real data 

will be reported. 

4 - Recent research and trends in maritime cargo and inventory 

routing 

Marielle Christiansen, Department of Industrial Economics and 


Technology, Alfred Getz vei 3, N-7491, Trondheim, Norway, 


We present the current status of models and solution methods for tactical planning 

problems in tramp and industrial shipping. Problems including both ship 

routing and inventory management considerations will be discussed, as well as 

routing and scheduling problems where specified cargoes are transported from 

given loading ports to discharge ports. Maritime transportation problems are 

rich and we discuss several important real life aspects. A brief overview of 

both exact methods and heuristics used for solving maritime cargo and inventory 

routing problems will be given. Finally, we present some trends regarding 

the research within maritime cargo and inventory routing. 


� WB-23 


6.2.49 

Price and Risk Forecasting in the Financial 

Sector 



Chair: Marcus Hildmann, Swissquant Group AG, Kuttelgasse 7, 

8001, Zürich, Switzerland, hildmann@swissquant.ch 




1 - Conditional Bootstrapping and Price Forecasting 

Marcus Hildmann, Swissquant Group AG, Kuttelgasse 7, 8001, 

Zürich, Switzerland, hildmann@swissquant.ch, Dejan Stokic, 

Florian Herzog 

We propose a bootstrapping concept of simulating the future prices, without a 

need for a priori setting specific data distribution, since assuming a data distribution 

in order to test the empirical data is potential source of errors. However, 

using simple bootstrapping destroys the path information of a time series. We 

present a bootstrapping based data mining method to simulate future electrical 

hourly price paths, which have the empirically observed path properties, calibrated 

by a nonlinear underlying process. Resulting simulated time series also 

include external seasonality information. 

2 - Neural network based search for mispriced options 

Dejan Stokic, Swissquant Group, Kuttelgasse 7, 8001, Zürich, 

Switzerland, stokic@swissquant.ch, Marcus Hildmann 

There are two general ways of speculating on financial options: by forecasting 

the future implied volatility or by searching for options that are heavily mispriced. 

We address the latter approach by developing a neural-network based 

data mining algorithm, which given the option price, price of the underlying, 

strike and moneyness, searches for options which are under- or over-priced. 

We train our model on the data collected in last 3 years. By using this nonparametric 

approach, none of the assumptions about the dynamics of underlying 

processes have to be specified. 

3 - Nonparametric method for economic indices construction. 

Application to analysis of capital and financial 

markets 

Ivan Kondrakov, Moscow Institute of Physics and Technology, 9, 

Institutskii per., 141700, Dolgoprudny, Moscow Region, Russian 

Federation, ivankondrakov@mail.ru, Alexander Shananin 

A nonparametric method for construction of the Konus-Divisia indices is considered. 

The method is based on the Afriat-Varian theorem and takes into account 

changes in the demand structure when the structure of prices is changing. 

The construction of indices is reduced to solving a linear inequalities system 

of a special type, for which a solution method of polynomial computational 

complexity is known. This nonparametric method can be used for analysis of 

market segmentation and prediction of demand. We discuss also the possibility 

of its using for analysis of capital and financial markets. 

4 - Estimation of investment project profitability in the 

modified Cantor-Lipman model 

Mikhail Vashchenko, Computing Center of the Russian 

Academy of Sciences, Vavilov St. 40, 119333, Moscow, Russian 

Federation, m_vashchenko@mail.ru 

The paper covers the methods for assessing the yield of investment projects, 

namely, considers the modified Cantor-Lipman model, which takes into account 

the probability of a crisis in the investments market and its impact on 

investor behavior. In such a formulation, the problem is reduced to the Bellman 

equation. We investigate the case when a cautious investment strategy avoiding 

bankruptcy is the optimum strategy, and concentrate on estimation of the 

investor’s capital growth in the dynamic system based on a cautious strategy. 

239


� WB-24 


6.2.50 

Automated Nurse Rostering 



Chair: Greet Vanden Berghe, Industrieel Ingenieur, KaHo 

Sint-Lieven, Gebr. Desmetstraat 1, 9000, Gent, Belgium, 

greet.vandenberghe@kahosl.be 

1 - Design and Implementation of Operational Research 

Techniques for Making Schedules of Nurses 

Lakhdar Djeffal, Universit Hadj-lakhdar, 05000, Batna, Algeria, 

lakdar_djeffal@yahoo.fr, Elamir Djeffal, Gilles Goncalves 

Scheduling the nursing staff is based on finding solutions to combinatorial nature, 

responding to multiple constraints. In this paper we describe an advanced 

method for finding a solution very close to the mathematical model of type 

Constraint Satisfaction Problem CSP with objective function. This method 

combines techniques from operations research with heuristics which allows 

it to be done at each node of the search tree. Our results show that the tool 

developed is effective in terms memory usage, decision support based on the 

preferences and wishes of assignment. 

2 - On Solving Real Nurse Rostering Problems 

Margarida Pato, ISEG (Technical University of Lisbon), CIO 

(University of Lisbon) and ISEG (Technical University of 

Lisbon), Rua do Quelhas, 6, 1200-781, Lisboa, Portugal, 

mpato@iseg.utl.pt, Margarida Moz, Pedro Caldeira 

We present a recently improved computational decision support system developed 

to determine rosters for nurses working in hospital units. The system 

includes input and output modules as well as an engine. Both input and output 

respect all the practical rostering issues of a real public hospital context and 

the engine is a standard exact algorithm used to solve the mixed binary linear 

programming model for the specific purpose. Some applications of the system 

with real data will be given. 

3 - Robust Optimization Approach to Scheduling Interns at 

Hospitals 

Maryam Ghotbaddini, Apt 4, No. 15, Tous Alley, Valiasr St., 

Tajrish Sq., Tehran, Tehran, Iran, Tehran, Tehran, Iran, Islamic 

Republic Of, ghotbaddini@gmail.com, Mohammad Javad 

Feyzollahi, Mohammad Modarres 

Scheduling interns at hospitals is an important problem which affects the utilization 

of interns, satisfactory of patients and hospital costs. In most of developed 

models for this problem, researchers neglected data uncertainty or their 

efforts resulted to complicated models, which are difficult to compute. It is 

obvious that in a real-world problem the exact number of daily patients is not 

known precisely. In this paper, we use a robust optimization method to address 

data uncertainty in scheduling interns at hospitals. 

4 - Constraint and precondition definitions for nurse rostering 

problems 

Greet Vanden Berghe, Industrieel Ingenieur, KaHo Sint-Lieven, 

Gebr. Desmetstraat 1, 9000, Gent, Belgium, 

greet.vandenberghe@kahosl.be, Burak Bilgin 

Many different variants of the nurse rostering problem exist. The difference 

mainly constitutes personnel and work characteristics, the objectives, and the 

hard and soft constraints. 

We concentrate on two particular modelling issues that require attention in order 

to make nurse rostering approaches re-usable in different settings. The 

first one considers precondition definitions. It deals with constraint evaluation 

across planning periods. Secondly, we present a model for dealing differently 

with idle days and absence requests. It offers an extension to existing constraint 

definitions. 

240 

� WB-25 


6.2.48 

ROADEF/EURO challenge senior session 1 



Chair: H. Murat Afsar, Industrial Systems, University of Technology 

of Troyes, 12 rue Marie Curie, BP 2060, 10010, Troyes, France, 

murat.afsar@utt.fr 

1 - A Guided-Construction approach to large scale power 

plant scheduling 

Cor Hurkens, Mathematics and Computer Science, Eindhoven 

University of Technology, POBox 513, 5600MB, Eindhoven, 

Netherlands, wscor@win.tue.nl 

We tackle the power plant scheduling problem of the 2010 ROADEF Challenge 

by a mixture of OR techniques: - tightening the outage opportunities; - optimizing 

a single plant schedule by means of dynamic programming; - enforcing 

feasibility by multi-plant scheduling via a time-index MIP formulation; - balancing 

production cost per scenario by a flow formulation. Solution variants 

are found by random sequencing of plants, as well as by varying the impact of 

modulation. We use CPLEX for solving the various subproblems. 

2 - A solution approach to the ROADEF/EURO challenge 

based on Benders decomposition 

Laurent Flindt Muller, DTU Management Engineering, 

Technical University of Denmark, Produktionstorvet, DTU — 

Bygning 424, 2800, Kgs. Lyngby, Denmark, lafm@man.dtu.dk, 

Richard Lusby, Bjørn Petersen 

A stochastic large-scale energy management problem is considered, which consists 

of planning refuelling and production levels for a set of power plants, such 

that demands are met across a set of scenarios at minimum cost. The algorithm 

is based on Benders Decomposition (BD), where the master problem entails 

finding shutdown dates and reload amounts, while each subproblem involves 

finding an optimal production plan for a scenario. Because of the nature of 

the constraints, the normal BD-algorithm is adapted to handle these. Effective 

preprocessing removes large parts of the solution space. 

3 - Matheuristic methods for large-scale energy management 

problems 

Mauro Dell’Amico, ICOOR, Italy, dellamico@icoor.it, José 

Carlos Díaz Díaz 

The algorithm we propose has an heuristic nature, but it makes use of sophisticated 

mathematical models and applies exact optimal solution approaches. Our 

method works in phases. - In the first phase, preprocessing techniques are used 

to strengthen and propagate the constraints. - In the second phase, a constructive 

heuristic schedules the outages and finds an admissible production asset. 

- The third phase is devoted to improve the incumbent solution through local 

search methods. - In the fourth phase, a MIP model which catches some of the 

most important constraints of the problem is solved through an iterative procedure 

which adds "dual cutting planes". - The last phase is a refinement of the 

best solution found. 

4 - Team S21: Using Integer Linear Model for Scheduling 

Outages 

Vincent Jost, Laboratoire d’Informatique, CNRS - Ecole 

Polytechnique, route de Saclay, 91128, Palaiseau, France, 

vjost@lix.polytechnique.fr, David Savourey, Nora Touati 

Moungla 

We split the problem into scheduling the outages and power affectation (for a 

given schedule). Currently, we use a 0-1 linear formulation for the scheduling 

of the outages taking into account CT 13 to 21. We heuristically add other 

constraints to ensure that enough fuel can be spent between two consecutive 

outages (constraint on min refueling and Amax). Starting with maximum production 

for type 2 power plants, we deal with modulation to avoid overproduction. 

We still need to find ways to evaluate the quality of a schedule. We 

discuss the relevance of the model, especially the stochastic part.

� WB-26 


3.1.11 

New Achievements in Game Theory I 

(cooperative and noncooperative) 



Chair: Mariana Rodica Branzei, Faculty of Computer Science, 

"Alexandru Ioan Cuza” University, 16, Berthelot St., 700483, Iasi, 

Romania, branzeir@info.uaic.ro 

Chair: Sirma Zeynep Alparslan Gok, Mathematics, Faculty of Arts 

and Sciences, Suleyman Demirel University, Faculty of Arts and 

Sciences, Suleyman Demirel University, 322260, Isparta, Turkey, 

zeynepalparslan@yahoo.com 




1 - A smuggling game with incomplete information 

Ryusuke Hohzaki, Dep. of Computer Science, National Defense 

Academy, 1-10-20 Hashirimizu, 239-8686, Yokosuka, 

Kanagawa, Japan, hozaki@cc.nda.ac.jp, Ryuichi Masuda 

Since the Dresher’s work, there have been many researches on the so-called 

inspection game or smuggling game. This report deals with a smuggling game 

of Customs and a smuggler by considering asymmetric information which the 

players obtain. We model the game as a game with incomplete information, 

which has not been studied so far. We derive a perfect Bayesian equilibrium 

for the game and analyze optimal decisions of the players by comparing with 

another model with complete information. 

2 - New results on rendezvous search on the interval 

John Howard, Operational Research, London School of 

Economics, Houghton Street„ London, WC2A 2AE, United 

Kingdom, j.v.howard@lse.ac.uk, Marco Timmer 

Two people are placed on a finite stretch of road (using independent draws 

from the same distribution), and must try to meet each other in the least possible 

expected time. This is a rendezvous search problem for which some optimal 

solutions are known when the players have to use identical strategies and the 

distribution is uniform. In this paper we characterise the complete set of solutions 

for the uniform case. 

Further, we derive a recurrence relation for solutions to this symmetric rendezvous 

problem for any initial distribution, so that all such problems can be 

solved by computer. 

3 - A Modified Beer Game with Two Decision-Makers at 

Each Stage 

Özlem Çoban, Industrial Engineering, Sabanci University, 


ozlemcoban@su.sabanciuniv.edu, Murat Kaya, Gurdal Ertek 

We conduct a modified version of the well-known beer game in which, at each 

stage, the order decision is jointly given by two subjects: the sales manager 

whose performance measure is the back order cost, and the production manager 

whose performance measure is the inventory holding cost. We analyze if 

this change in the decision structure will dampen the bullwhip effect. We also 

run regression analysis to determine the factors (such as the inventory on hand 

and pipeline inventory) that affect the subjects’ order quantity decisions. 

� WB-27 


8.2.06 

URBAN TRANSPORT SYSTEMS 



Chair: Maurizio Bruglieri, INDACO, Politecnico di Milano, Via 

Durando, 38/a, Milano, Italy, maurizio.bruglieri@polimi.it 


1 - Data Provision For Attended Home Delivery In Urban 

Areas 

Dirk Christian Mattfeld, Business Information Systems, 

Universitaet Braunschweig, Decision Support Group, 

Mühlenpfordtstraße 23, 38106, Braunschweig, Germany, 

d.mattfeld@tu-bs.de, Jan Fabian Ehmke 

Fast and reliable delivery of goods is a crucial part of service quality. In order 

to provide a high service quality in attended home delivery, customers expect a 

choice of narrow delivery time slots. Confirmed time slots have to be realized 

within cost efficient delivery tours. We demonstrate efficient and reliable vehicle 

routing in urban areas, which is based on empirical traffic data that can be 

utilized in time-dependent problem formulations. Computational experiments 

underline the benefits of time-dependent heuristics for the travelling salesman 

problem considering customer time windows. 

2 - Centralized Versus Decentralized Control - A Solvable 

Stylized Model in Transportation Logistics 

Olivier Gallay, STI-IMT-LPM, Ecole Polytechnique Fédérale de 

Lausanne (EPFL), Station 17, CH-1015, Lausanne, Vaud, 

Switzerland, olivier.gallay@epfl.ch, Max-Olivier Hongler, 

Michael Hülsmann, Philip Cordes, Richard Colmorn 

To analyze the potential outcomes resulting from interaction between autonomous 

decision-making “smart parts” in logistics networks, we propose 

here an exactly solvable stylized model that is able to quantify how much the 

dynamics can be enhanced by (fully decentralized) agent-based mechanisms. 

Cost functions are introduced in order to compare the performances of centralized 

versus decentralized organization and we are enable to conclude that 

for time horizons shorter than a critical value, multi-agent interactions generate 

smaller costs that an optimal effective central controller. 

3 - Improved probe vehicle measurements with point data 

Livia Mannini, Dep. of Civil Engineering, Roma Tre University, 

Italy, lmannini@uniroma3.it, Ernesto Cipriani, Stefano Gori 

This study deals with the problem of urban travel time forecast. In order to 

achieve this objective a procedure that combines micro and macro approaches 

is proposed: micro approach deals with microsimulation models feeded with 

micro (GPS probe vehicle) data; then, a macro approach is adopted to integrate 

micro data with aggregate measurements supported by fixed detectors. 

4 - A web-based carpooling service for universities: a case 

study in Milan 

Maurizio Bruglieri, INDACO, Politecnico di Milano, Via 

Durando, 38/a, Milano, Italy, maurizio.bruglieri@polimi.it, 

Alessandro Luè, Alberto Colorni 

Carpooling is a transport system based on a shared use of private cars. The 

mobility managers of the Università Statale and Politecnico di Milano universities 

are interested in promoting the use of such system among their students 

and employees. The paper presents an ongoing project to design, implement 

and test a car pooling service for such universities. The design will take into 

account how to introduce and promote the service, identifying regulation, incentives, 

modalities, and marketing actions. A web-based software tool will be 

implemented to manage the matching of the users. 

� WB-28 



Scheduling with Due Dates 



Chair: Igor Karpov, Laboratory 20, Institute of Control Sciences of 

the Russian Academy of Sciences, Russia, Russia, Moscow, 

Profsoyuznaya, 65, Moscow, Russian Federation, 

karpov_sh33_yar@mail.ru 

1 - Heuristics for the single machine weighted squared tardiness 


Jorge Valente, Faculdade de Economia - LIAAD - INESC Porto 

L.A., Universidade do Porto, Rua Dr. Roberto Frias, 4200-464, 

Porto, Portugal, jvalente@fep.up.pt, Jeffrey Schaller 

241


This paper considers the single machine scheduling problem with a weighted 

squared tardiness objective function. Several dispatching rules are proposed. 

These heuristics are tested on a set of randomly generated problems. The performance 

of the proposed dispatching rules is analysed, and for the smaller 

instances the heuristics are compared with optimum results. 

2 - Evaluation of Job Shop Scheduling Rules in Due Date 

Performance Optimization 

Fabio Pereira, Industrial Engineering Post Graduation Program, 

Nove de Julho University (Uninove), Francisco Matarazzo Av., 

612 São Paulo-Brasil, 05001100, São Paulo, São Paulo, Brazil, 

fabio.pea@gmail.com, Edna Barbosa, Michele Gonçalves, 

Marilda Fátima 

Job shop is small manufacturing operations environments which the production 

scheduling are characterized by, for example, the need to promise a competitive 

completion date estimate to the customer. In these conditions, the choice of the 

proper production schedule is healthy. This paper evaluates the effect of the 

main scheduling rules on due date performance in a job shop for optimization 

of both number of tardy jobs and the total tardiness. The optimization approach 

is based on the OptQuest tool for Arena, which allows searching for optimal 

solutions within simulation models. 

3 - A Preemption-Based Heuristic for the Single-Machine 

Total Weighted Tardiness Problem 

Halil ¸Sen, Industrial Engineering, Sabanci University, Sabanci 

University, Orhanli/Tuzla, 34956, Istanbul, Turkey, 

halilsen@sabanciuniv.edu, Kerem Bulbul 

We consider the non-preemptive single-machine total weighted tardiness 

(TWT) problem with general weights, processing times, and due dates. We 

first develop a family of preemptive lower bounds for this problem and explore 

their structural properties. Then, we show that the solution corresponding to the 

least tight lower bound features some desirable properties that can be exploited 

to build excellent feasible solutions to the original non-preemptive problem in 

short computational times. We present results on standard benchmark instances 

from the literature. 

4 - Polynomially solvable case of the NP -hard problem 

1|r_j|L_max 

Igor Karpov, Laboratory 20, Institute of Control Sciences of the 

Russian Academy of Sciences, Russia, Russia, Moscow, 

Profsoyuznaya, 65, Moscow, Russian Federation, 

karpov_sh33_yar@mail.ru, Alexander Lazarev 

We consider the classical NP -hard scheduling problem in strong sense 1 | 

r_j | L_max. New properties of optimal schedules are found. Polynomially 

case is selected when the release times (r_j), the processing time (p_j) and 

due dates (d_j) of jobs satisfy the relationships: d_j = αp_j + βr_j + C, 

p_j ≥ 0, α ∈ [0, 1], β ∈ [1, +∞), C – constant. An algorithm finds Paretooptimal 

sets of schedules for objective functions L_max and C_max that contains 

no more than n schedules. 

Paragraphs of the report: 

1.Properties of the problem. 

2.Makespan problem under a constraint on the maximum lateness. Algorithm 

and its evaluation. 

3.Pareto-optimal schedules for the objective functions C_max and L_max. 

Algorithm and its evaluation. 

� WB-29 


8.2.11 

Energy and Environmental Finance 



Chair: Silvana Stefani, Quantitative Methods for Economics and 

Business Sciences, Università Milano Bicocca, Piazza Ateneo Nuovo 

1 U7-4023, I-20126, Milano, Italy, silvana.stefani@unimib.it 

1 - Market power and optimal energy production under the 

EU ETS system 

242 

Daniele Felletti, Metodi Quantitativi, Università di 

Milano-Bicocca, via Bicocca degli Arcimboldi 8, Edificio U7, 

24040, Milano, Italy, daniele.felletti@tiscali.it, Silvana Stefani, 

Paolo Falbo 

A risk averse energy producer can switch production between two kinds of 

plants with different CO2 impact. We suppose he can partly transfer price 

shocks in production factors to the final price of electricity. The market power 

of the producer is modelled as a coefficient influencing the multivariate distribution 

of cost factors and energy prices. By simulation we analyse several 

market scenarios. Optimal production policies are obtained as efficient combinations 

of mean-variance of the profit function. 

2 - New Solution Tools to Unit Commitment in Power Production 

Planning 

Joao Pedro Pedroso, DCC - FC, Universidade do Porto and 

INESC Porto, Portugal, jpp@fc.up.pt, Ana Viana, Abdur Rais 

Unit Commitment is a classical problem in power production planning, consisting 

of deciding which electric generators must be committed to production in 

each period, and defining the production level at which each generator should 

operate. We present a branch-and-bound strategy for solving the problem, taking 

advantage of the specific structure of the problem: branch propagation is 

allowed only for children that, after look ahead, are found to satisfy all constraints. 

Careful bounding based on the quadratic structure of the problem is 

applied. 

3 - Emission markets: dynamics of market prices and a 

"fair’ price for CO2 

Silvana Stefani, Quantitative Methods for Economics and 

Business Sciences, Università Milano Bicocca, Piazza Ateneo 

Nuovo 1 U7-4023, I-20126, Milano, Italy, 

silvana.stefani@unimib.it, Daniele Felletti, Paolo Falbo 

A power producer can switch production between a polluting and a renewable 

plant. After a description of CO2 emission market in the first trading period, 

we propose an equation by which a CO2 "fair’ threshold price is found below 

which it is not convenient switching production from polluting to renewable. 

We show evidence for Germany and Italy. 

4 - Does speculation drive oil prices? A new evidence 

Cristina Bencivenga, Teoria Economica e Metodi Quantitativu, 

Universita’ di Roma "La Sapienza", Piazza Aldo Moro 5, 00155, 

Rome, Italy, Italy, c.bencivenga@dte.uniroma1.it, Umberto 

Triulzi, Rita D’Ecclesia 

Crude oil prices have been showing exceptional volatility with oil price moving 

from $ 50 per barrel to $ 147 per barrel (June 2005 - July 2008) to drop to $ 

50/bd on November 21, 2008. Oil price fluctuation really affects consumers, 

producers and marketers especially in terms of costs, incentives to invest in 

technology and trading strategies. At present the price of crude oil does not 

seem to be provided by the traditional relationship between supply and demand 

but it is affected by others factors as a dynamic financial market and changing 

political forces (Stevans L. K., Sessions D. N. (2008)). We assume that active 

financial traders operating in the oil market may cause large deviations of 

prices from fundamentals. In particular some "speculators’ i.e. large banks and 

hedge funds, not directly interested in the delivery of oil, may strongly affect 

this market. The recent rise in oil prices may be generated by both changes 

in market fundamentals and speculation (Kaufmann R. K., Ullman B. (2009)). 

Other economists (i.e., Krugman P. (2008)) sustain that the "oil bubble’ is not 

due to speculation but it may be a result of other variables linked to the growing 

consumption of emerging countries (i.e., China) and the increasing cost of exploration 

and drilling activities. This paper aims to identify the various factors 

that affect the dynamics of West Texas Intermediate (WTI) crude oil spot prices 

and empirically assess their role. The role of the speculative factor is investigated 

together with macroeconomic variable. Given the non-stationarity of the 

examined variables a VECM is adopted. The analysis is performed over the 

period 1993-2009 in order to capture possible changes in the dynamic. Over 

the entire period one cointegrating vector, i.e., a long run equilibrium is found 

while over the sub period 2001-2009 two long run equilibrium are detected. 

The results show that in the period 2001-2009 the variables chosen to measure 

the speculative factor seem to affect the price dynamic.

� WB-30 


8.2.13 

Risk Models in Finance 

Stream: Operational Research and Quantitative Models 

in Banking 


Chair: João Oliveira Soares, Engineering and Management, IST, Av. 

Rovisco Pais, 1049-001, Lisboa, Portugal, joaosoares@ist.utl.pt 

1 - Quantitative vs. Qualitative Criteria for Credit Risk Assessment* 

João Oliveira Soares, Engineering and Management, IST, Av. 

Rovisco Pais, 1049-001, Lisboa, Portugal, joaosoares@ist.utl.pt, 

Joaquim Pina, Margarida Catalão-Lopes 

The existing vast literature on credit risk assessment and default prediction provides 

models building mostly in quantitative indicators. We present the results 

of a survey carried out of experts from the main banks in Portugal, conveying 

evidence on the dominant procedures undertaken by the Portuguese banking 

system. Our analysis concludes on the relevance of qualitative criteria, particularly 

management’s experience and reliability, and on their significant negative 

correlation with banks’ default records. Within this context the paper reflects 

on the role of multi-criteria decision analysis (MCDA) models as a way to process 

credit risk assessment integrating qualitative and quantitative aspects. 

2 - Consolidation in the European Pharmaceutical Industry 

Alain Chevalier, FINANCE, ESCP-EAP, 79,Avenue de la 

République-PARIS-75011-FRANCE, 75011, PARIS, PARIS, 

France, chevalier@escp-eap.net 

The paper analyses questions related to M&A activity in the European pharmaceutical 

sector: (a) consolidation forces in the industry, (b) factors which affect 

the sector’s profitability, (c) potential evolutions in the R&D policies. Three 

M&A cases (Bayer-Schering, Sanofi-Aventis, several Novartis operations) are 

presented. An attempt is made to assess whether value was created by the acquisition 

and to identify the existence of abnormal returns. The results will be 

compared with those of the financial industry. 

3 - Incorporating Risk and Uncertainty into MCDM in practice 

Brendan O’Brien, University College Dublin, Sheheree, Loreto 

Road, 1111, Killarney, Co Kerry, Ireland, 

brendan.obrien@ucd.ie, Cathal Brugha 

Risk and uncertainty is incorporated in to a structured MCDM method using 

Direct-Interactive Structured-Criteria (DISC) Multi-Criteria Decision-Making 

(MCDM). This method specifically combines both financial and non financial 

aspects of a large decision into a single usable methodology that provides an optimal 

result even when there are very close alternatives and scores. We propose 

a method that is robust, usable and uses real criteria structures. The method is 

implemented in a real case, with actual criteria, actual scores, actual decision 

makers in a manner that is interactive and refines as the decision progresses. 

4 - The information content of option-based forecasts of 

volatility 

Silvia Muzzioli, Economics, University of Modena and Reggio 

Emilia, V.le Berengario 51, 41100, Modena, Italy, 

silvia.muzzioli@unimore.it 

The aim of this paper is to investigate and empirically test with market data the 

information content of option based forecasts of volatility. In particular we examine 

the predictive power of three different forecasts: Black-Scholes implied 

volatility, model free implied volatility proposed by Britten-Jones and Neuberger 

(2000), and corridor implied volatility introduced by Carr and Madan 

(1998). Moreover we compare the three option-based forecasts with historical 

volatility in order to see if they subsume all the information contained in the 

latter. 


� WB-31 


8.2.15 


routing II 



Chair: Geir Hasle, Applied Mathematics, SINTEF ICT, P.O. Box 124 

Blindern, NO-0314, Oslo, Norway, Geir.Hasle@sintef.no 

1 - Scheduling and routing of harvesting resources 




mikael.ronnqvist@nhh.no, David Bredström, Petrus Jönsson 

We describe an integrated scheduling and routing problem. Harvesting operations 

at forest harvest areas are done by harvesters and forwarders. Each team 

(a harvester and a forwarder) is assigned to a set of harvest areas. Each machine 

in a team and each harvest area have very different characteristics. The planning 

problem is complex and we propose a two phase solution method where 

we in the first we solve an assignment problem (teams to harvest areas) and in 

a second, the scheduling of each team over its assigned areas. We report results 

from a case with 46 machines and 968 harvest areas. 

2 - Solving The Dynamic Pickup and Delivery Problem with 

Time Windows Using Hybrid Local Search Metaheuristics 

Jawad Omari, Programmes, KADDB, Amman, Amman, Jordan, 

jawad82@batelco.jo 

A local courier firm offers same-day pickup and delivery services of small 

parcels using its own fleet. The service is booked in advance or requested during 

the day; each request has specific time periods to perform the service. The 

firm aims to minimize the total distance traveled by all vehicles. To do so the 

Dynamic Pickup and Delivery Problem with Time Windows model is used, and 

an online hybrid metaheuristic based on Variable Neighborhood Search, Tabu 

Search, and Guided Local Search is created to search for feasible solutions. 

The hybrid is tested on instances spanning over 30 days. 

3 - A Decision Support System for Cruise Yacht Scheduling 

and Routing 

Dimitris Paraskevopoulos, Management Science and Technology, 

Athens University of Economics and Business, Evelpidon 47A & 

Lefkados 33, Athens, Athens, Greece, dparaskevop@aueb.gr, 

Panagiotis Repoussis, Christos Tarantilis, George Ioannou 

This study presents a Decision Support System for cruise yacht scheduling and 

routing. The proposed DSS enables users to generate cruise plans (i.e. sequence 

of visits, arrival/departure dates and times) according to particular preferences. 

The latter constitutes a hard combinatorial optimization problem. To 

this end, the problem is mathematically depicted and a metaheuristic algorithm 

is designed and developed for solving it. Computational experiments illustrate 

the performance of the proposed methodology, while the deployment and operation 

of the DSS on a Greek company is also discussed. 

4 - Integrated Crew Pairing and Assignment by Column 

Generation and Dynamic Constraint Aggregation 

Issmail Elhallaoui, Math., Polytechnique, cp. 6079 succ. 

centre-ville, H3C 3A7, Montreal, Qué., Canada, 

issmail.elhallaoui@gerad.ca, Francois Soumis, Mohamed 

Sadoune, Guy Desaulniers 

The crew scheduling problem is commonly decomposed into two stages which 

are solved sequentially. Crew pairing generates a set of pairings covering all 

flight legs. Crew assignment generates anonymous blocks covering all pairings. 

The simultaneous problem generates a set of blocks covering all flight 

legs. It is a large set covering problem highly degenerated. We combine the 

column generation and the dynamic constraint aggregation methods to solve it 

and save up to 9% on real-life problems. 

243


� WB-32 


8.2.17 

Energy and technological system issues 


and Climate 


Chair: Vincent Mazauric, Strategy & Technology, Schneider Electric, 

38TEC/T3 Building, 37 Quai Paul-Louis Merlin, 38050, Grenoble, 

France, vincent.mazauric@schneider-electric.com 

1 - Energy transmission in electrical engineering: A multiscale 

approach 

Vincent Mazauric, Strategy & Technology, Schneider Electric, 

38TEC/T3 Building, 37 Quai Paul-Louis Merlin, 38050, 

Grenoble, France, vincent.mazauric@schneider-electric.com, 

Nadia Maïzi 

In order to address the abysmal lack of efficiency of the electrical system (73% 

of losses, 45% of CO2 emissions worldwide. . . !), the laws of electromagnetism 

are addressed through a thermodynamic-based optimal problem. This approach 

is shown supporting a multi-scale analysis where the brute minimization condition 

is replaced by embedded minimizations on the various scales excited by 

the power electrical system. Following the previous thermodynamic viewpoint, 

these various scales are reviewed from deep within the material to the whole 

electrical system. 

2 - Optimisation of a multi sources district heating network 

using Floyd and Moore algorithms 

Ghanassia Elie, EDF R&D, 75, Paris, elieghan@yahoo.fr 

District heating systems are usually designed for the distribution of centralised 

heat generation. However, mutualisation fits better to an urban network operating 

mainly from distributed renewable energies. 

This paper deals with multi sources network optimization. A mathematical 

model, using Moore algorithm in graph theory to optimize the energy transmitted 

per linear meter has been developed. It implies the utilization of a 

geographic system of information for a better acquisition of inputs and visualisation 

of outputs. The results are presented for a real case of a French city 

district. 

3 - Optimization of future power systems focusing on reliability 

of supply 

Mathilde Drouineau, Center for Applied Mathematics, Mines 

Paristech, 1 rue Claude Daunesse, BP N207, F-06904, Sophia 

Antipolis Cedex, France, 

mathilde.drouineau@mines-paristech.fr, Nadia Maïzi, Vincent 

Mazauric 

The overall efficiency of future power systems is expected to improve with renewable 

and distributed energy sources, as the level of losses induced by the 

Carnot cycles are decreased. However these sources challenge reliability of 

supply and may induce extra-losses. In this work, we apply variational principles, 

deduced from thermodynamics, to take reliability into account and assess 

the overall amount of losses. It appears that centralized power systems are disadvantaged 

by the efficiency of the Carnot cycles, whereas distributed systems 

are penalized by the losses induced by reliability. 

4 - The diversity of Canadian energy policies: an illustration 

of emerging opportunities using the new model 

TIMES-Canada 

Jean-Philippe Waaub, Geography, UQAM, CP 8888 succ. 

Centre-Ville, H3C 3P8, Montreal, Quebec, Canada, 

waaub.jean-philippe@uqam.ca, Kathleen Vaillancourt, Olivier 

Bahn, Richard Loulou 

Regarding energy production, Canada is an important player on the world 

scene while energy security represents a major challenge for non-producing 

provinces. Provincial energy systems are diversified and a national energy strategy 

is missing to optimize the conception of energy policies. Our objective is to 

analyze the role of renewable electricity using the new technology-rich model 

TIMES-Canada in two economic scenarios up to 2030. We perform sensitivity 

analyses on interesting matters for policy making such as electricity exports, 

oil sands development, liquefied natural gas imports, etc. 

244 

� WB-33 


8.2.19 

Carbon markets 



Chair: Pauline Barrieu, Statistics, London School of Economics, 

Houghton street, WC2A 2AE, London, United Kingdom, 

p.m.barrieu@lse.ac.uk 

Chair: Max Fehr, Statistics department, LSE, WC2A2AE, London, 

United Kingdom, m.w.fehr@lse.ac.uk 

1 - Existence of a supply function equilibrium for electricity 

markets 

Edward Anderson, Faculty of Economics and Business, 

University of Sydney, NSW 2006, Sydney, NSW, Australia, 

edward.anderson@sydney.edu.au 

Electricity markets can be modelled using supply function equilibrium (SFE). 

Often there are asymmetric SFE that are independent of the demand distribution, 

called ’Strong SFE’. We consider an asymmetric duopoly with fixed 

marginal costs and capacity constraints. When at low demand scenarios only 

the cheaper firm is used, and at high demand scenarios the more expensive firm 

is at its capacity, then a strong SFE may not exist. We demonstrate an equilibrium 

involving a vertical segment, which is not a strong SFE. In very general 

circumstances there is exactly one SFE in pure strategies. 

2 - Option Pricing in the European Unions Emission Trading 

Scheme 

Max Fehr, Statistics department, LSE, WC2A2AE, London, 


We propose a model for risk neutral futures price dynamics in the European 

Unions Emissions Trading Scheme (EU ETS). Historical price dynamics suggests 

that both allowance prices for different compliance periods and CER 

prices for different compliance periods are significantly related. To obtain a 

realistic price dynamics we take into account the specific details of the EU ETS 

compliance regulations, such as banking and the link to the Clean Development 

Mechanism (CDM), and exploit arbitrage relationships between futures on EU 

allowances and Certified Emission Reductions. 

3 - On fair pricing of emission-related derivatives 

Juri Hinz, Logistics, Zurich Universtity of Applied Sciences, 

IDP, Rosenstrasse 3, CH-8401 , Winterthur, Switzerland, 

hizr@zhaw.ch 

Tackling climate change is at the top of many agendas. In this context, emission 

trading schemes are considered as promising tools. The regulatory framework 

of an emission trading scheme introduces a market for emission allowances 

and creates need for risk management by appropriate financial contracts. In 

this work, we address logical principles underlying their valuation. 

4 - Pricing CO2 permits using approximation approaches 

Rudiger Kiesel, University Duisburg Essen, 47048 , Duisburg, 

Germany, Ruediger.Kiesel@uni-due.de 

Equilibrium models have been widely used in literature with the aim of showing 

theoretical properties of emission trading systems. This paper derives first 

a new equilibrium model. Second, it is shown that the theoretical permit price 

is related to changes in the expectation about how long regulated companies 

will need to exhaust the remaining permits. Third, by application to real data it 

demonstrates that emission trading systems are inherently prone to jumps. 

� WB-34 


8.2.23 

Fast algorithms for large matrix 

optimization problems 



Chair: Michel Baes, IFOR, ETH, HG.G.22.1, Ramistrasse 101„ 

8092, Zurich, Switzerland, michel.baes@ifor.math.ethz.ch

1 - Smoothing techniques for solving structured semidefinite 

programs 

Michael Bürgisser, IFOR, ETH, HG G 22.4, Rämistrasse 101, 

8092, Zürich, Switzerland, 

michael.buergisser@ifor.math.ethz.ch, Michel Baes 

We use smoothing techniques to solve approximately mildly structured 

semidefinite programs with many constraints. As smoothing techniques require 

a specific problem format, we introduce an alternative problem formulation 

that fulfills the structural assumptions. The resulting algorithm has a 

complexity that depends linearly both on the number of constraints and on the 

inverse of the accuracy. In the numerical experiments we performed, smoothing 

techniques compare favorably with interior-point methods for very large-scale 

instances. 

2 - Randomised first-order algorithms for large semidefinite 

optimization 

Michel Baes, IFOR, ETH, HG.G.22.1, Ramistrasse 101„ 8092, 

Zurich, Switzerland, michel.baes@ifor.math.ethz.ch 

Solving large-scale and dense semidefinite optimization problems is an extremely 

difficult task: the size of the problem rules out standard first-order 

schemes, let alone interior-point methods. Pursuing a research direction initiated 

by Arkadi Nemirovski, we develop and study some first-order methods 

where subgradient computations can be accelerated by randomization. We report 

some numerical experiments on this line. 

3 - An alternating direction algorithm for finding jointly 

sparse extremal eigenvectors 

Peter Richtarik, University of Edinburgh, United Kingdom, 

peter.richtarik@ed.ac.uk 

In this paper we consider the problem of finding jointly sparse minimal and 

maximal eigenvectors of a symmetric matrix. We first formulate the problem 

as a nonconvex optimization program with a sparsity-inducing penalty term and 

then propose and analyze a simple alternating direction algorithm for solving 

it. It turns out that the emergence of sparsity in the iterates of the method can be 

explained in geometrical terms. We finish the discussion by giving some preliminary 

computational results on random matrices and outline an application 

to compressed sensing. 

4 - Kernel-Based Interior-Point Methods for Monotone LCP 

over Symmetric Cones 

Goran Lesaja, Mathematical Sciences, Georgia Southern 

University, 203 Georgia Ave., 30460-8093, Statesboro, Georgia, 

United States, goran@georgiasouthern.edu, Kees Roos 

We present a generic interior-point method for monotone LCP over symmetric 

cones that is based on barrier functions which are defined by a large class of 

univariate functions called eligible kernel functions. Furthermore, the method 

uses Nesterov-Todd search directions. We provide a unified analysis of the 

method and give a general scheme on how to calculate the iteration bounds for 

the entire class. For some specific eligible kernel functions we match the best 

known iteration bound for large-step methods while for short-step methods the 

best iteration bound is matched for all cases. 

� WB-35 


6.2.46 

MINLP Reformulations and Applications 



Chair: Leo Liberti, LIX, Ecole Polytechnique, LIX, Ecole 

Polytechnique, 91128, Palaiseau, France, leoliberti@gmail.com 

1 - Expectation and Chance-constrained Models and Algorithms 

for Insuring Critical Paths 

Cole Smith, Industrial and Systems Engineering, University of 

Florida, PO Box 116595, 32611, Gainesville, FL, United States, 

cole@ise.ufl.edu, Siqian Shen, Shabbir Ahmed 


We consider a critical path network problem, where a project is completed only 

after a series of dependent tasks are finished. Task durations are uncertain, but 

can be insured a priori to mitigate potential delays. One must balance costs 

incurred in insuring arcs with penalties associated with late project completion 

times, where lateness penalties are assumed to be lower semi-continuous nondecreasing 

functions of completion time. We employ RLT to make the problem 

amenable to solution via Benders decomposition, and demonstrate the efficacy 

of our approach by testing on a test-set of problems. 

2 - Robust formulations for Mixed Integer Non-Linear Programs: 

uncertainty in routing problems 

Hassan Hijazi, Orange Labs & LIF, Orange Labs 

R&D/CORE-TPN, 38-40 rue du General Leclerc cedex 9, 92794, 

Issy-Les-Moulineaux, France, 

hassan.hijazi@orange-ftgroup.com, Pierre Bonami, Adam 

Ouorou 

In Telecommunications, operators usually use market surveys and statistical 

models in order to estimate the evolution of networks’ traffic, which often leads 

to weak management planning decisions. Motivated by this reality, we introduce 

elements of Robust Optimization theory for Mixed Integer Non-Linear 

Programs modeling multi-flow delay constrained routing problems. We write 

and compare different robust formulations offering different protection levels. 

Computational experiments are developed in order to evaluate the "price of robustness" 

and to assess the quality of proposed models. 

3 - The signomial global optimization algorithm — some 

recent advances 

Andreas Lundell, Process Design and Systems Engineering, Åbo 

Akademi University, Biskopsgatan 8, FIN-20500, Turku, 

Finland, andreas.lundell@abo.fi, Tapio Westerlund 

The signomial global optimization (SGO) algorithm can be used to solve 

MINLP problems containing nonconvex signomial functions to global optimality. 

In the algorithm, convex underestimators for the nonconvex signomial 

terms are obtained using single-variable power and exponential transformations 

approximated with piecewise linear functions. In this presentation, some 

recent advances in regard to this method are discussed, including how it is possible 

to solve problems where variables with nonpositive bounds need to be 

transformed using translations.in regards 

4 - Reduced Reformulation-Linearization Technique for 

Polynomial Programs 

Sonia Cafieri, LOTA, Ecole Nationale d’Aviation Civile, 31055, 

Toulouse, France, sonia.cafieri@recherche.enac.fr, Pierre 

Hansen, Lucas Létocart, Leo Liberti, Frederic Messine 

Reduced RLT (rRLT) is a special class of Reformulation-Linearization Technique 

(RLT). This reformulation was originally defined for nonconvex, both 

continuous and mixed-integer, quadratic programming problems subject to linear 

equality constraints. It is obtained by replacing some of the quadratic terms 

with suitable linear constraints, called rRLT constraints. We present an extension 

of the rRLT theory to the case of general polynomial programs. We also 

show a strategy to choose the basis of a matrix involved in the rRLT constraints 

generation so as to improve the chances of tightening the lower bound of the associated 

convex relaxation. This allows to improve the performance of a spatial 

Branch-and-Bound algorithm applied to nonconvex NLP and MINLP problems 

where such convex relaxation is exploited. 

� WB-36 


3.1.05 

News from Commercial MIP Solvers 



Chair: Daniel Junglas, CPLEX Development, IBM Deutschland, An 

den Frankengräbern 1, 55129, Mainz, Germany, 

daniel.junglas@de.ibm.com 

1 - Recent improvements to ILOG IBM CPLEX 

Daniel Junglas, CPLEX Development, IBM Deutschland, An 

den Frankengräbern 1, 55129, Mainz, Germany, 

daniel.junglas@de.ibm.com 

245


We report on features recently added to ILOG IBM CPLEX. This includes performance 

improvements in particular for parallel algorithms (MIP, barrier) and 

solvers for quadratic problems as well as improved stability analysis tools. We 

also present some of the improved interoperability features introduced in the 

latest versions of ILOG IBM CPLEX. 

2 - Recent improvements to FICO Xpress 

Oliver Bastert, FICO, Maximilianstr. 35a, 80539, Munich, 

Germany, oliverbastert@fico.com, Richard Laundy 

In this talk we take a look at some of the recent improvements to FICO Xpress. 

We discuss the addition of new modeling constructs to handle logical constraints 

and describe the new features for handling multiple solutions. On the 

performance side we present computational results showing the improvements 

to the Xpress parallel MIP solver and quadratic simplex solvers 

3 - Primal Heuristics and Related Features of the SAS/OR 

MILP solver 

Philipp Christophel, Operations Research R&D, SAS Institute, 

100 SAS Campus Drive, 27513, Cary, NC, Germany, 

Philipp.Christophel@sas.com 

This talk discusses new features and performance improvements of the 

SAS/OR MILP solver. It focuses on current research and development in the 

area of primal heuristics. This includes a discussion about concepts like heuristic 

strategies, types of primal heuristics, and solution pools. Furthermore it 

demonstrates how the implementation of these concepts influences the performance 

of the SAS/OR MILP solver and how it leads to new possibilities for 

SAS/OR users. 

4 - Preprocessing in Linear-Fractional Programming 

Anett Racz, Applied Mathematics and probability theory, 

University of Debrecen, Egyetm tér 1, 4032, Debrecen, Hungary, 

racz.anett@inf.unideb.hu 

Most of the professionally developed solvers automatically use preprocessing 

techniques to maintain numerical stability and improve performance. In this 

paper we describe the main results of our investigations connected with preprocessing 

techniques in Linear-Fractional Programming (LFP),which are based 

on the use of well-known preprocessing techniques ([Mészáros, Shul (2003)], 

[Andersen (1995)]) of linear programming and we adapt them to LFP Problems. 

� WB-37 


3.1.09 

MCDA and Public Administration 



Chair: Norese Maria Franca, DSPEA, Politecnico di Torino, C.so 

Duca degli Abruzzi nr. 24, 10129, Torino, Italy, Italy, 

mariafranca.norese@polito.it 

1 - Using a MCDA methodology for evaluating social acceptance 

of hydrogen technology in France: from collective 

perceptions to criteria 

Sébastien Damart, M-Lab (DRM) - Université Paris Dauphine / 

CNAM (Chaire EGSS), Université Paris Dauphine Place du Mal 

de Lattre de Tassigny, 75775, Paris, France, 

sebastien.damart@cnam.fr, Amidou Kpoumie, Benjamin 

Rousval, Alexis Tsoukiàs 

A MCDA methodology has been used within a research project on the evaluation 

of the social acceptance of hydrogen (H2) technology in France. At a 

first stage we collected shared beliefs of citizens about implications of the use 

of H2 technology. For that purpose focus groups have been organised leading 

to the construction of collective causal maps. During this stage, scenarios of 

what would be the evolution of H2 technology use in the next decades have 

been identified using mainly expert reports and results of already performed 

forecasts. In the second stage, the causal maps have been synthesised in order 

to highlight the underlying values and a set of criteria upon which the scenarios 

could be assessed. 

246 

2 - An integrated use of cognitive mapping and multi criteria 

models to support communication and innovation 

Maria Franca Norese, DISPEA Production Systems and 

Economics, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 

10129, Torino, Italy, mariafranca.norese@polito.it, Simone 

Griffa, Chiara Novello 

Land monitoring, to plan or control activities, is one of the main functions 

of the public administration. A new technology, which includes Unmanned 

Aerial Vehicle platforms, could be an interesting proposal but also a critical 

situation because several public and private organisations should be involved 

in the innovation process and different decisional uncertainties and complexities 

are present and could negatively impact the process. An integrated use of 

cognitive maps and multicriteria models allowed us to identify the key actors, 

orient the inquiry and organize all the information elements towards a global 

definition of some system alternatives and their evaluation and selection. 

3 - The Strategic Choice Approach (SCA) for structuring 

decisional problems in the context of public projects. 

The Turin East Ring Road case study. 

Diana Rolando, Casa Città, Turin Polytechnic, Istituto Galileo 

Ferraris, Corso Massimo d’Azeglio, 10125, Turin, Italy, 

diana.rolando@polito.it 

During the first phase of complex plans it is necessary to analyse uncertainties 

and risks associated with the project, to define strategic decisions as well 

as concrete solutions. SCA is a methodology that adopts a multicriteria approach 

to shape decision problems, design and compare solutions and control 

uncertainties, in order to assist decision makers from the involved organizations. 

Recently SCA has been tested on a complex public project in Turin, to 

support the decisions to be taken, the criteria for an ELECTRE application and 

the uncertainties to be analyzed by all the stakeholders. 

4 - Real Option Analysis as a Decision Aiding tool 

Giulia Lucertini, Università degli Studi di Padova, via venezia 1, 

35131, Padova, Italy, Italy, giulialucertini@hotmail.com, Chiara 

D’Alpaos, Alexis Tsoukiàs 

In our research we propose to use Real Options Analysis (ROA) as a decision 

aiding tool. In the presentation we first compare customary Cost Benefit Analysis 

(CBA) to ROA and we focus on the dynamic dimension of ROA which 

is missing to CBA. We then compare ROA to decision under risk and uncertainty 

emphasising the concept of information value as options value. We then 

explore the possibility of using ROA as a tool for generating alternatives in 

public policy evaluation. 

� WB-38 


6.2.44 

Lot-sizing models 



Chair: Wilco van den Heuvel, Econometric Institute, Erasmus 

University Rotterdam, Burg. Oudlaan 50, P.O. Box 1738, 3000DR, 

Rotterdam, Netherlands, wvandenheuvel@ese.eur.nl 

1 - Reformulations of the economic lot-sizing problem with 

remanufacturing 

Mathijn Retel Helmrich, Erasmus School of Economics, 

Erasmus University Rotterdam, Postbus 1738, 3000 DR, 

Rotterdam, Netherlands, retelhelmrich@ese.eur.nl, Wilco van 

den Heuvel, Raf Jans, Albert Wagelmans 

The classic lot-sizing problem has been extended with a remanufacturing option. 

In each period, we can choose to set up a process to remanufacture returned 

products or produce new items. These processes can have separate or 

joint set-up costs. We show that both cases are NP-hard. A "natural’ MIP formulation 

contains big M constraints. Therefore, we propose several alternative 

formulations, inspired by reformulations of the classic problem, namely Eppen 

& Martin’s shortest path reformulation, Van Vyve & Wolsey’s partial shortest 

path reformulation and the (l,S,WW) valid inequalities.

2 - Gray Markets and All Units Discounts 

Mengze Shi, rotman school of management, university of 

toronto, 105 st george street, m5s 3e6, toronto, ontario, Canada, 

mshi@rotman.utoronto.ca 

Gray market is the unauthorized channel of distribution for a supplier’s authentic 

products. This paper studies the gray market caused by a reseller responding 

to the supplier’s all-unit quantity discount offerings. We perform closed 

form analysis of the reseller’s dynamic optimal lot-size decisions in response 

to supplier policies. A novel feature of our solution is the precisely expressed 

relationship between holding costs and the supply of goods to the gray market. 

We also identify conditions under which a profit maximizing supplier is more 

likely to accommodate the gray market. 

3 - Effective replenishment policies for the multi-item dynamic 

lot-sizing problem with inventory limited 

Jose M Gutierrez, Estadística, Investigación Operativa y 

Computación, Universidad de La Laguna, Facultad de 

Matemáticas, Av. Astrofísico Fco. Sánchez s/n., 38271, La 

Laguna, Tenerife, Spain, jmgrrez@ull.es, Marcos Colebrook, 

Beatriz Abdul-Jalbar, Joaquín Sicilia 

We address the dynamic lot-sizing problem considering multiple items and 

storage capacity. Despite we can easily characterize a subset of optimal solutions 

just extending the properties of the single-item case, these results are not 

helpful to design an efficient algorithm. Accordingly, heuristics are appropriate 

approaches to obtain near-optimal solutions for this NP-hard problem. Thus, 

we propose a heuristic procedure based on the smoothing technique, which is 

tested on a large set of randomly generated instances. 

4 - Some lot-sizing models with perishable items 

Wilco van den Heuvel, Econometric Institute, Erasmus 

University Rotterdam, Burg. Oudlaan 50, P.O. Box 1738, 

3000DR, Rotterdam, Netherlands, wvandenheuvel@ese.eur.nl, 

Mehmet Onal, Edwin Romeijn 

We consider a lot-sizing model with perishable items. We assume that items deteriorate 

completely after a deterministic lifetime and can not be sold thereafter. 

Furthermore, the order in which items are distributed to the customer depends 

on the customer preference and the way items are exposed to the customer 

by the store manager. In this way items may be consumed in four different 

manners: (i) Last-Expired, First-Out, (ii) First-Expired, First-Out, (iii) First-In, 

First-Out, or (iv) Last-In, First-Out. We consider the complexity and propose 

algorithms for the different models. 

� WB-39 


6.2.45 

Scheduling and Pricing 



Chair: Wieslaw Kubiak, Faculty of Business Administration, 

Memorial University, Prince Philip Drive, A1B 3X5, St. John’s, NL, 

Canada, wkubiak@mun.ca 

1 - Dynamic resource scheduling 

Joanna Jozefowska, Institute of Computing Science, Poznañ 

University of Technology, Piotrowo 2, 60-965, Poznañ, 

Wielkopolska, Poland, jjozefowska@cs.put.poznan.pl, Lukasz 

Jozefowski, Wieslaw Kubiak 

A scheduling problem is considered where multiple users with various importance 

compete for a scarce resource and the resource allocation should guarantee 

appropriate service rate objectives for the users. We consider so called 

dynamic environment in which the parameters describing instances of the problem 

may change in an unpredictable way. The problem occurs within a broad 

spectrum of systems, including databases, mediabased applications, and networks. 

In this paper a scheduling algorithm based on the divisor methods of 

apportionment is presented for solving the problem. 

2 - An overview on shop scheduling with minimum time 

lags 

Djamal Rebaine, Informatique et mathématique, Université du 

Québec à Chicoutimi, 555, boul. de l’Université, G7H 2B1, 

Chicoutimi, Québec, Canada, drebaine@uqac.ca 


We consider the two-machine shop problems with unit-time operations and 

minimum time lags. The goal is to seek a schedule with a minimium makespan. 

In this talk we give an overview for the flow shop, open shop, single machine 

with coupled operations, and generalizations with parallel machines. We start 

with preliminary results. Then, for each of the mentioned models, we present 

results on well solvable cases and worst-case analysis. 

3 - Makespan Minimization of Multi-Slot Just-In-Time 

Scheduling on Single and Parallel Machines 

Dariusz Dereniowski, Algorithms and System Modeling, Gdańsk 

University of Technology, ul. Gabriela Narutowicza 11/12, 

80-233, Gdańsk, Poland, deren@eti.pg.gda.pl, Wieslaw Kubiak 

In this talk we address the problem of minimizing the number of slots and 

makespan of multi-slot just-in-time schedules. We are interested in the latter 

(more gerenal) optimization criterion. Several algorithmic results are presented, 

including efficient optimal algorithms for the single machine problem 

and for a multiple machine case when the processing time of each job does 

not exceed its due date; and a polynomial-time approximation algorithm for a 

general case on parallel machines. 

An interesting application of this task scheduling model is a routing problem in 

a slotted ring network. 

4 - Bullwhip effect emergence into an after-sales spare 

part service supply chain in Telecom firms - A complex 

system approach 

Mauricio Flores, Systems, Instituto Politecnico Nacional, 

Petirrojo #9, Las Alamedas, 52970, Atizapan de Zaragoza, 

Estado de Mexico, Mexico, fcmauricio@yahoo.com, Oswaldo 

Morales Matamoros, Ricardo Tejeida Padilla, Isaias Badillo Piña 

Telecom Equipment Manufacturers have a great opportunity to capture revenue 

and profit from telecommunications service providers by provide them 

after-sales spare part services of repairable units. The challenge is to match the 

supply process and demand process in order to support 99.999% of availability 

of the telecom network. Unfortunately these two processes cannot match 

perfectly, and the effect is the bullwhip effect. In order to cope with this problem, 

this paper studies the bullwhip effect emergence under a complex system 

approach. 

� WB-40 


6.2.52 

Optimal Control and Design in Applications 

Stream: Engineering Optimization 


Chair: Volker Schulz, University of Trier, 54296, Trier, Germany, 

Volker.Schulz@uni-trier.de 

Chair: Moritz Diehl, ESAT, K. U. Leuven, Kasteelpark Arenberg 10, 

3001, Leuven, Belgium, moritz.diehl@esat.kuleuven.be 

1 - Aerodynamic Shape Optimization under Uncertainty 

Claudia Schillings, University of Trier, Germany, 

claudia.schillings@uni-trier.de, Volker Schulz 

A novel approach towards stochastic distributed uncertainties is discussed for 

the specific application of shape uncertainties in aerodynamic design. While 

the random field of uncertainties is approximated by a goal-oriented Karhunen- 

Loeve expansion, an adaptively refined sparse grid is used to discretize the resulting 

probability space. Algorithmic approaches based on multiple-setpoint 

ideas in combination with one-shot methods will be presented as well as numerical 

results. 

2 - Large Scale Aerodynamic Shape Optimization 

Stephan Schmidt, University of Trier, Universitätsring 15, FB IV 

- Mathematics, 54296, Trier, Germany, 

Stephan.Schmidt@uni-trier.de, Volker Schulz 

Shape calculus techniques for aerodynamic shape optimization are presented 

that result in a shape derivative existing only on the surface of the aircraft. Thus, 

this adjoint based sensitivity can be computed very efficiently and a large deformation 

of the aircraft shape is possible using every CFD mesh node position 

as a design parameter. The resulting loss of regularity is treated by considering 

the shape Hessian, which is derived for Stokes and Navier-Stokes flows using 

operator symbols. 

247


3 - Double Smoothing Algorithm for a class of Optimal 

Control problems 

Olivier Devolder, CORE, Université catholique de Louvain 

(UCL), Louvain-la-Neuve, Belgium, 

olivier.devolder@uclouvain.be, François Glineur, Yurii Nesterov 

We consider an optimal control problem governed by a linear differential system. 

The constraints are formed by some convex bounds on the states at finite 

number of instants, and by point-wise convex constraints on the control. We 

tackle this class of problems by a dual approach, without preliminary discretization. 

Dualization of the state constraints leads to a non-smooth convex problem 

in finite dimension. We apply a double smoothing. Our strategy is supported 

by a worst-case complexity analysis. 

� WB-41 


3.1.06 

Long Term Financial Decisions 



Chair: Ursula Walther, Frankfurt School of Finance and 

Management, Sonnemannstrasse 9-11, 60314, Frankfurt/M, 

Germany, u.walther@frankfurt-school.de 

1 - Is Timing Money? The Return Shaping Effect of Technical 

Trading Systems 

Peter Scholz, Frankfurt School of Finance and Management, 

60314, Frankfurt am Main, Germany, 

p.scholz@frankfurt-school.de 

The success of technical trading systems still seems puzzling. Previous research 

is mainly based on historical backtests. This work investigates the hypothesis 

that trend following systems should pro 

t from autocorrelated returns. Therefore, we test trading systems on asset returns 

which were simulated by dierent parameterized stochastic processes. To 

evaluate the performance, the return distribution is compared to the buy-andhold 

strategy, applying concepts of stochastic dominance and expected utility. 

2 - Development of Service Quality (SQ) reference model 

in Private Banking (PB) 

Erkan Sengün, Institut für Management, Universität 

Koblenz-Landau, 56070, Koblenz, 

erkansenguen@uni-koblenz.de 

Is the high level of service quality the right way of success for the PB Market ? 

Or is "good’ Service quality far enough? The purpose of this research is to construct 

a measurement instrument to capture sustainable and optimal SQ in the 

German PB market. It allows to achieve competitive advantage and a better understanding 

of the subject perception (customer satisfaction, trust, commitment 

and customer loyalty, etc.). Due to the banker’s discretion there are almost no 

representative data about the wealth PB clients available. 

3 - Diversification effects of asset price process parameters 

— an empirical investigation 

Ursula Walther, Frankfurt School of Finance and Management, 

Sonnemannstrasse 9-11, 60314, Frankfurt/M, Germany, 

u.walther@frankfurt-school.de, Andryi Fetsun 

Higher moments of asset price distributions — especially skewness — have 

long been recognized as important characteristics in asset pricing and risk management. 

However, it is less well known that portfolio characteristics other than 

variance may not diversify on a portfolio level but even accumulate. We study 

this behavior based on a parameterized description of asset price processes using 

GARCH-type models and non-normal increments. We analyze historical 

buy-and-hold-portfolios of German stocks and also study aggregation effects 

of simulated returns. 

248 

� WB-42 


3.1.07 




Chair: Ali Eshragh Jahromi, School of Mathematics and Statistics, 



Ali.EshraghJahromi@unisa.edu.au 

1 - The impact of uncertainty in Decision Support for waste 

management organizations 

Ali Eshragh Jahromi, School of Mathematics and Statistics, 



Ali.EshraghJahromi@unisa.edu.au, Jerzy Filar 

By directive of the EU existing substances like plastics or glass waste have to be 

collected, recycled and commercialized as secondary raw materials. While the 

value of these substances is small, the expenses for recycling and transportation 

are considerable. Therefore waste management organizations are interested in 

minimizing their expenses for this process. For the case of deterministic input 

data a generalized two-stage assignment and transportation model for this realworld 

problem has been presented before. This model has been integrated into 

a custom-made DSS, and generates substantial savings for the waste management 

organization. In this talk we highlight the impact of uncertainty in input 

data for the problem under consideration. Furthermore we present an adaption 

of the previous model, which is able to cope with this kind of uncertainty. 

Finally we present computational results for the new model, and discuss the 

benefits of considering uncertainty. 

2 - A combined method to deal with conflicting software 

requirements 

Catarina Gomes, Uninova-CA3, Casa Emilio Cebola, Casal das 

Figueiras, 2970-261, Sesimbra, Portugal, 

catarina.alex.gomes@gmail.com, Rita Ribeiro 

The objective of this work is to contribute to the resolution of conflicting situations 

between software requirements (e.g security requirement may affect 

negatively the performance requirement) that may appear during software development 

life cycle. Thus, it is important to provide support regarding the 

decision of which requirements are more relevant and the priority order in its 

implementation. To this aim we used a search graph algorithm, called BFS 

(Best First Search), which seems appropriate for finding the ideal solution, i.e. 

the one responsible for satisfying or implementing all conflicting requirements, 

with the lower cost and pointing the ideal implementation path ordering. 

� WB-43 


8.2.02 

Sustainable Construction Processes 



Chair: Jana Šelih, Faculty of Civil and Geodetic, University of 

Ljubljana, Jamova 2, 1000, Ljubljana, Slovenia, jselih@fgg.uni-lj.si 

1 - Sustainability assessment of construction processes 

Jana Šelih, Faculty of Civil and Geodetic, University of 

Ljubljana, Jamova 2, 1000, Ljubljana, Slovenia, 

jselih@fgg.uni-lj.si, Aleksander Srdic 

Recently the definition of quality has been extended to a more comprehensive 

level, which also includes sustainable performance of products and processes. 

This is especially valid for construction. The purpose of the paper is to develop 

a rational and unambiguous method to assess the sustainable performance of 

the complete construction production chain. The developed multi-criteria decision 

method is robust and easy-to-use to facilitate the implementation of the 

method in practice. It is based on Life Cycle Assessment (LCA). A case study 

analysis is presented.

2 - New Classification of Construction Companies: Overhead 

Costs Aspect 

Rasa Apanaviciene, Dept of Civil Engineering Technologies, 

Kaunas University of Technology, Studentu str. 48-401, 

LT-51367, Kaunas, Lithuania, rasa.apanaviciene@ktu.lt, Ala 

Daugeliene 

The traditional classification of construction companies depending on the number 

of employees is not appropriate when analysing the competitiveness of construction 

companies. The collected data of Lithuanian construction companies 

was analysed by applying statistical methods and the construction companies 

were classified into competitiveness classes according to the relative value of 

the overhead costs. The new classification provides the basis for economical 

evaluation of the construction companies and modelling of their competitiveness 

in regard to the value of overhead costs as well as applying the competitive 

advantages for the estimation of construction bidding price. 

3 - ERP system implementation in Latvian construction 

company 

Andrejs Tambovcevs, Riga Technical university, Latvia, 

ata2000@inbox.lv 

ERP systems have the potential to integrate seamlessly organizational processes 

using common shared information and data flows. The purpose of the 

study is to investigate ERP system implementation process in the construction 

company in Latvia. The study briefly described the business processes involved 

in the company and illustrated how ERP systems could be implemented and 

the efficiency of management system consequently enhanced. This study also 

argues that ERP systems are important source of organizational change with 

major implications for the organization and management of work. 

4 - Environmental Management Systems Experience 

among Latvian Construction Companies 

Tatjana Tambovceva, Faculty of Engineering Economics and 

Management, Riga Technical University, Mezha street 1/7- 213, 

LV-1048, Riga, Latvia, tatjana.tambovceva@rtu.lv, Ineta Geipele 

The concept of sustainable development has become widespread amongst government 

agencies, politicians, corporations and other organizations around the 

world. The purpose of this study is to investigate experiences and effects of ISO 

14001 in Latvian construction companies. Our results show that the companies 

primarily expect an external recognition of the EMS activities. ISO 14001 often 

leads to reduced environmental impact, especially in the area of waste. The 

authors conclude that it is necessary to let all personnel to participate in work 

with the EMS as early as possible. 

� WB-44 


8.2.03 

Vector and Set-Valued Optimization III 



Chair: Enrico Miglierina, Dipartimento di Economia, Università 

dell’Insubria, via Monte Generoso 71, 21100, varese, Italy, 

enrico.miglierina@uninsubria.it 

1 - Characterizations of Calmness and Subregularity of 

Constraint Set Mappings 

Helmut Gfrerer, Institute for Computational Mathematics, 

Johannes Kepler University Linz, Altenbergerstr 69, A-4040, 

Linz, Austria, gfrerer@numa.uni-linz.ac.at 

The existence of nondegenerate multipliers in first-order necessary conditions 

at a local minimizer is related to the validity of some constraint qualification 

condition, for instance the property of metric subregularity of the constraint 

set mapping or equivalently, the calmness property of the solution mapping. 

In this talk we present characterizations of calmness/subregularity. We will 

see that there are some limitations when using exclusively first-order analysis, 

which can be bypassed assuming some part of the constraint mapping to be 

known subregular or by using second-order analysis. 

2 - Generalized convexity for multiobjective problems with 

conic constraints 

Beatriz HernÁndez, Economics, University Pablo de Olavide, 

Edificio N o 3, José Moñino - 2 a planta-despacho26, Ctra. de 


Utrera, Km. 1- 41013 Sevilla, 41013, SEVILLA, Spain, 

mbherjim@upo.es, Marko A. Rojas-Medar, Rafaela 

Osuna-Gómez, Antonio Rufián-Lizana 

Taking in mind Craven’s notion of K-invexity function and Martin’s notion of 

Karush-Kuhn-Tucker-invexity, we give a new notion of generalized convexity 

that is both necessary and sufficient to ensure every vector Karush-Kuhn- 

Tucker point is a weakly efficient solution for multiobjective problems with 

conic constraints. Moreover, it is the weakest to characterize the set of weakly 

efficient solutions.The notions and results that exist in the literature up to now 

are particular instances of the ones presented here. 

3 - Mountain Pass-type Theorem for Vector-Valued Functions 

Enrico Miglierina, Dipartimento di Economia, Università 

dell’Insubria, via Monte Generoso 71, 21100, varese, Italy, 

enrico.miglierina@uninsubria.it, Ewa Bednarczuk, Elena Molho 

The mountain pass theorem for scalar functions is a fundamental result of the 

minimax methods in variational analysis. Here we give a result that extends this 

approach to vector valued functions. Under suitable geometrical assumptions, 

we prove a theorem that ensures the existence of a critical point of the considered 

function f and we localize this point as a solution of a minimax problem 

for the function f. We remark that the considered minimax problem consists of 

an inner vector optimization problem and of an outer set-valued optimization 

problem. 

� WB-45 


8.2.12 

New Trends of Facility Logistics 

Stream: Facility Logistics 


Chair: Yugang Yu, Rotterdam School of Management, Erasmus 

University, T10-38,Burg. Oudlaan 50„ P.O. Box 1738, 3000 DR, 

Rotterdam, Netherlands, yyugang@rsm.nl 

1 - A Location Problem in Construction Management 

Huseyin Guden, Industrial Engineering, Baskent University, 

06531, Ankara, Turkey, hsyngdn@yahoo.com, Haldun Sural 

In construction projects (im)mobile concrete plants are located to build viaducts 

and tunnels on a line over a time horizon. There are costs of opening and moving 

a plant and transporting concrete to sites. The problem is to determine the 

number, type, and movement of the plants and to make the concrete production 

- allocation decisions so that the total cost is minimized. We develop two strong 

integer models for the problem. Using a real project data, we are able to solve 

the models to optimality. We also perform a sensitivity analysis on the solution 

and provide computational results. 

2 - Optimal Dislocation of Transport Junctions 

Vaclav Cempirek, Transport Technology and Control, Univerzita 

Pardubice, Studentstka 95, 53201, Pardubice, Czech Republic, 

hana.cisarova@upce.cz 

Optimization of transport junctions in Hub and Spoke system is very timely in 

the context of growing demands for quality and the speed of transport. The 

current trend of concentration of traffic flows into larger and economically less 

demanding transport session, it entails the issue of distribution hubs. There are 

many possible algorithms. Majority, these algorithms are heuristic. This contribution 

provides a detailed description of genetic algorithm that is suitable for 

solving the optimal location of transmission junctions. 

3 - Material handling systems - the future is now 

Dominik Berbig, Institut für Fördertechnik und Logistiksysteme, 

KIT, Gotthard-Franz-Straße 8, 76131, Karlsruhe, Germany, 

berbig@kit.edu, Christian Huber, Frank Schönung, Christoph 

Nobbe, Kai Furmans 

Requirements for material flow systems have undergone a great change. Flexibility, 

modularity and energy efficiency have become key aspects. There is a 

need for reconfigurable material flow systems and, consequently, adequate as 

well as compatible analysis. This paper deals twofold with this circumstance: 

In a first step, innovative (but close to industrial application) technologies to 

render manufacturing systems flexible are presented. In a second step, both, 

well-suited analysing techniques to quantify the benefits and recently developed 

simulative and analytical tools are illustrated. 

249


� WB-46 


8.2.14 

Stochastic Optimization Models 

Stream: Stochastic Programming 2 [c] 


Chair: Marta Kostrzewska, Institute of Mathematics, University of 

Silesia, ul. Bankowa 14, Katowice, 40-007, Katowice, Poland, 

marta.kostrzewska83@gmail.com 

1 - Strategic Capacity Planning under Dynamic Probabilistic 

Demand 

Ashis Chatterjee, Operations Management, Indian Institute of 

Management Calcutta, Joka, DH Road, 700104, Kolkata, West 

Bengal, India, ac@iimcal.ac.in, Dipankar Bose 

The existing single period models in Strategic Capacity Planning under demand 

uncertainty have been extended in this paper to a two period multi product 

model to capture seasonality of demand. Models have been developed for both 

Flexible and Dedicated Plant. While deciding on first period production, the 

actual demand is assumed as not known and demand for period 1 is known at 

the start of period 2. The resulting two-stage stochastic linear program has been 

solved by simulated data based optimization. Solutions have been examined to 

study the economics of Flexible and Dedicated plant. 

2 - The Problem of Minimum Cost Flow on Dynamic Generative 

Network Flows 

Seyed Ahmad Hosseini, Computer Science, University of 

Tehran, Tehran,University of Tehran, 0098, Tehran, Tehran, Iran, 

Islamic Republic Of, ahmad.s.hosseini@gmail.com 

This paper is concerned with a new class of network flows called dynamic generative 

network flows in which the flow commodity is dynamically generated 

at source nodes and dynamically consumed at sink nodes. It is assumed that 

the source nodes produce the flow according to time generative functions and 

the sink nodes absorb the flow according to time consumption functions. The 

minimum cost dynamic flow problem for a pre-specified time horizon T is defined 

and mathematically formulated and some methods are developed to solve 

the problem. 

3 - Stochastic single product network design model with 

lead time and safety stock considerations 

Mehdi SharifYazdi, Department of Industrial Engineering, 

University of Science and Culture, ’s Gravendijkwal 83, 3021 

EG, Rotterdam, Netherlands, mehdi.sharif@gmail.com, Leyla 

Ozsen, Behzad Bagheri 

The stochastic single product network design model with lead time and safety 

stock considerations (SSPNDLS) performs a tradeoff between capacity utilization 

and flexibility in long run. SSPNDLS simultaneously decides on locating 

DCs, selecting their capacity and allocating retailers to DCs regarding capacity 

and coverage constraints as well as demand and lead time uncertainty. The 

paper presents a nonlinear program minimizing total expected cost of location, 

pipeline inventory, and safety stock over a set of discrete scenarios and proposes 

a Lagrangian-relaxation-based algorithm. 

4 - Optimization methods for stochastic cost flow problem 

Marta Kostrzewska, Institute of Mathematics, University of 

Silesia, ul. Bankowa 14, Katowice, 40-007, Katowice, Poland, 

marta.kostrzewska83@gmail.com, Leslaw Socha 

The minimum cost flow problem in stochastic network is considered. The costs 

per unit of flow on the arcs of the network are assumed to be independent random 

variables and the criteria of minimization are the expected value and the 

second moment of the total cost of flow. This problem is transformed to the bicriteria 

minimum cost flow problem with the linear and the quadratic objective 

functions. Two sandwich methods for approximation of the efficient frontier of 

considered problem with the convergence proofs are proposed. Moreover, the 

approximate methods for the determination of the distribution of the total cost 

are studied. 

250 

� WB-47 


8.2.16 

Activities for Popularization of Science 

Stream: Young People for System Theory, Optimization 

and Education 


Chair: Kateryna Pereverza, Students Science Association, National 

Technical University of Ukraine, Kyiv, Revutskogo, 19/1, app. 282, 

02091, Kyiv, Ukraine, pereverza.kate@gmail.com 

1 - Promoting and Disseminating Science among Youth: 

Peculiarities and Perspectives 

Ielyzaveta Korotchenko, National Technical University of 

Ukraine "Kyiv Politechnic Institute", prospect Peremohy 37, 

03056, Kyiv, Ukraine, Korotchenko.liza@gmail.com 

The paper considers the role of promoting of science in Ukraine due to its 

significant role in the development of the country. Talented youth, involved 

into science, can make a breakthrough using new technologies and inventions. 

The affective ways of disseminating science, encouraging young people for 

scientific research, intellectual endeavors in general are viewed.Based on the 

experience of Students’ Science Association and real-life examples, different 

approaches, measures, agents, programs and initiatives for promoting of science, 

particularly OR, among youth are presented in this talk. 

2 - Organizing of Scientific and Educational Projects as a 

Way to Integrating Students into Scientific Community 

Alexej Orlov, Students Science Association, Institute for applied 

system analysis of National Technical University of Ukraine 

"KPI", Peremohy ave., 37, build. 1, room 299(15), 03056, Kyiv, 

Ukraine, orlov.alexej@gmail.com, Kateryna Pereverza 

In this report we will consider participation of young people in organization of 

scientific and educational projects as a way to integrating them into scientific 

community and motivating them to build academic careers. Usefulness of such 

approach would be demonstrated in example of projects of Students Science 

Association of NTUU KPI. Organizing team of SSA projects consist totally 

from KPI students. Their motivations to participate in project organizing, opportunities 

that open up for students and project management approach which 

is used at SSA are presented in this paper. 

3 - Activity of Student’s Organizations for Popularization of 

Science: Russian Experience 

Oleg Tumanov, Biology and soil, Kazan state university, 

Kremlevskaya st., 18, 420008, Kazan, Russian Federation, 

leick@inbox.ru 

Different student’s organizations and associations play a significant role in 

popularization of science especially among youth. Organization of scientific 

conferences and competitions by these institutions has a success in involving 

youth into science. A Russian experience of promoting science among 

youth will be presented. It will be illustrated with the examples of organizing 

projects aimed on science population and educational actions which have 

already demonstrated their effectiveness. 

4 - Leadership style and organizations’ strategic orientation 

Gabrijela Leskovar-Špacapan, Faculty of Economics and 

Business, University of Maribor, Razlagova 20, 2000, maribor, 

Slovenia, gabrijela.leskovar-spacapan@uni-mb.si 

The paper analyses whether strategic orientations of Slovenian organizations 

are supported by leadership style that influence employees’ perception of support 

for innovation. It is hypothesized that organizations with different strategic 

orientations exhibit different levels of employees’ perception of support for innovation 

as caused by differences in leadership style. Hypotheses were tested 

using variables measured through multiple items based on data from 195 organizations. 

The results confirm that leaders have significant impact on employees’ 

perception of support for innovation and thus on creativity and innovation.

� WB-48 


8.2.04 

Iterative Methods for Economic Models: 

Related Topics 

Stream: Iterative Methods for Economic Models 


Chair: Adriana Gnudi, Mathematics, Statistics, Computer science and 

Applications, University of Bergamo, Via dei Caniana, 2, 24127 , 

Bergamo, Italy, adriana.gnudi@unibg.it 

1 - Complex Dynamic Multi-level Networks 

Patrizia Daniele, Department of Mathematics and Computer 

Science, University of Catania, Viale A. Doria, 6, 95125, 

Catania, Italy, daniele@dmi.unict.it 

We shall consider a supply chain network model with three tiers of decisionmakers 

(manufacturers, retailers, and consumers) in the case when prices and 

shipments are evolving on time. Moreover, we assume that excesses of production 

and excesses of demand of the commodity are present. For such a 

framework we furnish, using the infinite dimensional duality theory, the equilibrium 

conditions for the representatives of each tier of the supernetwork, the 

time-dependent variational formulation governing the complete supply chain 

supernetwork, and we provide some existence theorems. 

EURO 24 - Lisbon 2010 WC-04 

Wednesday, 12:20 - 13:40 

� WC-02 


3.2.14 







1 - The Discretizable Molecular Distance Geometry Problem 

Nelson Maculan Filho, COPPE / PESC, Universidade Federal do 

Rio de Janeiro, Rio de Janeiro, RJ, Brazil, maculan@cos.ufrj.br 

The function of a protein is determined by its 3D structure. Nuclear Magnetic 

Resonance (NMR) experiments provide distances between some pairs of atoms 

of a protein. The Molecular Distance Geometry Problem (MDGP) consists in 

finding all the atomic positions of the molecule by exploiting the distances 

generated by the NMR experiments. In practice, the MDGP is solved by continuous 

optimization methods. We show that under a few realistic assumptions, 

the MDGP can be formulated as a search in a discrete space, where we call 

this MDGP subclass the Discretizable MDGP (DMDGP). We prove that the 

DMDGP is also NP-hard and we propose a Branch-and-Prune (BP). The BP 

algorithm performs remarkably well in practice in terms of speed and solution 

accuracy. We present computational results on several artificial and real-life 

instances. 

� WC-04 


3.2.13 

Facilities planning, design and management 






Chair: Eric Taillard, HEIG-Vd, Route de Cheseaux 1, Case Postale, 

1401, Yverdon-les-Bains, Vaud, eric.taillard@heig-vd.ch 

1 - An Estimation of Distribution Algorithm for Machine 

Part Cell Formation Problem 

Ibrahim Saber, Quantitative methods, FSEG, sfax, sfax, Tunisia, 

saber.ibrahim@gmail.com, Bassem Jarboui, Rebai Abdelwaheb 

We propose to apply a novel evolutionary algorithm called Estimation of Distribution 

Algorithm for the first time to the Part Cell Formation Problem. To improve 

solution, we added the Variable Neighborhood Search as a local search. 

Furthermore, we present two evaluation criteria called the Grouping Efficacy 

and the Percentage of Exceptional Elements to quantify the goodness of the 

obtained solutions. An extensive comparative study was elaborated with the 

existing literature. The obtained results have shown that the proposed EDA is 

very competitive against the previously best results. 

2 - A particle swarm optimization algorithm for a pick and 

place robotic system 

Slim Daoud, LOSI, universty of technology of Troys, 12 rue 

marie curie, 10000, troyes, France, slim.daoud@utt.fr, Farouk 

Yalaoui, Lionel Amodeo, Hicham Chehade, Thierry Girard 

In this work, we are interested in a robotic system which realizes pick and place 

operations. The robots must seize the products on a conveyor and deposit them 

on fixed points. This system was studied in 2000 by Mattone et al who developed 

different queuing strategies. To optimize the system, we have developed 

a particle swarm optimization algorithm which defines the suitable scheduling 

rules for each robot. The objective is to maximize the gripping rate. The obtained 

results are compared to an exact method and the results are promising. 

The developed method is applied on an industrial case. 

251

WC-05 EURO 24 - Lisbon 2010 

3 - Electromagnetic Meta-Heuristic Method for Solving 

Floorplanning Problem 

Seyyed Hassan Taheri, Mathematics, Khayyam University, 

Fallahi 1 - Shahrak Lashkar, 91775-1159, Mashhad, Iran, Islamic 

Republic Of, s_h_taheri@yahoo.com, Eric Taillard 

The floorplanning problem consists in finding the optimal positions for a given 

set of departments of fixed area within a facility such that the sum of distances 

between pairs of department that have a positive connection cost is minimized. 

This is a NP-hard combinatorial optimization problem. This paper applies 

the Electromagnetism-like Meta-Heuristic method with some local search to 

the floorpalnning problem. We provide results from several test problems that 

demonstrate the robustness of this approach across different problems and parameter 

setting. 

4 - Minimizing Working Time in the Straight Assembly Line 

with Deteriorating Tasks 

E. Maneshi, industrial engineering, iran university of science and 

technology, no. 2, 9.7 alley, 9 st, Sheikh-e-Bahai, Tehran, 

Tehran, Iran, Islamic Republic Of, ershad505@gmail.com, M.b. 

Aryanezhad, U. Bahalke, M. Karimi-Nasab, A.m. Yolmeh 

Assembly line scheduling is a category of assembly line balancing problems 

(ALBP), where the optimal schedule of the tasks and their assignments to different 

workstations should be determined simultaneously. Total working time 

of a product in the most industrial factories dealing with a set of deteriorating 

tasks is important because of its costs. So this paper introduces the effect of 

task deterioration in to the total working time of a product in a straight assembly 

line. Task deterioration means that a task processed later consumes more 

time than the same task when it is processed earlier. The following assumptions 

are considered in this research: - there are a fixed number of workstations in a 

simple straight assembly line, - there are a set of dependent tasks that should be 

assigned and scheduled to each workstation while considering the precedence 

relations between the tasks, - every task should be performed only one time 

in each cycle, - at each workstation, there is a multi-skilled machine able to 

perform all of the tasks assigned to that workstation, - there is a single product 

to be processed, - it is assumed that no machine breaks down when performing 

each task assigned to it. In other words, interruption is not allowed. On the 

other hand, pre-emption is not allowed, - setups are assumed to be negligible, - 

planned machine idleness is not allowed, but idleness could occur for the difference 

of the station time and the cycle time, - performance time of each task 

consists of two main parts: a fixed part, and a variable part which depends on 

the delay in the start time of the task and its deterioration coefficient, - production 

manager desires to obtain the optimal tasks’ schedule corresponding 

to minimum value of the total working time (because of tasks’ deterioration) 

while suggesting a virtual value for the worst cycle time. In other words, production 

manager does not accept tasks’ schedules and assignments with larger 

cycle time than what he proposes based on his previous experiences (as an upper 

bound for cycle time). 

Then the problem is formulated via a mathematical model. As the problem 

is strongly NP-Hard, a genetic algorithm is developed to solve the problem in 

large scales. Finally, several well-known test problems are solved for illustrating 

the proposed approach. 

� WC-05 


3.2.16 

Matheuristics 



Chair: Matteo Fischetti, DEI, University of Padova, Italy, 

matteo.fischetti@unipd.it 

Chair: Filipe Alvelos, Departamento de Produção e Sistemas, 

Universidade do Minho, Campus de Gualtar, 4710-057, Braga, 

Portugal, falvelos@dps.uminho.pt 

1 - A hybrid Multiple Particle Collision Algorithm applied to 

a source estimation inverse problem 

252 

Eduardo Luz, Applied Computing Grad. School, National 

Institute for Space Research, Av. Astronautas 1758, CAP/LAC, 

12227010, Sao Jose dos Campos, SP, Afghanistan, 

eduardofpl@gmail.com, José Carlos Becceneri, Haroldo F. 

Campos Velho 

This work presents the application of the Multiple Particle Collision Algorithm 

(MPCA), hybridized with the SIMPLEX algorithm, for estimating the location 

and the strength of a polluting source. The MPCA is a Metropolis based algorithm, 

broadly inspired in the scattering and absorption of particles while in nuclear 

reaction. The source estimation inverse problem is a non-linear optimization 

problem, where the objective function is given by the square difference 

between the pollutant concentration measured and the pollutant concentration 

computed from a Lagrangian particle dispersion model. 

2 - A Primal-Dual Local Search Heuristic for the Set Covering 

Problem 

Belma Yelbay, Sabanci University, Sabanci University Tuzla 

Campus, Istanbul, Turkey, byelbay@su.sabanciuniv.edu, S. Ilker 

Birbil, Kerem Bulbul 

We present a heuristic algorithm for solving the set covering problem effectively. 

It is based on the primal-dual approach which is commonly used for 

approximating NP-hard optimization problems. Unlike the traditional primaldual 

algorithms, our algorithm uses a dual variable selection method which 

performs well for cost and coverage correlated problems. We improve the solution 

quality through a local search procedure based on primal and dual variable 

fixing. We show that the proposed heuristic is able to produce good results in 

terms of both the solution quality and time. 

3 - A local search heuristic based on column generation 

Filipe Alvelos, Departamento de Produção e Sistemas, 

Universidade do Minho, Campus de Gualtar, 4710-057, Braga, 

Portugal, falvelos@dps.uminho.pt, Amaro de Sousa, Dorabella 

Santos, Carina Pimentel, Elsa Silva, J. M. Valério de Carvalho 

We propose a local search heuristic for integer programming models with a 

very large number of binary variables which has three main steps. In the first 

step, the problem is solved by column generation. In the second step, an initial 

solution is constructed by using primal and dual information from the restricted 

master problem. In the third step, based on the representation of a solution as 

a subset of columns, local search is applied. We illustrate the application of 

the heuristic in a multicommodity flow problem and in a two-dimensional bin 

packing problem. 

� WC-06 


8.2.30 

DEA Applications XI 



Chair: Mikulas Luptacik, Economics, Vienna University of 

Economics and Business Administration, Augasse 2-6, A-1090, 

Vienna, mikulas@chello.at 

1 - Changes in productivity of Spanish university libraries 

Clara Simon de Blas, Statistics & Operations Research, Rey Juan 

Carlos University, Departamental II, Desp. 251, C/Tulipan s/n, 

28933, Mostoles, Madrid, Spain, clara.simon@urjc.es, Jose 

Simon Martin, Alicia Arias 

We analyze productivity growth, technical progress, and efficiency change in 

a sample of 34 Spanish university libraries between 2003 and 2007. DEA and 

a Malmquist Index are combined with a bootstrap method to provide statistical 

inference estimators of individual productivity, technical progress, and efficiency 

change scores. To calculate productivity, a three-stage service model 

has been developed. The results indicate a growth in the internal productivity 

of the libraries and in the productivity of the service. 

2 - Initial Allocation of Emission Certificates Using Data 

Envelopment Analysis 

Mikulas Luptacik, Economics, Vienna University of Economics 

and Business Administration, Augasse 2-6, A-1090, Vienna, 

mikulas@chello.at 

In the paper we present a new approach for the initial allocation of the emission 

certificates. Using data envelopment analysis the amount of free allocated certificates 

to the plants is based on their eco-efficiency scores. We suppose that 

there is a centralized decision maker who supervises all the operating units. Applying 

and extending the centralized resource allocation model by Lozano-Villa 

(2004) the amount of free allocated certificates is based on their eco-efficiency 

such that the total amount of emissions is minimized.

3 - Academia and business world on bank performance: 

Are they on the same page? Developing a case for 

China 

Necmi Avkiran, UQ Business School, The University of 

Queensland, St Lucia Campus, 4072, Brisbane, Queensland, 

n.avkiran@uq.edu.au 

Majority of applications of DEA in banking do not test for the association of 

efficiency estimates with key performance indicators used by industry. As Chinese 

banks come under increasing scrutiny, identifying efficiency estimates’ associations 

with accepted financial measures of performance could guide benchmarking 

activities, pricing decisions and regulatory monitoring. Following a 

systematic test of super-efficiency models, super SBM emerges as the most 

significant model explaining the variation in the two industry ratios post-tax 

profit/average total assets, and return on average equity. 

4 - Evaluating mutual funds using robust nonparametric 

techniques 

Amparo Soler-Dominguez, Universitat Jaume I, 12071, 

Castellon, adomingu@cofin.uji.es, Juan Carlos Matallin-Saez, 

Emili Tortosa-Ausina 

Literature evaluating the performance of mutual funds using OR has evolved 

rapidly. The instruments applied (mostly DEA and FDH) have the ability of encompassing 

several dimensions of performance, but they have also some drawbacks 

that may have prevented a wider acceptance. In this article we apply 

not only the nonconvex counterpart of DEA, namely, FDH but also order- m 

and order- partial frontiers to a sample of Spanish mutual funds. The results 

obtained for both order-m and order- are quite useful, since a full ranking of 

performance is obtained. Although results hinge on the specified m and parameter. 

� WC-07 


8.2.47 

Machine Scheduling 

Stream: Project Management and Scheduling [c] 


Chair: Sinan Gürel, Middle East Technical University, METU 

Department of Industrial Engineering, 06531, Ankara, Turkey, 

sgurel@ie.metu.edu.tr 

1 - Single-machine scheduling with a general learning effect 

function 

Wen-Chiung Lee, Statistics, Feng-Chia University, 100 Wenhua 

Road, 407, Taichung, Taiwan, wclee@fcu.edu.tw 

Scheduling with learning effects has become a vivid area of research recently. 

However, it is assumed that the functions of the learning effects have specific 

forms in most of the existing models. In this paper, we propose a new model 

in which the actual job processing time is a general function of the processing 

time of jobs already processed and its scheduled position. It has the flexibility 

to describe different learning curves. Most of the models in the literature are 

special cases of our proposed model. We provide the optimal sequences for 

some single-machine problems. 

2 - Scheduling of jobs in a parallel machine problem with 

eligibility and release and queue times 

Manuel Mateo, Departament Business Administration, 

Universitat Politecnica Catalunya, Avda Diagonal, 647, 7th, 

E-08028, Barcelona, Spain, manel.mateo@upc.edu, Xavier 

Garriga, Imma Ribas 

The problem dealt is the scheduling of parallel machines with eligibility, i.e. 

not all the jobs can be manufactured in any machine. The previous and the 

next operations also introduce release times and queue times. The three-step 

algorithm proposed solves the problem of a set of n jobs to be scheduled on m 

parallel machines distributed among p levels, particularly three levels are studied. 

A machine can produce jobs of the same or a lower level. Any machine has 

the same processing time for a job. The objective is to find a feasible schedule 

with minimum completion time Cmax. 


3 - Finite-capacity-based due date setting: A computational 

study for flowshops 

Paz Perez Gonzalez, Industrial Management, University of 

Sevilla, Camino de los descubrimientos s/n, 41092, Sevilla, 

Spain, pazperez@esi.us.es, Jose M Framinan, Jose M. 

Molina-Pariente 

In this work we analyse a dynamic permutation flowshop shop floor, where 

new jobs arrive while old jobs are already scheduled with a committed due 

date, which must be considered as a deadline. The problem is to schedule the 

new jobs in order to determine the tightest possible due dates. In our paper 

we analyse the different problems that may arise depending on whether the old 

jobs can be rescheduled, or not. The distribution of solution of these problems 

are analyzed in order to study their structure and to help selecting the more 

suitable solution procedures to be applied. 

4 - Decompositions and conic formulations for railway design 

problems with convex congestions at yard locations 

Sinan Gürel, Middle East Technical University, METU 

Department of Industrial Engineering, 06531, Ankara, Turkey, 

sgurel@ie.metu.edu.tr 

Finding optimal yard and arc capacities are as critical as routing decisions in 

railway scheduling problems. Congestion realized at yard locations can be reduced 

by increased yard capacities which require high investment costs. Finding 

the yard locations at which the capacity should be increased simultaneously 

with routing decisions on railway network is a difficult problem. Nonlinear nature 

of congestion costs brings further difficulty in solving these problems. We 

propose using decomposition and conic formulation approaches for this problem 

and compare computational performance. 

� WC-08 


6.1.36 

Various Advances on Management and 

Scheduling I 



Chair: Erwin Pesch, FB 5 - Institute of Information Systems, 

University of Siegen, Hoelderlinstr. 3, 57068, Siegen, Germany, 

pesch@fb5.uni-siegen.de 




Chair: Edmund Burke, School of Computer Science & IT, University 

of Nottingham, Jubilee Campus, Wollaton Road, NG8 1BB, 

Nottingham, United Kingdom, ekb@cs.nott.ac.uk 

1 - Monthly tour scheduling models with mixed skills considering 

weekend off requirements 

Aiying Rong, Cemapre (Center for Applied Mathematics and 

Economics), ISEG-Technical university of Lisbon, Rua do 

Quelhas 6, 1200-781 , Lisbon, Portugal, arong@iseg.utl.pt 

This paper deals with the monthly tour scheduling problem with mixed skills 

considering the weekend off requirements, in contrast to the weekly planning 

horizon that is typical in most literature. Two model formulations are developed 

based on implicit programming techniques. One model uses a general 

integer programming formulation while the other one adopts a binary integer 

programming formulation. The effectiveness and efficiency of the two model 

formulations are illustrated and compared by the numerical tests based on realistic 

data sets. 

2 - Reducing schedule instability on the order level for the 

multi-item capacitated lot-size problem 

Andreas Cardeneo, Logistics Systems Engineering, FZI 

Forschungszentrum Informatik, Haid-und-Neu-Strasse 10-14, 

76131, Karlsruhe, Germany, cardeneo@fzi.de, Sebastian Fiedler 

253


Schedule stability is of increasing importance as it reduces dispatching time 

and harmonizes shop floor logistics. It has been subject of quite some publications 

on different types of MRP problems. Here, schedule stability for the 

MLCLSP in a multi-period rolling horizon setting, where demand is given by 

customer orders that change throughout the planning horizon, is addressed. 

A new penalty term measuring schedule instability and experimental results 

showing the impact of a selected set of environmental and design factors on 

schedule instability are presented. 

3 - Coordination of the branch closing times and scheduling 

the unloading operations in a parcel transfer center 

Kadir Ertogral, Industrial Engineering Department, TOBB 

University of Economics and Technology, Sogutozu cd. No:43, 

06560, Ankara, Turkey, kertogral@etu.edu.tr, Onur Dikmen 

Parcel transportation problems have received great deal of attention in the literature 

on different aspects, such as truck routing, network design, consolidation 

decisions, and operational planning in transfer centers. Our study deals with 

coordinating the branch closing times and scheduling the unloading operations 

of the trucks in a local transfer center of a national parcel carrier. The problem 

is modeled as an integer programming formulation. The formulation is 

based on a model for a parallel machine scheduling problem from the literature, 

with several modifications. Since the formulation of the problem is too 

big to solve to the optimality, we suggest a heuristic solution approach based 

on a linear programming relaxation of the model, and show the effectiveness of 

the heuristic approach through numerical experiments. 

� WC-09 


6.2.53 

Advanced Applications in Mathematical 

Programming 






Chair: Domingos Cardoso, Departamento de Matematica, 

Universidade de Aveiro, Campus Universitario de Santiago, 

3810-193, Aveiro, Portugal, dcardoso@ua.pt 

1 - Optimal Pricing, Marketing and Order Quantity for One 

Laptop Per Child 

Mostafa Ghasem Esfahani, Iran University of Science and 

Technology, Iran, Islamic Republic Of, 

mesfehani2002@yahoo.com, Seyed Jafar Sadjadi 

One of the primary concerns on One Laptop Per Child (OLPC) program is to 

find a suitable pricing strategy. The OLPC program provides various packages 

to penetrate different under developed countries. We present a mathematical 

model to determine the optimal price, marketing and ordering lot-size for this 

program. In this model, demand is considered as a function of price and marketing 

expenditure while the cost of production is a function of order size. 

The whole program is a kind of charity and the next goal is to maximize the 

total profit from the market. The resulted model is formulated as a Geometric 

Programming (GP). Therefore, we use the recent advances of optimization 

techniques called CVX to find the optimal solution. 

2 - An Upper Bound for Quadratic Multiple Knapsack Problem 

Tugba Saraç, Industrial Engineering Department, Eskisehir 

Osmangazi University, Meselik Kampüsü M3, 26480, Eskisehir, 

Turkey, tsarac@ogu.edu.tr 

The Quadratic Multiple Knapsack Problem (QMKP) with k knapsacks, each 

with its own capacity ck, asks to maximize a quadratic objective function subject 

to k+1 inequality constraints. Finding an upper bound is important for 

this kind of problems because in most cases, inexact methods are used to solve 

them and upper bounds help to know what the quality of the obtained solution 

is. Upper bounds are also used in branch-and-bound algorithm. In this study, 

an upper bound is proposed for QMKP and its performance is evaluated by 

using test instances taken from the literature. 

254 

3 - An Entropy-based Solver for Multidimensional Nonlinear 

Knapsack Problems 

Yuji Nakagawa, Department of Informatics, Kansai University, 

2-1-1 Ryouzenji-Cho, 569-1095, Takatsuki, Japan, 

nakagawa@res.kutc.kansai-u.ac.jp, Ross J. W. James, César 

Rego, Chanaka Edirisinghe 

We develop an implicit enumeration method for solving difficult linear and nonlinear 

multidimensional knapsack problems where branching is accomplished 

based on the sub-problem complexity. Using the concept of entropy in information 

theory, we develop a (sub)problem-difficulty metric that is used to devise 

decision rules for problem partitioning within implicit enumeration. Comparisons 

with state-of-the-art solvers show that our method is extremely efficient. 

4 - A mathematical approach to seek the natural and practical 

piano fingering 

Keisuke Hotta, Faculty of Information and Communication, 

Bunkyo University, 1100 Namegaya, 253-8550, Chigasaki, 

Kanagawa, Japan, khotta@shonan.bunkyo.ac.jp 

Given a piano score, the piano fingering problem is to decide the natural and 

reasonable fingering for the piano performance. The problem is not only the 

allocation of fingers to notes in score, but also some kind of melody analysis 

of music. Considering several costs based on that, I modeled the problem as 

0,1-IP, and tried to seek the practical fingering. The determination of the fingering 

is based on a lot of rules, which several piano players and researchers 

have mentioned. However, some of them seem like a contradiction. I also seek 

the useful and helpful rules for the fingering. 

� WC-11 


8.2.38 


Technology III 









1 - Innovative Training Methodologies for Transport Trainers 

Ivan Kolarov, Machine Dsign, HST, Geo Milev Str. N 158, 1574, 

Sofia, Bulgaria, ikolarov@vtu.bg 

An investigation of the main ICT application for continuing training of drivers 

and transport specialists in accordance with Directive 2003/59/EC is presented. 

The best practices of innovative methods for training used by transport specialists 

in 6 European countries (ES, EL, IT, LT, UK and BG) are analyzed and the 

needs of trainers in the sector are determined. Perspective training methodologies 

with application for creation of Virtual Learning Community for the needs 

of sector are specified. Results of this investigation will be used for creation of 

a tool for transport trainers. 

2 - Residual intelligibility estimation method on a safety 

area border 

Elena Arkhypova, Institute of Physics and Technology, National 

Technical University of Ukraine ’Kyiv Polytechnic Institute’, 37 

Prospect Peremogy, Kiev 03056, Ukraine, str.Malyshko 

21-b,ap.77 Kiev 02192, Ukraine, Kiev, Ukraine, leoo@zeos.net, 

Vladimir Zhuravlev 

It is realized analytical and experimental researches which solve the task that 

consists in an objective estimation of one-dimensional speech masking efficiency 

parameter on a safety area border. This parameter estimation method 

is developed on the basis of the correlation coefficient statistical characteristics 

analysis. Articulation words tables are developed for the first time with regard 

to statistical parameters of the Ukrainian language phonetic structure. Articulation 

tests which confirm adequacy of the offered method are carried out.

3 - Optimization of time-triggered communication protocols 

by means of Resource Constrained Project 

Scheduling with Temporal Constraints 

Zdenek Hanzalek, DCE, CTU, Karlovo nam 13, 121 35, Prague, 

Czech Republic, zdenek.hanzalek@fel.cvut.cz, Premysl Sucha 

The objective is to choose schedule of the communication protocols and to minimize 

the makespan of time-triggered messages in order to maximize the space 

for other messages. Messages are characterized by release dates, deadlines, 

end-to-end deadlines, synchronisation and precedence relations. The resources 

can be characterized as one processor (single-channel Flexray) or dedicated 

processors (Profinet) or identical processors (two-channel TTP). In the case of 

wireless communications the message transmission occupies several collision 

domains and therefore it is a multiprocessor task. 

4 - Evaluate Outsourcing Design Systems Performance 

based on GRA 

Cheng-Ru Wu, Yuanpei University, Taiwan, 

alexru00@ms41.hinet.net, Chiu-Chin Chen, Che-Wei Chang 

This study focused on using the grey relational analysis (GRA) decision support 

tools for the Taiwan bureau’s control information systems (CIS) analysis 

the outsourcing design information systems performance case studies. GRA in 

the combination-based approach, this article provides decision-makers are doing 

outsourcing design, to more practical and accurate in line with the standard 

structure of domestic industries to enhance the control of outsourcing design 

the overall effectiveness of information systems. 

� WC-12 


8.2.39 

AHP 07 


Processes 


Chair: Gulgun Kayakutlu, Industrial Engineering, Istanbul Technical 

University, Macka, 34367, Istanbul, Turkey, gkayakutlu@gmail.com 

1 - Applying the Delphi and AHP for Selecting the Optimal 

International Exhibition Agency 

Pi-Fang Hsu, Department of Communications Management, 

Shih Hsin University, No.1, Lane 17, Mu-Cha Road, Sec.1, 

11604, Taipei, Taiwan, celina9@ms26.hinet.net, Hsin-Yu Chiang 

This study develops a model for selecting international exhibition agency. First, 

the proposed model adopts the modified Delphi method to identify suitable criteria 

for evaluating exhibition agency. Next, the research model applies the 

analytic hierarchy process (AHP) to calculate and rank the optimal exhibition 

agency. Additionally, the example of a renowned Taiwanese bedding company 

is used to demonstrate the process of exhibition agency selection using this 

model. The proposed model helps enterprises effectively select media agency, 

making it highly applicable in academia and commerce. 

2 - Influential Causes of Bullwhip Effect in Automotive 

Supply Chain 

Fatih Tas, Industrial and Mechanical Engineering, Istanbul 

Technical University, Pasalimani Cad. Selehattin bey apt., 34/6 

Uskudar, 34472, Istanbul, Turkey, tasfatih88@hotmail.com, 

Gulgun Kayakutlu 

Bullwhip effect is an important obstacle for supply chain success. Researchers 

are focused on six causes of bullwhip effect: demand fluctuations, order batching, 

shortage gaming, price fluctuation, lead time, and level of echelons in 

the supply chain. This study aims to determine the sequence of importance 

of these causes in the automotive industry using Analytic Hierarchy Process 

(AHP) method. Supply chain managers of both spare parts chain and finished 

products chain are interviewed. The level of echelons is found to be the most 

important cause in this industry unlike other industries. 

3 - Correction Approach for Solving the Pairwise Comparison 

Matrix Inconsistency 

Dmitry Borodin, Business Information and ICT, University 

College of Ghent, Schoonmeersstraat 52, 9000, Ghent, Belgium, 

dmitriy.borodin@hogent.be, Viktor Gorelik 


The work proposes a correction approach to solve the problem of the inconsistency 

of the pairwise comparison matrix. Two correction techniques forcing 

such matrices to the consistency are demonstrated together with the results of 

computations and their comparison with other different approaches. Pairwise 

comparisons are central in mathematics for the measurement of intangible factors, 

as proven in the AHP/ANP theory (T. Saaty). To recover the scale for 

decision making, it is necessary to solve the respective eigenvalue problem, 

which brings us to the pairwise comparison inconsistency. 

4 - Interactive Comparison Support with Comparison Pattern 

Search for Analytic Hierarchy Process 

Yumi Tadano, Graduate School of Information Science and 

Technology, Hokkaido University, North 14, West 9, Kita-ku, 

060-0014, Sapporo, Hokkaido, Japan, 

yumi-hr@complex.eng.hokudai.ac.jp, Hidenori Kawamura, Keiji 

Suzuki 

The analytic hierarchy process (AHP) is a method for decision making. The 

AHP calculate overall evaluations according to structure a hierarchy of the 

problem and compare two elements of the hierarchy’s each level. Therefore, 

the comparisons of all pairs are difficult when evaluating many alternatives. We 

propose a comparison support method for evaluating many alternatives when 

decision maker needs to decide the highest priority alternative. The comparison 

support method stops pairwise comparisons when the highest priority alternative 

is found even if all comparisons aren’t compared. 

� WC-13 


2.2.21 

Location and Network Design 



Chair: Ivana Ljubic, Department of Statistics and Decision Support 


Austria, ivana.ljubic@univie.ac.at 

1 - Lagrangean Decomposition for an Adaptive Location- 

Distribution Problem 

Bernard Gendron, DIRO/CIRRELT, Université de Montréal, C.P. 

6128, succ. Centre-ville, H3C 3J7, Montréal, Québec, Canada, 

gendron@iro.umontreal.ca, Paul-Virak Khuong, Frédéric Semet 

We consider a location problem motivated by a case study for a multi-channel 

retailing company, which sells a wide variety of products via Internet, mail order 

catalogs, and stores. Since most items to deliver are small or medium-size 

parcels, consolidation is a major concern which is addressed by designing a 

multi-echelon distribution system. This system is adaptive, in the sense that 

terminals and satellites can be opened or closed easily according to demand 

fluctuations. We introduce a Lagrangean decomposition approach embedded 

in a branch-and-bound scheme, which can deliver provably optimal solutions 

to large-scale instances of the problem. 

2 - The Hop-Constrained Connected Facility Location 

Problem 

Stefan Gollowitzer, Department of Statistics and Decision 

Support Systems, University of Vienna, Bruennerstraße 72, 1210, 

Vienna, Austria, stefan.gollowitzer@univie.ac.at, Ivana Ljubic 

The Connected Facility Location (ConFL) problem models the Fiber-To-The- 

Curb strategy for broadband local access networks. The problem generalizes 

the Uncapacitated Facility Location and the Steiner tree problem. We consider 

a variant of ConFL, in which the number of edges between a predefined root 

and each open facility is limited. We provide a theoretical and computational 

comparison of MIP models for this problem. Finally, we show how to model 

the problem on layered graphs. Our computational study shows the computational 

advantage of the latter approach over classical MIP models. 

255


� WC-14 


2.2.15 

Stochastic Methods in Finance and 

Economics 



Chair: Ricardo Josa-Fombellida, Estadística e Investigación 

Operativa, Universidad de Valladolid, Facultad de Ciencias, Paseo 

Prado de la Magdalena, s/n, 47005, Valladolid, Spain, 

ricar@eio.uva.es 

Chair: Juan Pablo Rincon-Zapatero, Economia, Universidad Carlos 

III de Madrid, c/ Madrid, 126, 28903, Getafe, Madrid, Spain, 

jrincon@eco.uc3m.es 

1 - Consumption, portfolio and life insurance rules for 

time-inconsistent decision makers 

Jesús Marín-Solano, Matemàtica Econòmica, Financera i 

Actuarial, Universitat de Barcelona, Av. Diagonal, 690, 

Barcelona, Spain, jmarin@ub.edu, Jorge Navas, Oriol Roch 

The assumption of a constant discount rate of time preference has been questioned 

by empirical findings about how agents change their preferences over 

time. However, the use of standard optimal control techniques for problems 

with other discount factors apart from the exponential function with a constant 

rate can lead to time inconsistent solutions. In this paper we solve a problem 

of consumption, portfolio and life insurance rules for a decision maker with a 

non-constant discount rate of time preference. 

2 - Effect of Term Structure of Futures Price on Spot Procurement 

Policies 

Ankur Goel, Operations, Case Western Reserve University, 

10900 Euclid Avenue, Peter B Lewis Building, 44106, 

Cleveland, OH, ankur.goel2@case.edu, Genaro Gutierrez 

We characterize optimal procurement policy of commodity from the spot market 

under the paradigm of fluctuating spot prices and stochastic demand. In 

particular, we explore the effect of additional term structure information of futures 

prices on the procurement policies from the spot market. In this regards, 

we compare one-factor stochastic price model with a two-factor model. In addition, 

we also explore the benefits of frequent calibration of stochastic price 

process on the procurement cost structure of a firm. 

3 - Risk-Sensitive Discounted and Average Criteria in Finite 

State Markov Decision Chains 

Karel Sladky, Department of Econometrics, Institute of 

Information Theory and Automation, Academy of Sciences of 

the Czech Republic, Pod Vodarenskou vezi 4, 18208, Prague 8, 

Czech Republic, sladky@utia.cas.cz 

The usual optimization criteria for Markov decision chains as total discounted 

or average costs cannot reflect variability-risk features of the problem. To this 

end, we focus attention on policies yielding minimal risk-sensitive costs, i.e., if 

the stream of discounted or undiscounted one-stage costs is evaluated, instead 

of linear by an exponential utility function. Necessary and sufficient optimality 

conditions for discounted and average risk-sensitive criteria as well as conditions 

guaranteeing independence of average costs on starting state will be 

discussed. 

4 - Replacement Investment under Tax Uncertainty 

256 

Joao Zambujal-Oliveira, Engineering and Management, 

Technical University of Lisbon, Av. Rovisco Pais„ Lisboa, 

1049-001, Lisbon, Portugal, j.zambujal.oliveira@ist.utl.pt 

This paper examines the asset replacement problem to investigate the optimal 

level under an uncertain tax environment and considering a depreciation policy. 

Based on the concept of equivalent annual cost, a minimization model, applied 

to the real options paradigm, allows innovative evaluations for flexibility in 

the replacement process analysis. This model improves previous ones since 

it considers a semi-autonomous process for salvage value. Assembled over a 

partial differential equation framework, the model integrates several processes 

(GBM, mean reversion and jump process), which are integrated in a cost function 

that supports replacement decisions under varying tax environment. The 

general analytical and particular numerical solutions differ significantly from 

those observed in previous models, providing evidences of over valuated levels 

of replacement and corroborating that different types of uncertainties can produce 

non-monotonous effects on the optimal replacement level. The outcome 

is a new and stronger approach to the EAC literature, supplying an algorithm 

conditioned by a variable salvage value and changes on tax regime. 

� WC-15 


2.2.12 

Arc Routing Problems 

Stream: Vehicle Routing [c] 


Chair: Luis Gouveia, DEIO, University of Lisbon, Campo Grande, 

Bloco C6, 1749-016, Lisbon, Portugal, legouveia@fc.ul.pt 

1 - Rural Postman and related Arc Routing Problems 

Ana Maria Rodrigues, INESC Porto/ ISCAP-IPP, INESC Porto, 

Campus da FEUP RuaDr. Roberto Frias, n 378, 4200-465, Porto, 

Portugal, amr@inescporto.pt, José Soeiro Ferreira 

The presentation addresses the Rural Postman Problem (RPP) and related Arc 

Routing Problems (ARPs). The RPP (NP-Hard problem) has the objective of 

determining a minimum cost circuit traversing the required edges at least once 

(not all the edges of the graph are required). Applications appear in many contexts: 

mail delivery, garbage collection, street cleaning and cutting problems. 

The idea is to characterize, illustrate links and hierarchies and propose an organization 

to facilitate the understanding and integrate all the related ARPs. A 

reference to the solutions methods will also be made. 

2 - Models for the Sectoring-Arc Routing Problem (SARP) 

Ana Catarina Nunes, Dep. Métodos Quantitativos, ISCTE - IUL 

/ Centro IO, Portugal, Av. das Forças Armadas, 1649-026, 

Lisbon, Portugal, catarina.nunes@iscte.pt, Cândida Mourão 

The Sectoring-Arc Routing Problem (SARP) models the activities associated 

with the streets of large urban areas, such as waste collection. The SARP is 

defined over a mixed graph. Its aim is to identify a given number of similar 

sectors (sub-graphs) and to build a set of collecting trips in each sector, such 

that the total duration of the trips is minimized. Each sector is collected by one 

vehicle and each of its trips can not exceed a given load. Linear mixed integer 

programming formulations are presented for the SARP, and computational 

results over a set of benchmark problems are reported. 

3 - Bounds for the Mixed Capacitated Arc Routing Problem 

Cândida Mourão, Dep. Matemática, Instituto Superior de 

Economia e Gestão / Centro IO, Rua do Quelhas, 6, Gabinete 

203, 1200-781, Lisboa, Portugal, cmourao@iseg.utl.pt, Luis 

Gouveia, Leonor S.Pinto 

Mixed Capacitated Arc Routing (MCARP) models are widely used in distribution 

or collection problems where vehicles with limited capacity perform 

certain activities that are continuously distributed along some pre-defined links 

(routes, streets) of an associated mixed network. We present methods to obtain 

bounds for the problem. The quality of the methods is tested on some benchmark 

instances. We discuss and compare it with the best known method from 

the MCARP literature used for medium and large sized instances. 

4 - A Lower Bound and a Hybrid ILS-VND for a Location 

Arc Routing Problem 

Labadi Nacima, ROSAS, University of Technology of Troyes, 

12, rue Marie Curie, BP2060, 10010, Troyes, France, 

nacima.labadi@utt.fr, Jan Melechovsky

This note deals with a variant of the Capacitated Location Arc Routing Problem 

(CLARP). The problem is defined on a two level undirected weighted graph, 

with plants, potential depots and customers (edges with positive demands). The 

goal is to determine the depots to open and for each depot, the plants from 

whose it would be provided to service the edges assigned to it; in order to minimize 

the total cost. An integer linear programming is developed and its linear 

relaxation is used to derive a valid lower bound. To solve the problem, A hybrid 

ILS VND is also designed. 

� WC-16 


2.2.14 

Improving Real-time Railway Operations 

Stream: Public Transport [c] 


Chair: Dennis Huisman, Econometric Institute, Erasmus University, 

Rotterdam, Netherlands, huisman@ese.eur.nl 

1 - Robust Train Scheduling based on the UK Network 

Banafsheh Khosravi, School of Management, University of 

Southampton, Southampton, United Kingdom, 

B.Khosravi@soton.ac.uk, Julia Bennell, Chris Potts 

We consider the computationally hard problem of scheduling trains for large or 

complicated networks. The problem can be formulated as a job shop scheduling 

problem subject to a set of operational and safety constraints. Our proposed 

model makes use of a modified disjunctive graph. The objective is to minimize 

the total weighted tardiness to avoid delay propagation in a passenger network 

which is based on the railway system in the UK. We develop a solution method 

inspired by the successful results of local search procedures to solve job shop 

scheduling problems. 

2 - Dispatching and coordination in real-time railway traffic 

management 

Francesco Corman, Department of Transport and Planning, Delft 

University of Technology, -, Delft, Netherlands, 

f.corman@tudelft.nl, Andrea D’Ariano, Dario Pacciarelli, Marco 

Pranzo 

Train conflict resolution in multiple dispatching areas either requires to solve 

a large centralized problem or to solve and coordinate several smaller problems. 

This paper compares centralized and distributed approaches for dispatching 

trains in a large Dutch railway network. Computational experiments show 

that centralized control outperforms the distributed one when the timetable provides 

good guidance for dispatching. On the other hand, distributed control is 

the most effective approach when severe disruptions require deep modifications 

to the timetable in order to recover feasibility. 

3 - Standardisation of service intention and interfaces for 

automatic train scheduling 

Sabrina Wiedersheim, D-Math, ETH Zurich, Institute for 

Operation Research, ETH Zurich, HG G21.2, Raemistr. 101, 

8092, Zurich, Zurich, Switzerland, wsabrina@ethz.ch, Gabrio 

Curzio Caimi, Leo Kroon, Marco Laumanns, Dick Middelkoop 

In the last years, many train scheduling models and algorithms have been developed. 

However, the specification of input data and the interfaces between 

different scheduling steps often differ. In particular, the description of the commercial 

requirements of a timetable, the so-called service intention, varies from 

approach to approach. We present how to resolve this discrepancy by introducing 

standards for the specification of input and output data. This way, collaboration 

between different countries and research groups becomes easier and 

comparison between different methods is more meaningful. 

4 - Delay Management with Passenger Re-Routing 

Twan Dollevoet, Econometric Institute, Erasmus University of 

Rotterdam, Burgemeester Oudlaan 50, P.O. Box 1738, 3000 DR, 

Rotterdam, Netherlands, dollevoet@ese.eur.nl, Dennis Huisman 

Delay management determines which connections should be maintained in case 

of a delayed feeder train. Current delay management models assume that the 

delay for a passenger who misses a connection equals the cycle time of the 

timetable. In reality, passengers will adjust their route when a connection is 

dropped. We propose a model that takes passenger re-routing into account explicitly. 

We have developed several methods that solve the resulting integer 

programs in a short amount of time. The quality of these methods is compared 

using real-life instances from Netherlands Railways. 


� WC-17 


1.3.14 

Regression and Its Application 



Chair: Pakize Taylan, Mathematics, Dicle University, 21280, 

Diyarbakır, Diyarbakir, Turkey, ptaylan@dicle.edu.tr 

1 - Time Series Forecasting Using Fuzzy Time Series 

Approach, Neural Network Models and Regression 

Splines 

Akhlitdin Nizamitdinov, Statistics, Anadolu University, Anadolu 

University Faculty of Science, Department of Statistics, 

Eskisehir, Turkey, ahlidin@gmail.com, Memmedaga Memmedli, 

Ozer Ozdemir 

Neural network models, regression splines and fuzzy time series approach have 

important affects to improve forecasting of time series. Hence, in this study, we 

aim to improve time series forecasting by using a fuzzy time series approach, 

regression splines such as P-Spline and Cubic Smoothing Splines and neural 

network models such as Generalized Regression Neural Networks (GRNN), 

Linear Network, Multi Layer Perceptron (MLP) and Radial Basis Function 

(RBF). All results are shown by using them for daily, weekly and monthly 

close prices of Istanbul Stock Exchange (ISE) national-100 index. 

2 - Support Vector Machine for Time Series Regression 

Ange-Michel Poli, Université de Corse, 20250, Corte, France, 

poli.ange-michel@wanadoo.fr 

Support Vector Machines (SVMs) have been extensively used in classification 

and regression. In this talk we will show how SVMs can be used to predict 

spcific aggregated values from a time series. Applications in finance will also 

be discussed. 

� WC-18 


1.3.15 

Stochastic Models and Optimization 






1 - Design and Planning of Green Supply Chains: A Fuzzy 

Approach 

Tânia Pinto_Varela, UMOSE-LNEG, Estrada do Paço do 

Lumiar, 1649-038, Lisbon, Portugal, Tania.pinto@ineti.pt, Ana 

Paula Barbósa-Póvoa, Augusto Novais 

Green Supply Chains can be seen as logistic structures that guarantee production 

and distribution of products in an environmental friendly manner. To pursue 

this goal companies must invest on the optimal design/planning of their 

logistic structures, while considering the trade-off between profit and environmental 

impacts. This is addressed here using a generic and uniform mathematical 

framework (Resource-Task-Network). For this bi-level optimization a 

symmetric fuzzy linear programming approach is used, where those objectives 

are replaced by a new one embodying a compromise between them. 

2 - On the Structural Analysis of Inventory Problems with 

Inventory-dependent Demand 

Frank Y Chen, Department of Systems Engineering & 

Engineering Management, Chinese University of Hong Kong, 

Hong Kong, China, yhchen@se.cuhk.edu.hk 

We consider how inventory should be managed when inventory affects sales in 

a periodic review setting. Models to be analyzed include those with a general 

demand function (by which demand varies with the stocking or displayed quantity)and 

a general cost structure. Capitalizing on our prior experience gained 

in related inventory-pricing control problems, we are able to stablish structural 

results for these models. 

257


� WC-19 



Large-scale Mixed Optimization Problems 



Chair: F.-Javier Heredia, Statistics and Operations Research, 

Universitat Politècnica de Catalunya, Jordi Girona 1-3, Campus Nord, 

C5, 08034, Barcelona, Catalunya, Spain, f.javier.heredia@upc.edu 

1 - Solving tough instances of the controlled tabular adjustment 

problem 

José Antonio González Alastrué, Dept. Estadística e 

Investigación Operativa, UPC, Edificio C5, Campus Nord, Jordi 

Girona 1-3, 08034, Barcelona, Spain, 

Jose.A.Gonzalez@upc.edu, Jordi Castro 

Controlled tabular adjustment (CTA) is an efficient method for disclosure control 

of tabular data released by national statistical agencies (NSA). CTA results 

in mixed integer linear problems which are difficult to be solved by generic 

solvers; in some instances, it is even difficult to find a feasible solution. In this 

work we present several techniques (e.g., block coordinate descent, SAT procedures 

for initial feasible points, etc) that may drive a difficult instance towards 

a quality solution within a reasonable time limit (required by NSAs in the real 

world). 

2 - Perspective cuts for solving the optimal electricity market 

bid problem with bilateral contracts 

Eugenio Mijangos, Applied Mathematics and Statistics and 

Operations Research, UPV/EHU, P.O. Box 644 – Dept. 

Matematica Aplicada y E.I.O. (UPV/EHU), 48080, Bilbao, 

Spain, eugenio.mijangos@ehu.es, F.-Javier Heredia 

The electric market regulation in Spain (MIBEL) establishes the rules for bilateral 

contracts in the day-ahead optimal bid problem. Our model allows a 

price-taker generation company to decide the unit commitment of the thermal 

units, the economic dispatch of the bilateral contracts and the optimal sale bids 

observing the MIBEL. The uncertainty of the spot prices is represented through 

scenario sets. We solve this model as a deterministic MIQP problem by using 

perspective cuts to improve the performance of Branch and Cut approach. Numerical 

results are reported. 

3 - On the use of Probabilistic Algorithms in Nonsmooth 

Optimization 

Angel A. Juan, Computer Science, Open University of Catalonia, 

Rambla Poblenou, 156, 08018, Barcelona, Spain, 

ajuanp@gmail.com, Bernardo Ruiz, Helena Ramalhinho 

Lourenço, Javier Faulin, Carlos Mendez 

This presentation discusses the use of probabilistic or randomized algorithms 

for solving nonsmooth optimization problems. We then propose the use of 

probability distributions, such as the Geometric or the Triangular ones, to add 

a biased random behavior to classical heuristics such as the Clarke and Wright 

Savings heuristic for the Vehicle Routing Problem or the NEH heuristic for the 

Flow Shop Scheduling Problem. By randomizing these heuristics, a large set of 

alternative good solutions can be quickly obtained in a natural way. Some specific 

examples of this technique are analyzed to illustrate the main ideas behind 

this approach. 

4 - Endogenous model for medium-term electricity generation 

planning in liberalized mixed markets 

Laura Marí, Statistics and Operations Research, Univ. 

Politecnica de Catalunya, C Jordi Girona 1-3, Campus Nord - Ed 

C5, Dptx 201, 08034, Barcelona, Spain, laura.mari@upc.edu, 

Narcis Nabona 

Mixed electricity markets have pool auction and bilateral contracts. In the 

medium term stochastic parameters and the probabilistic matching of load have 

to be modeled. The maximization of revenue by participant generation companies 

leads to the minimization of the difference of positive definite quadratics 

subject to many linear inequality and equality constraints. The equilibrium situation 

of the market is obtained by using the Nikaido-Isoda relaxation algorithm 

of successive optimizations, where reverse convex constraints are employed to 

deal with non convex objective function. 

258 

� WC-20 


1.3.33A 

Discrete and Global Optimization 

Stream: Discrete and Global Optimization 





Chair: Valentin Weber, G-SCOP, 46 avenue Félix Viallet, 38000, 

GRENOBLE, France, valentin.weber@g-scop.grenoble-inp.fr 

1 - Robust Quadratic Assignment Problem and a Heuristic 

to Solve It 

Mohammad Javad Feyzollahi, Industrial Engineering dept., 

Sharif University of Technology, Industrial Eng. Dept., Sharif 

University of Technology, Azadi Ave.,Tehran, Iran, 11365-8639, 

Tehran, Tehran, Iran, Islamic Republic Of, 

feyzollahi@gmail.com, Maryam Ghotbaddini, Mohammad 

Modarres 

The Quadratic Assignment Problem models many real-world problems. The 

TSP, maximal clique and graph partitioning can be formulated as a QAP. It 

is obvious that in a real-world problem the exact amount of workflow among 

facilities, traffic in computing process and data transferring and distances between 

cities in the TSP model are not known precisely. We use a robust discrete 

optimization method to address data uncertainty in QAP, present a heuristic 

method to solve it and analyze the result by simulation. 

2 - Optimal adjust of continuous and discrete variables in 

non-linear programming problems 

Edilaine Soler, Escola de Engenharia de São Carlos, 

Universidade de São Paulo, Av. Trabalhador São-Carlense, 400, 

Centro, 13560-970, São Carlos, SP, Brazil, edilaine@sc.usp.br, 

Geraldo R. M. da Costa 

In this work a method is proposed for solving non-linear programming problems 

containing continuous and discrete variables. A penalty function, which 

penalizes the objective function when discrete variables assume non-discrete 

values, is presented. By including this penalty function into the objective function, 

a non-linear programming problem with only continuous variables is obtained 

and the solution of this problem is equivalent to the solution of the initial 

problem that contains discrete and continuous variables. The proposed algorithm 

was applied to the Optimal Power Flow problem. 

3 - A method for approximating general pairwise comparison 

matrices by consistent matrices 

Janos Fulop, Research Group of Operations Research and 

Decision Systems, Computer and Automation Research Institute, 

Hungarian Academy of Sciences, Kende u. 13-17, H-1111, 

Budapest, Hungary, fulop@sztaki.hu 

In multiattribute decision making, pairwise comparison (PC) matrices are applied 

to derive implicit weights for a given set of decision alternatives. A class 

of the approaches is based on the approximation of the PC matrix by a consistent 

matrix. We consider PC matrices without the reciprocity condition, and 

approximate them by consistent matrices in the least-squares sense. We transform 

the problem into the form of separable programming, and give sufficient 

conditions of the convexity of the objective function over the feasible set. For 

the general case, we propose a branch-and-bound method. 

4 - Challenging instances for discrete optimization problems 

Valentin Weber, G-SCOP, 46 avenue Félix Viallet, 38000, 

GRENOBLE, France, valentin.weber@g-scop.grenoble-inp.fr, 

Nadia Brauner, Yann Kieffer 

Difficulty in complexity theory reflects worst case performances. However the 

practice of benchmarking for NP-hard problem is usually based on libraries 

of sample instances or random generators. We can question the relevance of 

such instances, in particular, whether they illustrate the actual difficulty of the 

problem. In our study, we look into some of these sample instances, analyze 

their hardness and suggest more challenging instances. This leads us to define 

several criteria to approach the informal concept of instance difficulty.

5 - Effective Branching Rules for Linear Integer Programming 

Models 

Elias Munapo, Decision Sciences, UNISA (SOUTH AFRICA), 

Preller Street, Pretoria, P. O. Box 392, UNISA 0003, 27, 

Pretoria, Gauteng, South Africa, munape@unisa.ac.za 

The paper presents an effective branching on constraint technique for linear integer 

programming models. In this technique a combination of two or more 

basic variables are used in branching rather than the usual single variable that 

is common in the available branch and bound related schemes. The proposed 

procedure has the advantage that it accelerates convergence if used within the 

context of the branch and cut algorithm. Searches on massive sub-problems 

that are usually associated with integer problems are minimized. 

� WC-21 


6.2.47 

Optimization Algorithms III 






1 - A Hierarchical Segmentation Algorithm for Image Processing 

Edwin Zarrazola, Estadística e Investigación Operativa, 

Universidad Complutense de Madrid, Facultad de Ciencias 

Matemáticas, Plaza de Ciencias 3, 28040, Madrid, 

ezarrazo@mat.ucm.es, Daniel Gomez Gonzalez, Javier Montero, 

Javier Yannez 

Segmentation algorithms are well known in the field of image processing. In 

this work we propose an efficient and polynomial algorithm for image segmentation 

based on the coloring of an associated network to the digital image. 

The main difference with the classical segmentation algorithms is in the output 

given by the segmentation process. Since the classical output for segmentation 

algorithms give us the homogeneous regions in the image, our proposal is to 

produce an hierarchical information (in a similar way as a dendrogram does 

in classical clustering methods) of how the groups are formed in the image, 

from the initial situation in which all pixels are in the same group to the final 

situation in which the whole image in divided in the minimal information units. 

2 - Efficient direct-search solvers for single and multiobjective 

derivative-free optimization 

Ana Luisa Custodio, Dept. Mathematics, New University of 

Lisbon, Quinta da Torre, 2829-516, Caparica, Portugal, 

alcustodio@fct.unl.pt, Jose Aguilar Madeira, Humberto Rocha, 

A. Ismael F. Vaz, Luís Nunes Vicente 

Many practical optimization problems involve expensive functions for which 

derivatives are unavailable or unreliable. 

We describe two solvers for such type of problems, based on direct-search 

methods organized around a search and a poll step. The first solver (sid-psm) 

applies to the single objective case and incorporates model-based techniques in 

both steps. The second solver (dms) is designed to compute the Pareto front in 

the multiobjective case without any function aggregation. 

Extensive numerical results show that both solvers are efficient. 

3 - An agent-based model of an eco-product market with 

social interactions and dynamic game pricing schemes 

Edward Thommes, Univesrity of Guelph, N1G 2WQ1, Guelph, 

Ontario, ethommes@uoguelph.ca 

We present an AB model of an eco-product market from a system design perspective, 

to investigate ways in which such a market can be made to emerge and 

develop. The model extends a static formulation of differentiated product markets 

to include social interactions among consumer classes. We study changes 

in response to influences such as new product introduction. Analysis of the 

model is conducted with multiple "personalities" of consumers. We also consider 

a dynamic game analysis perspective for pricing schemes of eco-products 

on markets simulated as above. 


4 - Optimization methods and software in study of billiard 

aerodynamics and reverse Magnus effect in free molecular 

flow 

Tatiana Tchemisova, Departmento of Mathematics, University of 

Aveiro, Campus Universitario de Santiago, 3810-193, Aveiro, 


We study Magnus effect for a spinning disc moving in a very rarified medium. 

We proposed a new approach that is based on examining the shape of the body’s 

cavities and numerical simulation of optimal mass distribution for any particular 

case of a shape, as well as for the general case. On the base of the results 

we conclude that the reverse Magnus effect is more common phenomenon and 

is due to two factors: lateral friction to the particles, and multiple collisions of 

particles with the body originating from the fact that the body’s surface is not 

convex but contains microscopic cavities. 

� WC-22 



LNG transportation 



Chair: Henrik Andersson, Department of Industrial Economics and 


Technology, Gløshaugen, Alfred Getz vei 3, NO-7491, Trondheim, 

Norway, Henrik.Andersson@iot.ntnu.no 

1 - Branch-and-Price for creating an Annual Delivery Program 

(ADP) of Multi-Product Lique 

ed Natural Gas 

Jørgen Glomvik Rakke, IØT, NTNU, Alfred Getz veg 3, 

Sentralbygg I, 7491, Trondeheim, Sør-Trøndelag, Norway, 

jorgen.rakke@iot.ntnu.no, Marielle Christiansen, Henrik 

Andersson, Guy Desaulniers 

We consider a ship routing and inventory management problem for one of the 

world’s largest producers of LNG. The problem is to manage the producer’s 

inventory and fleet of ships to create an ADP that respects the long-term contracts. 

A MIP formulation, the Basic Voyage Model (BVM), based on pregeneration 

of all scheduled voyages will be presented. The LP-relaxation of 

the BVM is weak. To reduce the gap, the BVM is reformulated using contract 

delivery patterns generated by column generation. The sub-problems do not 

possess the integrality property, which might help raise the bound. 

2 - Long term planning in the LNG value chain 

Kristin Tolstad Uggen, Applied economics and operations 

research, SINTEF Technology and society, S.P. Andersens v 5, 

7465, Trondheim, Norway, ktu@sintef.no, Adrian Werner, Marte 

Fodstad, Arnt-Gunnar Lium 

We will present a strategic planning tool for the LNG business with planning 

horizons of 10-20 years. The model has a company focus, and makes decisions 

on how to expand the existing portfolio with potential investment opportunities 

in new LNG terminals, vessels and different types of purchase- and salescontracts. 

Since future energy prices are uncertain, the model has stochastic 

natural gas and LNG prices both in contracts and on spot markets. Different 

modeling issues (especially transportation and contracts) and first results will 

be presented. 

3 - A branch-and-price-and-cut algorithm for a maritime 

liquefied natural gas inventory routing problem 

Henrik Andersson, Department of Industrial Economics and 


Technology, Gløshaugen, Alfred Getz vei 3, NO-7491, 

Trondheim, Norway, Henrik.Andersson@iot.ntnu.no, Marielle 

Christiansen, Guy Desaulniers 

A branch-and-price-and-cut algorithm for a maritime liquefied natural gas 

(LNG) inventory routing problem will be presented. A heterogeneous fleet 

of ships is used to transport LNG from liquefaction plants to regasification terminals. 

The production at the plants and the sale at the terminals are decisions. 

A decomposition of the problem is done where voyages are used to describe 

the movement of the ships. The proposed valid inequalities are derived using 

the heterogeneity of the fleet, only full loading and unloading of ship tanks and 

the variable production and consumption. 

259


� WC-23 


6.2.49 

Data Mining in Portfolio Analysis 1 



Chair: Vadim Mottl, Intelligent Systems, Computing Center of the 

Russian Academy of Sciences, Vavilov St. 40, 119333, Moscow, 

Russian Federation, vmottl@yandex.ru 

Chair: Evgeny Bauman, Markov Processes International, 25 Maple 

ave, 07901, Summit, New Jersey, United States, 

evbauman@markovprocesses.com 

1 - A mathematical and algorithmic framework for dynamic 

returns-based style analysis of investment portfolios 

Vadim Mottl, Intelligent Systems, Computing Center of the 


Russian Federation, vmottl@yandex.ru, Olga Krasotkina, 

Michael Markov, Ilya Muchnik 

Transactions of an investment portfolio are, as a rule, hidden from public. The 

only available information is the time series of its daily or monthly returns. The 

idea of Dynamic Investment Style Analysis is monitoring the hidden capital 

sharing in a portfolio as estimating a time-varying regression model of its periodic 

returns with respect to those of asset classes the portfolio might contain. 

Under a natural Markov assumption on the trading process, the sweep method, 

pair-wise separable quadratic programming and Kalman-Bucy filter-smoother 

are equivalent means of solving the problem. 

2 - Machine-learning for dynamic style analysis of hedge 

funds 

Olga Krasotkina, Tula State University, 300600, Tula, Russian 

Federation, ko180177@yandex.ru, Vadim Mottl, Michael 

Markov, Ilya Muchnik 

Immediate application of Dynamic Investment Style Analysis to a time series 

of peroodic returns on the portfolio under monitoring is inevitably concerned 

with making additional decisions on, first, the active subset in the assumed set 

of asset classes the portfolio might contain, and, second, partitioning it into 

two further subsets of really traded assets and buy-and-hold ones. This necessity 

turns the statistical problem of time-varying regression estimation into that 

of Machine Learning, which implies finding the most appropriate data model 

among a priori models of growing complexity. 

3 - Downside Risk Optimization via Quasi-Gradient Algorithm 

Evgeny Bauman, Markov Processes International, 25 Maple ave, 

07901, Summit, New Jersey, United States, 

evbauman@markovprocesses.com, Michael Markov 

H. M. Markowitz introduced the mean-variance analysis as an instrument for 

forming well diversified portfolios. He concluded that the most theoretically 

robust measure was semi-variance. Semi-variance (downside risk) is the expected 

value of the squared negative deviations about a specified "target’ rate 

of return. There are a lot heuristic algorithms to optimize downside risk. G. 

M. de Athayde suggested a quasi-gradient algorithm but there is no theoretical 

proof of its convergence. We developed a modification of his algorithm and 

proved its convergence. 

4 - Peer group analysis as a clustering problem of portfolio 

trackers 

Marcus Hildmann, Swissquant Group AG, Kuttelgasse 7, 8001, 

Zürich, Switzerland, hildmann@swissquant.ch 

In this work we are combining two very important notions in risk management: 

peer group analysis of portfolio managers and index or portfolio tracking. First, 

we introduce an optimization method based on minimizing the tracking error, 

which given the initial set of assets, finds the minimal number of assets 

needed for portfolio return time-series reconstruction. Based on these results, 

we search further for an optimal clustering of portfolios into different peer 

groups. The latter is used for objective evaluation of a selection of portfolio 

managers. 

260 

� WC-24 


6.2.50 

Crew Scheduling 



Chair: Silke Jütte, Supply Chain Management and Management 

Science, University of Cologne, Albertus-Magnus-Platz, 50923, 

Cologne, Germany, silke.juette@uni-koeln.de 

1 - Integrated airline crew scheduling by a dynamic constraint 

aggregation method 

Mohammed Saddoune, École Polytechnique de Montréal and 

GERAD, 700 cremazie Est, Appt 1A, h2p-1g2, montréal, QC, 

Canada, mohammed.saddoune@polymtl.ca, Guy Desaulniers, 

Issmail Elhallaoui, Francois Soumis 

The bidline airline crew scheduling has been traditionally decomposed into two 

stages (crew pairing and crew assignment) that are sequentially solved. We propose 

a model that completely integrates both stages and solve it in a single step 

using a combined column generation/dynamic constraint aggregation method. 

Tested on real data, the proposed approach reduces the total cost by 4.76% 

compared to the sequential approach with an increase of computational time 

by a mean factor of 3.8 which is much smaller than that obtained with column 

generation solely (around 200 for the smallest instances). 

2 - Pricing by Local Search for the Airline Crew Pairing 

Problem 

Nimet Aksoy, Industrial Engineering, Sabanci University, 

Sabanci University, Orhanli Tuzla, 34956, Istanbul, Turkey, 

nimetaksoy@su.sabanciuniv.edu, S. Ilker Birbil, Kerem Bulbul, 

Husnu Yenigun 

Traditional column generation (CG) approaches to the airline crew pairing 

problem formulate the pricing subproblem as a multi-label shortest path problem 

(MLSP) typically solved over a flight network. The MLSP suffers from an 

exponential complexity even for moderate size flight networks. We propose a 

CG method, where we first look for negatively priced pairings by a local search 

(LS) mechanism in a (partial) duty network. We resort to the MLSP when the 

LS fails and also generate new duties in the process. Numerical results are 

presented that attest to the efficiency of our approach. 

3 - Ground Crew Rostering with Work Patterns at a Major 

European Airline 

Richard Lusby, Department of Management Engineering, 

Technical University of Denmark, 2800, Kgs Lyngby, Denmark, 

rmlu@man.dtu.dk, Anders Dohn, Troels Martin Range, Jesper 

Larsen 

We consider an important staff rostering problem arising in the ground operations 

of a major European airline. The so-called Ground Crew Rostering 

Problem with Work Patterns, entails assigning a set of employees to a set of 

shifts spaced over a given daily time horizon so that the robustness of the final 

roster is maximized. We present a cutting stock based formulation and propose 

a column generation solution approach that utilizes a rolling time horizon to 

find an efficient set of roster lines. Encouraging numerical results are given for 

real-life data supplied by a major European Airline. 

4 - "Divide-and-Price": A Decomposition Algorithm for 

Solving Huge Railway Crew Scheduling Problems 

Silke Jütte, Department of Supply Chain Management and 

Management Science, University of Cologne, 


silke.juette@uni-koeln.de, Ulrich Thonemann 

The railway crew scheduling problem consists of generating crew duties to operate 

a set of trains at minimal cost. Typically, schedules covering thousands 

of trains need to be generated in very short time to allow for disruptions in the 

operations. We present an algorithm which has proven to significantly reduce 

runtimes of current solution approaches while providing schedules of reasonable 

quality. The planning problem is decomposed into overlapping regions 

which are optimized simultaneously. Trains belonging to several regions are 

assigned to the region where they are covered at lowest cost.

� WC-25 


6.2.48 




Chair: Ender Ozcan, Computer Science, University of Nottingham, 

Jubilee Campus, Wollaton Road, NG8 1BB, Nottingham, United 

Kingdom, exo@cs.nott.ac.uk 

1 - A hybrid iterative decomposition approach to ROADEF 

2010 

Cristiano Nattero, DIST, University of Genova, Via all’Opera Pia 

13, 16145, Genova, GE, Italy, cristiano.nattero@gmail.com, 

Davide Anghinolfi, Luca Maria Gambardella, Roberto 

Montemanni, Massimo Paolucci, Nihat Engin Toklu 

We propose a hybrid iterative decomposition approach based on mathematical 

programming and local search (LS) improvement. The decomposition separates 

the generation of schedules of the plant outages from the generation of 

optimal production plans. The approach consists of three phases. First a quality 

feasible starting solution is obtained solving a sequence of relaxed mixed 

integer programs, then this solution is improved by a LS where also a tabu list 

is used, and finally, the outage schedule selection and time step production plan 

is produced. 

2 - A mathematical-programming-based solution approach 

for the EDF energy management problem 

Francois Soumis, GERAD, 3000 Cote Ste-Catherine, H3T 2A7, 

Montreal, Québec, Canada, francois.soumis@gerad.ca, Guy 

Desaulniers, Michel Gendreau, Louis-Martin Rousseau, François 

Lessard 

To solve this problem that involves nonlinear and disjunctive constraints, we 

propose a mathematical programming approach based on an integer linear 

multi-commodity network flow model with side constraints. In this model, 

different production cycle possibilities for the power plants of type 2 are represented 

by arcs. The expectation of the production cost of the type 1 power 

plants for a period is given by a piecewise linear convex function, yielding up 

to 50000 pieces. Starting with a small subset of these pieces, additional pieces 

are added as needed. 

3 - High-performance local search for a large-scale energy 

management problem 

Frédéric Gardi, e-lab, Bouygues SA, 40 rue Washington, 75008, 

PARIS, France, fgardi@bouygues.com, Karim Nouioua 

We present the algorithm which we have implemented in the context of the 

ROADEF/EURO Challenge 2010, for solving a large-scale energy management 

problem addressed by EDF. This algorithm is a pure local-search heuristic, 

whose design and implementation follows the methodology presented in 

the past authors’ works: no decomposition of the problem is done, and no particular 

metaheuristic is used. In this way, we have concentrated our work on 

the design of moves and on the algorithms which are behind their evaluation. 

� WC-26 


3.1.11 

Models of Cooperative Games: Theory and 

Applications 






1 - Cooperative Games under Interval Uncertainty: On the 

Convexity of the Interval Undominated Cores 

Sirma Zeynep Alparslan Gok, Mathematics, Faculty of Arts and 

Sciences, Suleyman Demirel University, Faculty of Arts and 



zeynepalparslan@yahoo.com, Mariana Rodica Branzei, Oana 

Branzei 

This paper extends interval-type core solutions for cooperative interval games 

by discussing the set of undominated core solutions which consists of the interval 

nondominated core, the square interval dominance core, and the interval 

dominance core. The interval nondominated core is introduced and it is shown 

that it coincides with the interval core. A straightforward consequence of this 

result is the convexity of the interval nondominated core of any cooperative 

interval game. 

2 - Application of cooperative TU-game solution concepts 

to a partition function form game 

David Bartl, Department of Mathematics, University of Ostrava, 

30.dubna 22, 701 03, Ostrava, Czech Republic, bartl@osu.cz 

The classical cooperative TU-game solution concepts (core, bargaining set, ...) 

are defined for games the coalitional function form. We want to apply them 

to a cooperative TU-game which is given in the strategic or normal form. For 

that reason, contemplating the solution concepts, we extend their definition to 

the games in the partition function form (Lucas & Thrall, 1963), into which the 

strategic form game can be converted easily. The proposed methodology, apart 

from defining the gamma- or delta-core, which are known, enables to define, 

e.g., gamma- or delta-bargaining set etc. 

3 - A value for interval games with multi-choice coalitions: 

Axiomatic characterizations and applications 

Mariana Rodica Branzei, Faculty of Computer Science, 

"Alexandru Ioan Cuza” University, 16, Berthelot St., 700483, 

Iasi, Romania, branzeir@info.uaic.ro, Hao Sun, Genjiu Xu 

We consider cooperative games arising from situations where agents can participate 

at several levels in a joint venture and where one can only predict upper 

and lower bounds for the outcome of agents’ collaborative actions. This model 

of games fits many OR situations. To assist managers in scientific decisionmaking 

regarding reward/cost sharing problems with interval data and multiple 

levels of cooperation game-theoretic solution concepts are helpful. We propose 

an interval value based on multi-choice coalition payoffs, axiomatically 

characterize it and illustrate its use in practice. 

� WC-27 


8.2.06 

TRANSPORTATION PLANNING 



Chair: Jorge Freire de Sousa, Industrial Management Unit, Faculty of 

Engineering University of Porto, Rua Roberto Frias, 4200-465, Porto, 

Portugal, jfsousa@fe.up.pt 

1 - A New Model for Road Accident Prediction, Based on 

an Adaptive Neuro-fuzzy Inference System 

Mario Mellano, Dept. of Roads and Transportation, Technical 

University of Bari, 70125 Bari, Bari, mmella@poliba.it, Mauro 

Dell’Orco 

Due to the complexity of a road transportation system, studying either system 

decay, or crashes as its consequence, is not a simple task. Several road accident 

prediction models have been developed to investigate effects of many 

independent variables on crashes. In this paper, we have implemented a neurofuzzy 

inference engine to obtain crash prediction, and have tested the model on 

a four-lane median-divided Italian extra-urban road. The proposed algorithm 

has been tested through a simulation. Results have shown the goodness of the 

proposed model, and a remarkable significance of drivers’ short-term memory 

on accidents. 

2 - Discrete events simulation of a flexible service for people 

transportation 

Pasquale Carotenuto, Istituto per le Applicazioni del Calcolo "M. 

Picone", Consiglio Nazionale delle Ricerche, via dei Taurini 19, 

00185, Roma, RM, Italy, carotenuto@iac.cnr.it, Giovanni 

Storchi, Artem Serebriany 

261


The realization of innovative transport services, require increasingly greater 

flexibility and inexpensiveness of the service. In many cases the solution is 

to realize demand responsive transportation system. In this work, we address 

a Demand Responsive Transport Service able to manage incoming transport 

demand using a solution based on heuristics algorithm to solve a Dial-a-Ride 

Problem instance. The solutions provided by the heuristics are simulated in 

a discrete events environment in which it is possible to reproduce the programmed 

activities and evaluate some performance indicator. 

3 - Using travel time predictions for planning and control 

at public transport companies 

Jorge Freire de Sousa, Industrial Management Unit, Faculty of 

Engineering University of Porto, Rua Roberto Frias, 4200-465, 

Porto, Portugal, jfsousa@fe.up.pt, João Mendes-Moreira 

In the last decade public transport companies (PTCs) made an important effort 

on gathering data about the actual work of their vehicles using Automatic Vehicle 

Location systems. PTCs have data about what happened but they are not 

being able yet to convert this data in useful information for the planning tasks, 

namely: trips definition, buses and drivers duties, and the assignment of duties. 

A major indicator used in these tasks is the expected travel time (TT). In this 

talk we explain how TT predictions can be used in the planning and control 

tasks for typified situations. 

� WC-28 



Scheduling with Lags and Setups 



Chair: Alessandro Condotta, University of Leeds, United Kingdom, 

scs6ac@leeds.ac.uk 

1 - On the complexity of the hybrid flow shop with adjustment 

Jan Pelikan, Econometrics, University of Economics Prague, W. 

Churchill sq. 4, 13067, Prague, Czech Republic, pelikan@vse.cz 

The paper describes a case study of job scheduling in a mechanical-engineering 

production plant with a goal to minimise the overall processing time, or 

makespan. The production jobs are processed by machines, and each job is 

assigned to a certain machine for technological reasons. Before processing a 

job, the machine has to be adjusted; there is only one adjuster, who adjusts all 

of the machines as necessary. This problem is treated as a hybrid two-stage 

flow-shop: the first stage of the job processing is represented by the machine 

adjustment for the respective job, and the second stage by the processing of 

the job itself on the adjusted machine. A mathematical model is proposed, a 

heuristic method is formulated. Partition problem and 3-partition problem are 

reduced onto hybrid flow shop with adjustment so this is a proof that the flow 

shop with adjustment is NP hard in strong sense. 

2 - On the complexity of scheduling with negative time lags 

Adel Manaa, Operational Research, LIP6, 104 avenue du 

Président Kennedy, Bureau 538, 57016, Paris, France, 

adel.manaa@lip6.fr, Claire Hanen 

We investigate the scheduling problems with negative time lags to minimise 

the makespan. Unlike an ordinary precedence constraint, a negative time lag 

l between two tasks i and j, means that j cannot start earlier than —l time 

units before i. In such problems, the precedence graph may have cycles which 

may lead to unfeasibility. We identify polynomially solvable cases for different 

precedence graph structures on a single machine and on parallel machines. We 

also prove that for cyclic and star-like graphs, the problem with negative time 

lags is NP-complete. 

3 - Real-Time Scheduling of Failure-Prone Manufacturing 

Systems with Setup Times and Costs 

Fernando Tubilla, Mechanical Engineering Department, MIT, 23 

Sidney Street 208, 02139, Cambridge, MA, United States, 

ftubilla@mit.edu, Stan Gershwin 

We develop a new closed loop policy for scheduling manufacturing systems 

with setup times/costs and subject to random machine failures. The policy is 

exhaustive, prioritized, and easily tunable. After stating and motivating the 

policy, we use Lyapunov functions to obtain an easy-to-evaluate and not-tooconservative 

sufficient condition that ensures bounded cycle times for all items. 

We then show that the policy has very good performance with respect to long 

term average costs and transient costs. Finally, we discuss how our deterministic 

results apply to the failure-prone stochastic system. 

262 

4 - On scheduling multiple-operation jobs with time-lags 

Alessandro Condotta, University of Leeds, United Kingdom, 

scs6ac@leeds.ac.uk, Natalia Shakhlevich 

In many real-world scheduling problems, a job consists of multiple operations 

that are required to be processed with given time-lags in between. In this talk 

we review models and algorithms known in literature for scheduling multipleoperation 

with time-lags. We give an insight into the difficulties of finding 

efficient heuristics for NP-hard cases and present a new complexity result for 

the case with two operations per job where the sequence of all first operations is 

given. As example of a real-world application, we present an advanced model 

of scheduling patients in a chemotherapy clinic. 

� WC-29 


8.2.11 

Portfolio Selection 



Chair: David Wozabal, Business Adminstration, University of 

Vienna, Bruenner Str. 72, A-1210, Vienna, Austria, 


1 - Set-Portfolio Selection with the Use of Market Stochastic 

Bounds 

Sergio Ortobelli, MSIA, University of Bergamo, Via dei 

Caniana, 2, 24127, Bergamo, Italy, sol@unibg.it, Enrico 

Angelelli, Daniele Toninelli 

This paper proposes an ex-post comparison of portfolio selection strategies applied 

to some preselected assets among more than 1300 stocks of the US Market. 

In particular, for any portfolio selection problem we preselect few assets 

that optimize the association with the market stochastic bounds and present the 

highest ex-ante reward-risk performance. Then we propose the comparison of 

the ex-post final wealth obtained with the optimization of reward-risk functionals 

that use the stochastic bounds of the preselected assets. We consider 

strategies where the investors recalibrate weekly their portfolios and we compare 

the wealth obtained under the assumption that returns follow a Markov 

chain. 

2 - Mining financial time series: New insights from modelbased 

clustering methods 

José G. Dias, Quantitative Methods & UNIDE, ISCTE, Edifício 

ISCTE, Av. das Forças Armadas, 1649-026, Lisboa, Portugal, 

jose.dias@iscte.pt, Jeroen K. Vermunt, Sofia B. Ramos 

In recent years large amounts of financial data have become available for analysis. 

We propose to explore returns from 21 stock markets by model-based 

clustering of regime switching models. These models allow the relaxation of 

traditional assumptions such as conditional Gaussian returns. The data mining 

approach handles simultaneously the heterogeneity across stock markets and 

time. The results show a clear distinction between groups of stock markets, 

each one characterized by different regime switching states that correspond to 

different expected return-risk patterns. 

3 - Optimizing Value-at-Risk using the Difference of Convex 

Algorithm 

David Wozabal, Business Adminstration, University of Vienna, 

Bruenner Str. 72, A-1210, Vienna, Austria, 


Value-at-Risk (VaR) is an integral part of contemporary financial regulations. 

This paper treats a Value-at-Risk constrained Markowitz style portfolio selection 

problem when the distribution of returns are given in the form of finitely 

many scenarios. The problem is non-convex but can be reformulated as a difference 

of convex program. We apply the difference of convex algorithm to solve 

the problem. Numerical results comparing the solutions found by the DCA to 

the respective global optima for relatively small problems as well as numerical 

studies for large real life problems are discussed.

4 - Usage of Conditional CAPM method for explaining portfolio 

returns: An application for services, financial and 

industrial sectors’ returns in Istanbul Stock Excange 

Serap Atbas, Social Sciences Institute, Turkey, 

satbas@gmail.com, Melik Kamisli, Guven Sevil 

In the light of studies in the literature it has been concluded that, the static 

CAPM can not fully explain portfolio returns, the beta coefficients of the assets 

do not remain constant over time. When it is taken into consideration that 

the expected return and beta that has changed over time, conditional CAPM 

method better explains the portfolio return. The aim of this study is to test 

whether it is better explained or not the return on services, financial and industrial 

sectors by using Conditional CAPM methods in Turkey between the years 

03/01/2001 and 08/07/2009. 

� WC-30 


8.2.13 

Advances in Quantitative Credit Risk 

Modeling: Change We Need 


in Banking 


Chair: Bart Baesens, Decision Sciences and Information Mangement, 

K.U.Leuven, Naamsestraat 69, B-3000, Leuven, Leuven, 

bart.baesens@econ.kuleuven.ac.be 

Chair: David Martens, Decision Sciences and Information 


3000, Leuven, Belgium, David.Martens@econ.kuleuven.be 

Chair: Christophe Mues, School of Management, University of 

Southampton, SO17 1BJ, Southampton, United Kingdom, 

C.Mues@soton.ac.uk 

1 - Modeling credit rating migrations dependent on the 

business cycle 

Wouter Verbeke, Faculty of Economics and Business, 

K.U.Leuven, Naamsestraat 69, B-3000, Leuven, Belgium, 

wouter.verbeke@econ.kuleuven.be, Koen Berteloot, David 

Martens, Gerd Castermans, Tony Van Gestel, Bart Baesens 

This study introduces a modeling methodology, based on standard ordinal logistic 

regression, to estimate the credit rating migration matrix as a function 

of rating dynamics and macroeconomic indices. This allows to stress test and 

analyze the impact of economical downturn conditions on credit ratings, and to 

assess the risk related to a credit portfolio. A formal evaluation methodology is 

proposed that permits to interpret intuitively the difference between estimated 

and historic migration matrices. The proposed methodology is applied to model 

US corporate rating migrations from 1984 to 2008. 

2 - Rating Banks: a myth resolved 

Elisabeth Van Laere, Accounting & Finance, Vlerick Leuven 

Gent Mgt School, Reep 1, 9000, Gent, Belgium, 

elisabeth.vanlaere@vlerick.be, Bart Baesens 

One of the main challenges in developing bank regulation is the fact that risks 

taken in the process of financial intermediation are hard to observe. In order 

to reduce this lack of transparency CRAs provide information that can help to 

evaluate the risk. There is evidence of more split ratings over financial institutions 

suggesting that banks are more difficult to rate because of their opaqueness. 

In this paper the main research question is "What are the key determinants 

driving bank ratings and how robust are these?’ This question will be addressed 

using new variables and techniques. 

3 - Regression model development for Exposure at Default 

(EAD) 

Iain Brown, School of Management, University of Southampton, 

University of Southampton, University Road, SO17 1BJ, 

Southampton, United Kingdom, I.Brown@soton.ac.uk, 

Christophe Mues, Lyn Thomas 


In this presentation we propose a comprehensive and robust model for predicting 

the exposure at default (EAD). For off-balance-sheet (credit cards) to 

calculate the EAD one requires the committed but unused loan amount times a 

credit conversion factor (CCF). Ordinary least squares, logistic and cumulative 

logistic regression models are analysed with the main aim of finding the most 

robust and comprehensible model for the prediction of the CCF. A real-life data 

set with monthly balance amounts for clients over the period 2001-2004 is used 

in the building and testing of the regression models. 

4 - Competing Risks Survival Model for Residential Mortgage 

Loans 

Mindy Leow, School of Management, University of 

Southampton, University of Southampton, University Road, 

SO17 1BJ, Southampton, m.leow@soton.ac.uk, Christophe 

Mues, Lyn Thomas 

With access to a dataset on defaulted mortgages kindly provided by a major 

UK bank, we develop a competing risk survival analysis model to predict the 

time taken for a defaulted mortgage loan to go to some event (repossession or 

closure), which would allow for a more accurate prediction of loss as these periods 

could vary from months to years depending on the health of the economy. 

Besides loan-related characteristics, we incorporate a time-dependent macroeconomic 

variable to track the year-on-year change in the national HPI. 

� WC-31 


8.2.15 


routing III 



Chair: Geir Hasle, Applied Mathematics, SINTEF ICT, P.O. Box 124 

Blindern, NO-0314, Oslo, Norway, Geir.Hasle@sintef.no 

1 - A Novel Heuristic for Sequencing Mixed-Model Assembly 

Lines 

Rico Gujjula, Department of Production Management, TU 

Berlin, TU Berlin - H 95 - H 9155, Straße des 17. Juni 135, 

10623, Berlin, Berlin, Germany, rico.gujjula@tu-berlin.de, 

Sebastian Werk, Hans-Otto Guenther 

Many methods which were developed to sequence mixed-model assembly 

lines, e.g. in the automotive industry, fail to cope with real-life problem instances. 

We propose a novel heuristic solution procedure which is able to 

handle large and supposedly difficult problem instances. Test scenarios considering 

real-life aspects from car manufacturers were generated to evaluate 

the performance of the heuristic for realistic problem instances. It is shown 

that the proposed heuristic significantly outperforms simple priority rule-based 

methods and requires only reasonable computational effort. 

2 - Production scheduling in a multi-page invoice printing 

system 

Jiyin Liu, Business School, Loughborough University, Ashby 

Road, LE11 3TU, Loughborough, Leicestershire, United 

Kingdom, j.y.liu@lboro.ac.uk, Yan Zhang, Zhili Zhou 

We study the production scheduling problem in a practical multi-page invoice 

printing system that comprises three stages: the stencil preparation stage, 

the page printing stage and the invoice assembly stage. The last two stages 

have parallel machines and the second stage involves two types of sequencedependent 

setups. The objective is to minimize the makespan. We first formulate 

the problem into a mixed-integer linear programming model and then 

solve it using a hybrid genetic algorithm. Numerical experiment indicates that 

the algorithm can solve the problem efficiently and effectively. 

3 - Self-Optimization Aspects for Dynamic Scheduling 

Ivo Pereira, ISEP, Portugal, iasp@isep.ipp.pt, Ana Madureira 

Scheduling is a critical function that is present throughout many industries and 

applications. A great need exists for developing scheduling approaches that can 

be applied to a number of different scheduling problems with significant impact 

on performance of business organizations. A challenge is emerging in the 

design of scheduling support systems for manufacturing environments where 

dynamic adaptation and optimization become increasingly important. At this 

scenario, self-optimizing arise as the ability of the agent to monitor its state and 

performance and proactively tune itself to respond to environmental stimuli. 

263


� WC-32 


8.2.17 

Long-term modelling for agriculture and 

forestry 


and Climate 


Chair: Nicklas Forsell, CMA, CMA, France, 

nicklas.forsell@cma.ensmp.fr 

1 - Flexible potential of renewable energy sources in longterm 

energy planning models 

Nicklas Forsell, CMA, CMA, France, 

nicklas.forsell@cma.ensmp.fr 

An approach is presented for considering renewable energy sources in longterm 

energy planning models such as MARKAL-TIMES. The proposed approach 

focuses on modeling of forest potentials, and specifically considers the 

development of periodic unused forest potentials. Instead of exogenously defining 

a single "business-as-usual’ forest potential scenario, the model endogenously 

constructs a forest potential from a set of exogenously defined feasible 

forest potential scenarios. The proposed approach is demonstrated utilizing a 

case study of the future development of the Swedish energy system, specifically 

focusing on the future use of forest biomass. 

2 - Retribution of carbon sinks in forest biomass Vs stimulation 

of timber consumption: what impacts on the 

French forest sector? 

Philippe Delacote, INRA- LEF, France, 

philippe.delacote@eui.eu, Sylvain Caurla 

Forests contribute to climate change mitigation by sequestering carbon in forest 

biomass and in wood products and by substituting high energy products with 

wood products such as fuelwood instead of fossil fuel or construction wood 

instead of concrete, what we call substitution effect. Climate policies can thus 

focus on these effects. We assess potential impacts of substitution and stock 

policies on the French forest sector and on carbon accounting. Stock policies 

would be difficult to implement as long as carbon price does not reach levels 

comparable to timber prices. 

3 - Large-Scale Modelling of Global Food Security under 

Extreme Weather Events 

Sabine Fuss, Forestry, IIASA, Schlossplatz 1, -, 2361 , 

Laxenburg, -, Austria, fuss@iiasa.ac.at, Jana Szolgayova, Petr 

Havlik, Michael Obersteiner 

Climate change is supposed to impact food security more by increased frequency 

of extreme wheather events than by changes in average temperature and 

precipitation. We adapt the GLOBIOM model to investigate the impact of yield 

stochasticity when ensuring food security is an explicit constraint. Under business 

as usual, there is a trade-off between the insurance of food security and the 

pressure on average prices. Rural development policies facilitating switches in 

management systems (rainfed/irrigated) and trade liberalization may improve 

this. 

� WC-33 


8.2.19 

Risk and Uncertainty in Energy Models 



Chair: Dominik Möst, Universität Karlsruhe, Institut for Industrial 

Production (IIP), Hertzstraße 16, 76131, Karlsruhe, Germany, 

Dominik.Moest@kit.edu 

1 - Optimal gas procurement for a power producer using 

stochastic programming 

264 

Miguel Carrión, Electrical Engineering, University of Castilla - 

La Mancha, Avda Carlos III, s/n, Campus Fábrica de Armas, 

45071, Toledo, miguel.carrion@uclm.es, Antonio J. Conejo 

We consider the problem of a power producer that seeks to determine the 

monthly allocation of natural gas for its combined cycle gas turbines (CCGTs) 

within a planning horizon of one year. A stochastic programming model is used 

to address this problem where the pool price is characterized as a stochastic 

process using a set of scenarios. The monthly allocation of natural gas and the 

on/off status of the units are here-and-now decisions, whereas the pool trading 

is a wait-and-see decision. The performance of the proposed model is assessed 

by means of a realistic case study. 

2 - The impact of renewable capacities on the conventional 

power plant system in 2030 in Germany 

Dominik Möst, Universität Karlsruhe, Institut for Industrial 

Production (IIP), Hertzstraße 16, 76131, Karlsruhe, Germany, 

Dominik.Moest@kit.edu, Massimo Genoese, Wolf Fichtner 

One major contemporary strategic challenge for electricity supply systems is 

the integration of substantial amounts of renewable energy sources. Within the 

German sustainable energy strategy a target of 80% of renewables in electricity 

production is stipulated. This contribution therefore derives the residual load 

for the year 2030 and 2050 based on different correlations between the load and 

the renewable feed-in. This residual load is the starting point for a peak-load 

pricing model, which determines an optimal power plant mix under various 

renewable feed-in scenarios. 

3 - Extending Dynamic Programming’s Simulations and 

Regressions Approach to Value and Control a Hydropower 

System 

Michel Denault, HEC Montréal, Canada, 

michel.denault@hec.ca, Jean-Guy Simonato, Lars Stentoft 

We investigate the control problem of managing a hydropower to optimize its 

value. Exogenous stochastic state variables taken into consideration are the 

power spot price, the water inflows and the power demand. One endogenous 

state variable is the water level in the dam. We model the problem as a dynamic 

program and solve it with a combination of simulations and value function approximations, 

generalizing the well-known American option pricing techniques 

that use simulation and regressions. The use of simulation paths for all state 

variables allows considerable flexibility. 

4 - Extension Planning for Generation Units in a Hydrothermal 

Portfolio 

Ulf Kasper, Institute of Power Systems and Power Economics, 

RWTH Aachen University, Schinkelstraße 6, 52062, Aachen, 

Germany, uk@iaew.rwth-aachen.de, Albert Moser 

The increasing demand of system reserve and the development of spot prices 

in Europe motivate power generation and trading companies to consider an extension 

of their generation portfolios. In order to review the efficiency of such 

projects their effect on the company’s entire hydrothermal portfolio has to be 

taken into account. The Institute of Power Systems and Power Economics of 

RWTH Aachen University has developed a practically approved program for 

optimizing power generation and trading on markets for scheduled energy and 

reserve to simulate such effects on generation portfolios. 

� WC-34 


8.2.23 

Duality Problems 



Chair: Iqbal Husain, Mathematics, Jaypee College of Engineering 

and Technology, Dept. of Mathematics,Jaypee College of 

Engineering and Technology, A-B Road, Raghogarh, Guna.MP, 

476-226, Bhopal, MadhyaPradesh, India, ihusain11@yahoo.com 

1 - On weak conjugate maps, Fenchel and Fenchel Lagrange 

Duality, constructed by weak conjugate maps 

Ilknur Atasever, Department of Mathematics, Anadolu 

University, Anadolu Universitesi Fen Fakultesi Matematik 

Bolumu, 26470, Eskisehir, Turkey, iatasever@anadolu.edu.tr, 

Yalcin Kucuk, Mahide Kucuk

Augmented Lagrangian dual problem, which is constructed by using weak conjugate 

maps, was studied by Azimov and Gasimov for nonconvex optimization 

problems. In this work, by using weak conjugate maps Fenchel, and Fenchel 

Lagrange dual problems are constructed. Definition of stability of primal problem 

with respect to Fenchel and Fenchel Lagrange duals are given and it is 

proved that stability of the primal problem implies the strong duality. Furthermore, 

necessary and sufficient conditions for stability of primal problem with 

respect to Fenchel and Fenchel Lagrange dual problems are given. 

2 - The relations between the optimal objective values of 

the Fenchel,FenchelLagrange and Lagrange dual problems,constructed 

by weak conjugate maps, and optimality 

conditions 

Yalcin Kucuk, Department of Mathematics, Anadolu University, 

anadolu Universitesi Fen Fakultesi Matematik Bolumu, 26470, 

Eskisehir, ykucuk@anadolu.edu.tr, Ilknur Atasever, Mahide 

Kucuk 

Augmented Lagrangian dual problem, which is constructed by using weak conjugate 

maps, and optimality conditions were studied by Azimov and Gasimov 

for nonconvex optimization problems. In this work, the relations between the 

optimal objective values of Lagrange , Fenchel and Fenchel Lagrange dual 

problems, which are constructed by using weak conjugate maps, are given. Furthermore, 

optimality conditions for Fenchel and Fenchel Lagrange dual problems 

are presented. 

3 - Dynamic games and symmetric duality in variational 

problems 

Iqbal Husain, Mathematics, Jaypee College of Engineering and 

Technology, Dept. of Mathematics,Jaypee College of 


476-226, Bhopal, MadhyaPradesh, India, ihusain11@yahoo.com 

Certain single and multiple dynamic games are modeled and shown to be equivalent 

to pairs of symmetric dual single objective variational problems and symmetric 

dual multiobjective variational problems which have more general formulations 

than the traditional ones. Further for these pairs, the usual duality 

results are established under certain generalized convexity conditions and it is 

also shown that our results can be considered as dynamic generalizations of 

those of the corresponding static cases through natural boundary valued problems. 

4 - Mixed type symmetric and self duality for multiobjective 

variational problems with support functions 

Iqbal Husain, Mathematics, Jaypee College of Engineering and 

Technology, Dept. of Mathematics,Jaypee College of 


476-226, Bhopal, MadhyaPradesh, India, 

ihusain11@yahoo.com, Rumana Mattoo 

Mixed type symmetric dual pair for multiobjective variational problems containing 

support functions is formulated. This mixed formulation unifies two 

existing pairs Wolfe and Mond-Weir type symmetric dual multiobjective variational 

problems containing support functions. For this pair of mixed type nondifferentiable 

multiobjective variational problems, various duality theorems are 

established under convexity-concavity and pseudoconvexity-pseudoconcavity 

of certain combination of functionals appearing in the formulation. A self duality 

theorem under additional assumptions on the kernel functions that occur 

in the problems is validated. A pair of mixed type nondifferentiable multiobjective 

variational problem with natural boundary values is also formulated to 

investigate various duality theorems. 

� WC-35 


6.2.46 

Integer Nonlinear Programming 



Chair: Christoph Buchheim, Fakultät für Mathematik, Technische 

Universität Dortmund, Germany, 

christoph.buchheim@tu-dortmund.de 

1 - Integer Programming Subject to Monomial Constraints 

Dennis Michaels, Institute for Operations Research, ETH 

Zuerich, Raemistrasse 101, 8092, Zuerich, Switzerland, 


dennis.michaels@ifor.math.ethz.ch, Christoph Buchheim, Robert 

Weismantel 

We investigate integer programs containing equality constraints where the left 

hand sides consist of one monomial. Due to the number-theoretic nature of 

these constraints, standard methods based on linear algebra cannot be applied 

directly. Instead, we present a reformulation resulting in integer programs with 

linear constraints and polynomial objective functions, using prime decompositions 

of the right hand sides. Moreover, we show that minimizing a linear 

objective function with nonnegative coefficients over bivariate constraints is 

possible in polynomial time. 

2 - A Fast Algorithm for Convex Quadratic Integer Programming 

with Applications in Electronics 

Christoph Buchheim, Fakultät für Mathematik, Technische 

Universität Dortmund, Germany, 

christoph.buchheim@tu-dortmund.de, Alberto Caprara, Andrea 

Lodi 

We present a branch-and-bound algorithm for minimizing a convex quadratic 

objective function over integer variables subject to convex constraints. In a 

given node of the enumeration tree, corresponding to the fixing of a subset of 

the variables, a lower bound is given by the continuous minimum of the restricted 

objective function. We improve this bound by exploiting the integrality 

of the variables using suitably-defined lattice-free ellipsoids. 

Experiments with instances of the closest vector problem as well as instances 

arising in an application in electronics show that our approach is very fast on 

both unconstrained problems and problems with box constraints. The main 

reason is that all expensive calculations can be done in a preprocessing phase, 

while a single node in the enumeration tree can be processed in linear time in 

the problem dimension. 

3 - N-ary Relation Persistencies for Unconstrained 

Quadratic Binary Optimization : Generation, Implications, 

Linearization. 

Serigne Gueye, Laboratoire de Mathématiques Appliquées du 

Havre (LMAH), Université du Havre, 25 rue Philippe Lebon, BP 

540, 76058, Le Havre Cedex, France, 

serigne.gueye@univ-lehavre.fr, Philippe Michelon 

We propose and exploit some persistencies on product of several variables for 

the Unconstrained Quadratic Boolean Problem (UQBO). These persistencies 

are generalization of known results on one or a product of two variables derived 

from the so-called first and second-order derivatives. From a set of persistencies 

some others may be deduced with various techniques using graph 

theory or Constraint Programming. Generation and deduction of persistencies 

gives an iterative proceduce that may solve entirely low density instances. For 

harder cases, it is possible to derive valid inequalities added in any linearization 

scheme. Numerical results are presented. 

4 - The Power of SDP Relaxations - Computing Strong 

Bounds for QAPs and Graph Problems 

Hans Mittelmann, School of Math&Stats, Arizona State 

University, Box 871804, 85287-1804, Tempe, AZ, United States, 

mittelmann@asu.edu 

As is well-known semidefinite relaxations of discrete optimization problems 

yield excellent bounds on their solutions. We present two examples from our 

recent research. The first addresses the quadratic assignment problem and a formulation 

is developed which yields the strongest lower bounds known for larger 

dimensions. Utilizing the latest iterative SDP solver and ideas from verified 

computing a realistic problem from communications is solved for dimensions 

up to 512. The second area is the computation of bounds in graph problems. 

A strategy based on the Lovasz theta function is generalized to compute lower 

bounds for the chromatic number of graphs and upper bounds on the spherical 

kissing number utilizing SDP relaxations. Multiple precision SDP solvers are 

needed and improvements on known results for kissing numbers in dimensions 

up to 23 are obtained. This is joint work with Jiming Peng and Frank Vallentin. 

� WC-36 


3.1.05 

Container Terminal Applications 



Chair: Ulrich Dorndorf, INFORM GmbH, Pascalstr. 23, 52076, 

Aachen, Germany, udorndorf@acm.org 

265


1 - Scheduling freight trains in rail-rail transshipment 

yards 

Erwin Pesch, FB 5, University of Siegen, Hoelderlinstr. 3, 

57068, Siegen, Germany, erwin.pesch@uni-siegen.de, Nils 

Boysen, Florian Jaehn 

Transshipment yards, where gantry cranes allow for an efficient transshipment 

of containers between different freight trains, accelerate container handling, so 

that multiple smaller trains with equal destination can be consolidated to a reduced 

number of trains. An important problem continuously arising during the 

daily operations of a transshipment yard is the train scheduling problem, which 

decides on the succession of trains at the parallel railway tracks. This problem 

with a special focus on resolving deadlocks and avoiding multiple crane picks 

per container move is investigated. 

2 - Optimized load planning of trains in intermodal transportation 

Florian Bruns, Institute of Mathematics, Technical University of 

Clausthal, Germany, flbruns@uos.de, Sigrid Knust 

We study the problem of load planning for trains in intermodal container terminals. 

The objective is to assign load units to wagons of a train such that the 

utilization of the train is maximized, and setup and transportation costs in the 

terminal are minimized. Contrary to previous approaches additionally weight 

restrictions for the wagons are integrated into our model. We present two different 

integer linear programming formulations and test them on some real-world 

instances. It is shown that even non-commercial MIP-solvers can solve our 

models to optimality in reasonable time. 

3 - Scheduling triple cross-over stacking cranes in a container 

yard 

Frank Schneider, INFORM -GmbH, Pascalstrasse 23, 52076, 

Aachen, Germany, frank.schneider@inform-ac.com 

We describe an approach for scheduling triple cross-over rail mounted stacking 

cranes at a real container yard with automated container storage blocks with 

asynchronous hand over at the transfer areas at both block front ends. The problem 

is characterized by frequent long crane moves that make job assignment 

and crane routing particularly challenging as a tight synchronization between 

the cranes is required. Simulation results show that our method performs significantly 

better than commonly used heuristics, leading to a productivity gain 

of more than twenty percent. 

� WC-37 


3.1.09 

Territorial Decision Making 



Chair: Florent Joerin, ESAD, Université Laval, Pavillon 

Félix-Antoine-Savard, 2325, rue des bibliothèques, G1V 0A6, 

Quebec City, Quebec, Canada, Florent.Joerin@esad.ulaval.ca 

1 - Cohesive groups for cohesive representations : how to 

structure the problem of urban sprawl ? 

Florent Joerin, ESAD, Université Laval, Pavillon 

Félix-Antoine-Savard, 2325, rue des bibliothèques, G1V 0A6, 

Quebec City, Quebec, Canada, Florent.Joerin@esad.ulaval.ca, 

Pierre Rondier 

This contribution illustrates the use of social representations towards the enrichment 

of problem structuring methods. Our goal is to propose a participatory 

approach throughout facilitating the emerging of a collective shared representation 

of complex problems. Organizing stakeholders in subgroups with varied 

levels of cohesiveness enables to identify with transparency the core problem 

representations. Using media interaction in the written press and collective 

cognitive mapping, we compare how cohesiveness structures the social representations 

of Urban Sprawl in the Quebec metropolitan Area 

2 - Approaching Investments On Gas Transmission Networks 

With Robustness 

266 

Slawomir Pietrasz, CRIGEN-DETI-PSO, GDF SUEZ, 361 av du 

President Wilson, 93210, St-Denis-La-Plaine, France, 

slawomir.pietrasz@gdfsuez.com 

Making investment decisions on regional transmission networks that no decision 

maker would regret is becoming a challenge for the French gas operator 

GRTgaz. Focusing on the evolving context of investment studies we identify 

uncertain technical, economic and strategic parameters who play a significant 

role in the out coming investment proposition. We analyse what is at stake when 

it comes to robustness and flexibility by introducing a risk measure. Eventually 

we offer a set of new R&D approaches, that we hope, will enable to shift 

present operational models into future robust decision rules. 

� WC-38 


6.2.44 

Perishable item inventory management 



Chair: Stefan Minner, Business Administration, University of 

Vienna, Brünner Strasse 72, 1210, Vienna, Austria, 

stefan.minner@univie.ac.at 

1 - Computing replenishment cycle policy parameters for a 

perishable item with non-stationary stochastic demand 

under service level constraints 

Karin Pauls-Worm, Operations Research and Logistics, 

Wageningen UR, Hollandseweg 1, 6706KN, Wageningen, 

Netherlands, karin.pauls@wur.nl, Roberto Rossi, René Haijema, 

Jack van der Vorst 

Inventory management of perishable products is one of the key challenges in 

food industry. The decision maker has to determine the timing and production 

quantity of each replenishment, in order to minimise expected costs, guarantee 

a service level and avoid excessive waste. We formulate a mixed-integer 

linear programming model to compute optimal replenishment cycle policy parameters 

for an item with a fixed lifetime of any length and a stochastic erratic 

demand under service level constraints. The model keeps track of the ages of 

the items in stock and uses a FIFO policy. 

2 - Modelling Two-Echelon Serial Inventory Systems with 

Perishable Items 

Fredrik Olsson, Industrial Management and Logistics, Lund 

University, Ole römers väg 1, Box 118, SE-221 00, 22100, Lund, 

Sweden, Fredrik.Olsson@iml.lth.se 

This paper deals with a continuous review, two-echelon serial inventory system 

with perishable items. Transportation times and the lifetime of items are fixed. 

When the age of an item has reached its lifetime, the item is useless and consequently 

discarded from the system. The downstream location faces Poisson 

demand, and demand that cannot be met immediately is backordered. We develop 

an efficient approximate technique for evaluation of (S-1,S) policies. In 

a simulation study we evaluate the quality of our approximation. The results 

show that our technique works well in most cases. 

3 - A one warehouse, multiple retailer distribution inventory 

system for perishable products 

Stefan Minner, Business Administration, University of Vienna, 

Brünner Strasse 72, 1210, Vienna, Austria, 

stefan.minner@univie.ac.at 

There exist large research streams for durable product, multi-echelon systems 

and single echelon inventory systems for perishable products. The challenge 

especially in the latter case is that base-stock policies are no longer optimal. 

However, the combination of both problems most relevant in food distribution 

has not received much attention. We present a periodic review multi-echelon 

model and determine optimal order quantities and the allocation of incoming 

orders to the retailers under service level constraints. In a numerical study, we 

compare different rationing policies and analyze the impact of perishability on 

the optimal degree of inventory centralization. 

4 - Computing an optimal ordering policy and deriving an 

day dependent (s, S) rule for perishables with fixed ordering 

costs. 

René Haijema, Operations Research and Logistics, Wageningen 

University and Research center, Hollandseweg 1, 6706 KN,

Wageningen, Netherlands, rene.haijema@wur.nl, Jan Van Der 

Wal, Nico van Dijk 

For a product with a fixed shelf life, we compute a cost optimal stock-age dependent 

ordering policy, assuming fixed ordering costs and unit outdating and 

shortage costs. An optimal policy is computed by solving a periodic Markov 

decision problem. By simulation of the optimal policy we easily derive simpler 

rules, e.g an (s, S) rule with day dependent parameter values. For a case on 

blood products, the (s, S) rule appears to be close to optimal, whereas orderup-to 

S rules perform 10-20% worse. The optimal (MDP) policy is computed 

thus for deriving and benchmarking simpler rules. 

� WC-39 


6.2.45 

Discrete-continuous scheduling 



Chair: Jan Weglarz, Institute of Computing Science, Poznan 

University of Technology, Piotrowo 2, 60-965, Poznan, 

jan.weglarz@cs.put.poznan.pl 

1 - Discrete-continuous project scheduling — a review of 

recent results 

Grzegorz Waligora, Institute of Computing Science, Poznan 

University of Technology, Piotrowo 2, 60-965, Poznan, 

Wielkopolska, Poland, grzegorz.waligora@cs.put.poznan.pl, Jan 

Weglarz 

Discrete-continuous project scheduling problems are considered, in which nonpreemptable, 

precedence-related activities simultaneously require for their processing 

units of some discrete resources and an amount of a single continuous, 

renewable, and limited resource. The processing rate of each activity is an increasing, 

continuous function of the amount of the continuous resource allotted 

to the activity at a time. A review of the results obtained for two scheduling 

criteria: makespan minimization and net present value maximization, under 

different processing rate functions, is presented. 

2 - Metascheduling of workflow jobs in a computational 

grid 

Marek Mika, Institute of Computing Science, Poznan University 

of Technology, ul Piotrowo 2, 60-965, Poznan, Poland, 

Marek.Mika@cs.put.poznan.pl, Jan Weglarz 

Workflow jobs belong to the most complex computational jobs, and are often 

performed in computational grids. We assume that grid resources are divided 

into computational and network ones. Two types of tasks are considered: computational 

tasks requiring computational resources and transmission tasks requiring 

network resources. Scheduling jobs on a grid are made at two levels: 

global for the entire grid and local for a given site. We propose an algorithm 

for the metascheduler which operate at the global scheduling level in order to 

find a feasible resource allocation that minimize the makespan. 

3 - Optimal allocation of a non-renewable continuous resource 

for jobs performed sequentially 

Rafal Rozycki, Institute of Computing Science, Poznan 

University of Technology, ul.Piotrowo 2, 60-965, Poznan, 

Poland, rafal.rozycki@cs.put.poznan.pl 

In the paper a variant of discrete-continuous scheduling problem is considered. 

We assume that a given set of jobs of different characteristics has to be 

performed sequentially using a single non-renewable continuous resource with 

limited capacity. We propose a general method for finding an optimal continuous 

resource allocation for different scheduling criteria. Moreover, some cases 

where a solution can be found analytically, are showed as well. 

4 - A longitudinal study of the production scheduling literature 

Jose M Framinan, Industrial Management, University of Seville, 

School of Engineering, Ave. Descubrimientos s/n, E41092, 

Seville, Spain, framinan@us.es, Paz Perez Gonzalez, Jose M. 

Molina-Pariente 


In this work we analyze the evolution of the production scheduling literature. A 

number of well-known OR-journals are identified and the published papers on 

production scheduling are analyzed in an automatic manner in order to present 

how this field has been reflected (in these journals) over the years and the topics 

that have received most attention. A number of conclusions can be obtained 

from this work, as the relative weight of the field has greatly varied over the 

years from one journal to another, and the topics treated show an extremely 

scattered trend. 

� WC-40 


6.2.52 

Recent Advances in Engineering 

Optimization 






1 - Predispatch of Interconnected Hydroelectric Power 

Systems with Network Topology Modification 

Silvia Carvalho, University of Campinas, Brazil, 

silvia@densis.fee.unicamp.br, Aurelio Oliveira, Christiano Lyra 

In this work the primal-dual interior point method is used to minimize the DC 

predispatch generation costs and transmission losses on hydroelectric power 

systems with previously scheduled maneuvering. A matrix structure study is 

performed, to consider the changes that occur in the system along the planning 

period. This alternative is compared with a straight optimization approach, 

which does not take advantage of the problem special structure. Numerical 

experiments with IEEE and Brazilian power systems show the speed and robustness 

of the problem approach. 

2 - Improvements on Hyperbolic Smoothing Clustering 

Method for Application with Graphics Processing Unit 

Ricardo Farias, PESC, COPPE - UFRJ, Centro de Tecnologia, 

Bloco H, Sala 319, Cidade Universitaria, Rio de Janeiro, RJ, 

Brazil, 21.941-972, Rio de Janeiro, Rio de Janeiro, Brazil, 

rfarias@gmail.com, Sergio B. Villas-Boas, Adilson Elias Xavier, 

Marcelo Signorelli 

The minimum sum-of-squares clustering problem solved by Hyperbolic 

Smoothing Clustering Method adopts a smoothing strategy that solves a sequence 

of low dimension differentiable unconstrained optimization problems. 

We propose the GPU Hyperbolic Smoothing Clustering Method which contains 

parameters to control the balance of data-processing and data-transfer between 

CPU and GPU. The influence of the parameters of our method is investigated 

about its influence on performance. 

3 - Modelling the conditional variance of wind farm output 

utilising realised volatility 

John Boland, School of Mathematics and Statistics, University of 

South Australia, Mawson Lakes Blvd., 5095, Mawson Lakes, 

South Australia, Australia, john.boland@unisa.edu.au 

In time series analysis, independence means the series and the squared series 

show no autocorrelation. If there is no autocorrelation of the series but there is 

of the squared series, then weak dependence. This occurs when one examines 

wind farm output with for forecasting the level of output. To construct error 

bounds on the forecast when one has high frequency data also available, realised 

volatility is used to estimate the variance. We model 10 second data as 

an AR(3) process to derive an expression for the variance at a 5 minute time 

scale in terms of the autoregressive coefficients. 

4 - Investigating Optimal EOL Recall Policy for Environmentally 

Conscious Supply Chain Network 

Vildan Ozkir, Barbaros Bulvarý Yýldýz Teknik Universitesi 

Endustri Muhendisligi Bolumu Beþiktaþ, 34349, Istanbul, 

Turkey, vildanozkir@gmail.com, Huseyin Basligil 

267


Recent european legislations about sustainability and environment, SC problems 

become more complex due to the necessity of considering much more 

decision variables. As new technology generation life cycle become shorter, 

consumption rate of valuable resources increases. These huge valuable rubbish 

heaps yield considerable amount of economic loss and environmental damage. 

Shortly, this study investigates the most profitable way of environmentally conscious 

SC design. We propose a MILP model including the idea of collecting 

more end-of-life products by adjusting incentive prices. 

� WC-41 


3.1.06 

Forestry Management and Long Term 

Financial Decisions 



Chair: Klaus Seeland, Institute for Environmental Decisions IED, 

ETH Zürich, Natural and Social Science Interface, Sonneggstr. 33, 

SOL E 6, CH-8092, Zurich, Switzerland, klaus.seeland@env.ethz.ch 

1 - Applying Mean-Variance rules in the allocation of land 

use, to avoid deforestation processes in Southern 

Ecuador 

Baltazar Calvas, Institute of Forest Managment, Technische 

Universität München, Am Hochanger 13, 85354, Freising, 

Bayern, Germany, calvas@forst.wzw.tum.de, Thomas Knoke 

We used Monte Carlo Simulations to generate frequency distribution for the 

NPV of four land use options, addressing prices and productivity risks. Land 

use options (pasture, reforestation, intercropping and silvopasture) were them 

combined to maximise the Reward-to- variability ratio. The optimum combination 

was 17 % pasture, 41 % reforestation, 18 % intercropping and 24 % 

silvopasture. The mixed land use improved NPV and risk compared to the 

conventional single pasture system. 

2 - Urban forest carbon balance optimization - the case of 

Lahti, Finland 

Antti Mäkinen, Department of Forest Sciences, University of 

Helsinki, Latokartanonkaari 7, Helsingin Yliopisto, 00014 , 

Helsinki, antti.makinen@helsinki.fi 

We present a case study of urban forest carbon balance optimization from Lahti 

in Southern Finland. A forest simulation and optimization model was coupled 

with a dynamic soil carbon model for predicting the effect of forest management 

on soil carbon. The optimized management scenarios were (i) maximized 

carbon stock and (ii) minimized carbon release into the atmosphere taking into 

account the wood-based product lifecycle and energy production potential. 

3 - Real option valuation of forest plantation investments 

in Brazil 

Markku Kallio, Department of Business Technology, Helsinki 

School of Economics, Runeberginkatu 14-16, FIN-00100, 

Helsinki, kallio@hse.fi, Markku Kuula 

Decisions concerning investments in the real assets commonly involve uncertainties, 

flexibilities and market imperfections. In case of investments in emerging 

markets, risks relate to political, environmental and economic uncertainties, 

for instance. In this paper, we present how a firm, The Forest Company, may 

analyze eucalyptus pulpwood plantation investments in Brazil employing real 

option valuation. We propose a computational method for valuation of real 

options which allows incomplete and imperfect markets. 

4 - Regional Forest Organizations and their Financial Impact 

on Forestry and Small Forest Owners in Central 

Switzerland 

Klaus Seeland, Institute for Environmental Decisions IED, ETH 

Zürich, Natural and Social Science Interface, Sonneggstr. 33, 

SOL E 6, CH-8092, Zurich, Switzerland, 

klaus.seeland@env.ethz.ch 

The capacity for innovation by the establishment of RO and their impact on 

regional economic development has been investigated by forest expert interviews 

and a questionnaire survey among small forest owners in the Canton of 

Lucerne. The net financial return from their forest holding and marketing effectiveness, 

proportion of certified forest and wood products the RO members 

are significantly better off after only a short period of time as compared to the 

non-organized. 

268 

� WC-42 


3.1.07 

Graph Cleaning 

Stream: Graph Searching and Guarding 


Chair: Pawel Pralat, Department of Mathematics, West Virginia 

University, 26505, Morgantown, WV, United States, 

pralat@math.wvu.edu 

1 - The Fast Searching Problem 

Boting Yang, Dept. of Computer Science, University of Regina, 

3737 Wascana Parkway, S4S 0A2, Regina, Saskatchewan, 

Canada, boting@cs.uregina.ca 

Edge searching is a graph problem that corresponds to cleaning a contaminated 

graph using the minimum number of searchers. We define fast searching as a 

variant of this extensively studied problem. Fast searching corresponds to an 

internal monotone search in which every edge is traversed exactly once and 

searchers are not allowed to jump. We give an introduction to fast searching, 

describe the relationship between graph cleaning and fast searching, and 

present some recent progress in computing fast search numbers. 

2 - Cleaning Graphs with Brushes 

Margaret-Ellen Messinger, Mathematics Department, Mount 

Allison University, New Brunswick, Canada, 

wynn11@hotmail.com 

Initially we assume all edges and vertices of a network are contaminated or 

"dirty’. Brushes are placed on some vertices and at each step, one vertex is 

"cleaned’ whereupon it sends a brush along each dirty incident edge (cleaning 

those edges). A brush may not traverse an already cleaned edge. One challenge 

is to determine the minimum number of brushes needed to clean a graph: 

the brush number of a graph. We’ll explore that challenge and discuss some 

bounds on the brush number. 

3 - Minors for digraphs 

Paul Hunter, Computing Laboratory, Oxford University, Parks 

Road, OX1 3QD, Oxford, United Kingdom, 

paul.hunter@comlab.ox.ac.uk 

What is an appropriate definition for a digraph minor? In this talk I answer 

this question in the most vague way possible, but highlight the difficulties of 

providing anything more definitive. 

4 - Copwin Edge Critical Graphs 

Shannon Fitzpatrick, Mathematics and Statistics, University of 

Prince Edward Island, 550 University Avenue, C1A4P3, 

Charlottetown, PE, Canada, sfitzpatrick@upei.ca 

In the game of Cop and Robber, a single cop tries to apprehend a robber as 

they alternately move along edges of a reflexive graph. Graphs in which the 

cop always wins are called copwin graphs. A Copwin Edge Critical graph, 

with respect to edge addition (deletion), is graph that is not itself copwin, but 

theaddition (deletion) of any edge results in a copwin graph. In this talk, I 

will discuss some of the properties of Copwin Edge Critical graphs and give a 

characterization of those that are also planar. 

� WC-43 


8.2.02 

Sustainable Development in Civil 

Engineering and Multiattribute 



Chair: Tatjana Vilutiene, Dept. of Construction Technology and 

Management, Viulnius Gediminas Technical University, Sauletekio 

ave. 11, LT01001, Vilnius, Lithuania, tatjana.vilutiene@vgtu.lt

1 - Birstonas first eco-town in Lithuania: from the idea to 

the Eco-life Project 

Tatjana Vilutiene, Dept. of Construction Technology and 

Management, Viulnius Gediminas Technical University, 

Sauletekio ave. 11, LT01001, Vilnius, Lithuania, 

tatjana.vilutiene@vgtu.lt, Vaidotas Šarka, Darius Bieksa, Edita 

Šarkien 

The activities in ECO-life project intend to demonstrate the use of a practical 

and participatory approach, which can be used as an outset for selected community 

planning and implementation of sustainable urban areas. The goals of 

communities within ECO-life project cover the integration of specific innovations 

like new decisions for building envelope, new metering and supply system, 

polygeneration technologies and "whole town approach’. The paper also 

analyses the barriers which have to be overcome in Lithuania and possibilities 

to apply the multi-criteria decision aid. 

2 - Multi-criteria assessment of design projects’ managers 

by applying AHP and Additive Rational Assessment 

(ARAS) methods 

Zenonas Turskis, Dept. of Construction Technology and 

Management, Vilnius Gediminas Technical University, 

Sauletekio ave. 11, LT01001, Vilnius, Lithuania, 

zenonas.turskis@st.vgtu.lt, Povilas Vainiunas, Edmundas 

Kazimieras Zavadskas, Jolanta Tamosaitiene 

Construction planning processes are extremely important. Project manager 

must be well experienced in all stages of project implementation. Projects’ 

managers characteristics are considered to be less or more important for the 

effective project. Qualifying of design projects’ managers is a vital part of construction 

process. For managers’ assessment and decision supporting Additive 

Rational Assessment method (ARAS) was applied. The model presented in this 

study shows that the ARAS method aggregated together with the AHP method 

and expert questioning are effective tool. 

3 - Multi-objective stochastic simulation-based optimisation 

applied to supply chain planning 

Liana Napalkova, Department of Modelling and Simulation, 

Riga Technical University, 1, Kalku Street, LV-1658, Riga, 

Vidzeme, Latvia, Liana.Napalkova@rtu.lv, Galina Merkuryeva 

The multi-objective stochastic simulation-based optimisation problem with 

constraints and mixed decision variables is investigated. To solve the problem, 

an approximate Pareto-optimal front is generated. The hybrid two-phase 

optimisation method integrates evolutionary computation and response surfacebased 

methodology. A multi-objective genetic algorithm is used for a global 

search of Pareto-optimal solutions, whereas RSM-based linear search allows 

local improving the solutions. The hybrid method is applied to simulation optimisation 

of multi-echelon cyclic planning parameters. 

4 - Water supply regional management by stochastic programming 

Leonidas Sakalauskas, Statistical MOdelling, Institute of 

Mathematics&Informatics, Akademijos 4, 2600, Vilnius, 

Lithuania, sakal@ktl.mii.lt, Kestutis Zilinskas 

Effective planning of resources management is important factor for facilitating 

socio-economic development and eco-environmental sustainability. Such a 

planning effort is complicated with a variety of uncertain, dynamic and nonlinear 

factors as well as their interactions. In this study, a stochastic quadratic 

programming method is developed for reflecting dynamics of uncertainties of 

water supply systems based on a continuously distributed set of scenarios as 

well as tackling nonlinearities in the objective function to reflect the effects 

of marginal utility on system benefits and costs. The method developed can 

support the analysis of various policy scenarios that are associated with different 

levels of economic consequences when the promised targets are violated. 

The developed method is applied to a case study of planning water resources in 

management and development of regional ecological sustainability. The results 

have been generated and are helpful for decision makers in not only identifying 

desired resources- allocation strategies but also gaining insight into the tradeoff 

between economic objective and eco-environment violation risk. 


� WC-44 


8.2.03 

Dynamical Systems and Mathematical 

Modelling in OR I 

Stream: Dynamical Systems and Mathematical 

Modelling in OR 


Chair: Selma Belen, Mathematics and Computer, CAG University, 

Adana-Mersin Karayolu Uzeri, Yenice-Tarsus, 33800, TARSUS / 

Mersin, Turkey, selmabelen@cag.edu.tr 







1 - Electrical Energy Optimization in Spacecraft Attitude 

Control System Using Proper Configuration of Actuators 

M. Navabi, Shahid Beheshti University, Iran, Islamic Republic 

Of, civil.space.edu@gmail.com, N. Nassiri 

Today’s microsatellites (less than 100 kg in mass) are increasingly considered 

for scientific and commercial purposes. In designing attitude control system 

and choosing actuators is a viable task to consider energy storage, because 

microsatellites do not have abundant energy storage and rely on solar panels 

and rechargeable batteries. Here, a study of an ACS based on optimal control 

theory, has been done to achieve high precision control performance and minimize 

the electrical energy consumption. A numerical example is presented in 

this paper to demonstrate the solution technique. 

2 - Organization and Innovation: a System Dynamics Approach 

Boada Alejandro, Facultad Administracion de Empresas, 

Universidad Externado de Colombia, Calle 12 No. 1-17 Este, 

Bogota, Colombia, alejandro.boada@uexternado.edu.co, Nancy 

Mahecha, Maria Teresa Sequeda, Saldana Carolina 

This research seeks to identify inhibitors and motivator for innovation inside 

the companies; based on general systems theory and complex thought, the Organization 

is presented as a social system consisting of subsystems and interrelations. 

We develop a model, based on the systems dynamics, to represent 

the interaction patterns and behaviours in order to introduce organizational innovation 

in companies. By analyzing and modelling subsystems and their interactions, 

researchers present a different approach to the conventional way of 

relation innovation and organization. 

3 - Superneutrality on the transition path in a cash-inadvance 

model with recursive utility 

Kenji Miyazaki, Faculty of Economics, Hosei University, 4342 

Aihara, 194-0298, Machida, Tokyo, Japan, 

miya_ken@hosei.ac.jp 

This paper investigates whether a change in rates of monetary supply enhances 

rates of capital accumulation in a cash-in-advance monetary model with recursive 

utility. The paper discovers that the sign of the effect of growth rate of 

money on capital accumulation depends not only on the curvature of felicity 

but also on the slope and the curvature of the discounting rate function. When 

the discounting rate decreases with consumption, and the discounting rate and 

felicity is sufficiently concave, then inflation deteriorates capital accumulation 

on the transition path. 

4 - Modeling the Diffusion of Chain Emails 

Alexandros Kainich, National Technical University of Athens, 

15780, Athens, Greece, alexkainich@hotmail.com, John 

Coletsos 

We study the diffusion of chain emails using data from a single e-mail account. 

The social network formed has 15,058 nodes. We propose a way of organizing 

data and of modeling the spread of emails according to the cascade model. 

Still, we apply the algorithm of k-best for identifying 8 nodes that can provide 

the greatest expected distribution and propose a way to improve the speed of 

the program. Finally, we present the social network graph, explain the results 

of the algorithm and describe the usefulness. The search for k-best nodes can 

be very useful in marketing. 

269


� WC-45 


8.2.12 

Category and Shelf Space Management 

Stream: Demand and Supply in Consumer Goods and 

Retailing 


Chair: Heinrich Kuhn, Operations Management, Catholic University 

of Eichstaett-Ingolstadt, Auf der Schanz 49, 85049, Ingolstadt, 

Bavaria, Germany, heinrich.kuhn@ku-eichstaett.de 

Chair: Michael Sternbeck, Logistics, Catholic University of 

Eichstaett Ingolstadt, Auf der Schanz 49, 85049, Ingolstadt, 

michael.sternbeck@ku-eichstaett.de 

1 - Coordinating category management and store operations 

through extended planograms 

Rob Broekmeulen, OPAC, TU Eindhoven, P.O. Box 516, Pav. 

E10, 5600 MB, Eindhoven, -, Netherlands, 

r.a.c.m.broekmeulen@tue.nl 

Almost half of the shelf allocations in planograms of retail stores are insufficient 

from a logistic costs perspective. During the allocation process, category 

managers focus on increasing product variety and profitability, taking 

high shelf availability for granted. Too low allocations require unrealistic instore 

replenishment frequencies, resulting in higher costs and/or lower service 

than expected. We propose a coordination mechanism between category management 

and store operations based on differentiated service levels to address 

this issue. 

2 - Retail demand and supply planning with shelf space 

and price management 

Alexander Hübner, Operations Management, Catholic University 

Eichstaett-Ingolstadt, Auf der Schanz 49, 85049, Ingolstadt, 

Germany, alexander.huebner@ku-eichstaett.de, Heinrich Kuhn 

Integrating consumers into demand and supply chain planning is the key lever 

to increase value chain efficiency and performance in retail. Consumers’ demand 

for better service-levels and prices, whereas retailers respond with increasing 

product variety, becoming more price competitive and striving towards 

higher service levels. These have greatly increased the complexity of managing 

the business. We develop a corresponding demand and supply chain planning 

framework. With price optimization we will exemplify the interrelated planning 

issues, to set price levels, and allocate shelf space. 

3 - Shelf Space Driven Assortment Planning for Seasonal 

Consumer Goods 

Joern Meissner, Management Science, Lancaster University 


Kingdom, j.meissner@lancaster.ac.uk, Kevin Glazebrook 

We considers the operations of a "fast-fashion" retailers. Zara and others have 

developed and invested in merchandize procurement strategies that permit lead 

times as short as two weeks. Our research focuses on the use of the most 

valuable resource of such a retailer: shelf space. We investigate the use of 

multi-armed bandits to model the assortment decisions under demand learning. 

The learning aspect is captured by a Bayesian Gamma-Poisson model. We propose 

a knapsack based index heuristic that results in policies that are close to 

theoretically derived upper bounds. 

4 - Simulating shoppers‘ behavior for managers‘ training. 

Maximo Bosch, Industrial Engineering, University of Chile, 

Republica 701, Santiago, Chile, mbosch@dii.uchile.cl, 

Alejandra Puente 

Modern stores are organized in Categories, each one treated as an independent 

business unit. The Category Manager is in charge of pricing, spacing, assortment, 

and promotional decisions. A simulation game based on shoppers’ behavior 

under changes in these decisions was developed to support the training 

of Category managers. This is a noncompetitive web based simulator incorporating 

recent shoppers‘ behavior theory. 

270 

� WC-46 


8.2.14 

Numerical methods for derivatives pricing 

and hedging 

Stream: Numerical Methods in Finance 


Chair: Michèle Vanmaele, Applied Mathematics and Computer 

Science, Ghent University, Krijgslaan 281, S9, 9000, Gent, Belgium, 

michele.vanmaele@ugent.be 

1 - Pricing American Call Options under the Assumption of 

Stochastic Dividends 

Susanne Kruse, S-University of Applied Sciences, Hochschule 

der Sparkassen-Finanzgruppe, Simrockstr 4, 53179, Bonn, 

Germany, susanne.kruse@dsgv.de, Marlene Mueller 

In financial mathematics stock prices are usually modeled directly as a result 

of supply and demand and under the assumption that dividends are paid continuously. 

In contrast economic theory gives us the dividend discount model 

assuming that the stock price equals the present value of its future dividends. 

These two models need not to contradict each other - in their paper Korn and 

Rogers (2005) introduce a general dividend model preserving the stock price to 

follow a stochastic process and to be equal to the sum of all its discounted dividends. 

In our talk we specify the model of Korn and Rogers in a Black-Scholes 

framework in order to derive a closed-form solution for the pricing of American 

Call options under the assumption of a known next dividend followed by 

several stochastic dividend payments during the option’s time to maturity. 

2 - Hedging in the interest rate market and the commodity 

market 

Nele Vandaele, Applied Mathematics and Computer Science, 

Ghent University, Krijgslaan 281 (S9), 9000, Gent, Belgium, 

nele.vandaele@ugent.be, Kathrin Glau, Peter Leoni 

The theory concerning quadratic hedging (QH) strategies is rather elaborate, 

while there is almost no literature available on applications to concrete examples. 

In this talk we determine QH strategies for interest rate derivatives and 

for derivatives typical appearing in the commodity market. In the interest rate 

(resp. commodity) market, we determine the (adjusted) delta-hedge and the 

QH strategy for a swaption (resp. an option written on several assets, while we 

can only invest in a weighted combination of the underlyings). A crucial step 

in the numerical implementation is the application of the Fourier transform. 

3 - Vanna-Volga methods applied to FX derivatives: from 

theory to market practice 

Griselda Deelstra, Mathematics, Actuarial Sciences, Université 

Libre de Bruxelles (U.L.B.), Boulevard du Triomphe, CP 210, 

1050, Brussels, Belgium, griselda.deelstra@ulb.ac.be, Frédéric 

Bossens, Gregory Rayée, Nikos Skantzos 

We study Vanna-Volga methods which are used to price first generation exotic 

options in the Foreign Exchange market. They are based on a rescaling of the 

correction to the Black-Scholes price through the so-called ’probability of survival’ 

and the ’expected first exit time’. We offer a justification of the core 

technique for the case of vanilla options and show how to adapt it to the pricing 

of exotic options. Our results are compared to a large collection of indicative 

market prices and to more sophisticated models. Finally we propose a simple 

calibration method based on one-touch prices. 

4 - Local Volatility Pricing Models for Long-dated FX 

Derivatives 

Gregory Rayee, Mathematics, Actuarial Sciences, Université 

Libre de Bruxelles (U.L.B.), 55 chemin de bas ransbeck, 1380, 

Lasne, Belgium, grayee@ulb.ac.be, Griselda Deelstra 

We study the local volatility function associated to a 3-factor pricing model 

for FX derivatives. The spot FX rate is governed by a geometric Brownian 

motion with a local volatility, while domestic and foreign interest rates follow 

a Hull-White one-factor Gaussian model. The model is suitable to price and 

hedge long-dated FX derivatives. We derive the local volatility function and 

obtain different approaches to calibrate the local volatility on the FX option’s 

market. Then, we derive a calibration method for one extension which allows 

the volatility of the spot FX to have stochastic behavior.

� WC-47 


8.2.16 

Data Mining in Early Warning Systems I 

Stream: Data Mining in Early Warning Systems 


Chair: Tobias Klatt, Department of Economic Studies, Göttingen 

University, Platz der Göttinger Sieben 3, 37073, Göttingen, Germany, 

Tobias.Klatt@wiwi.uni-goettingen.de 

1 - A trapezoidal Fuzzy DEMATEL-approach to assess dependencies 

among sub-indicators of a hierarchical disaster 

indicator model for indirect industrial disaster vulnerability 

Michael Hiete, Institute for Industrial Production (IIP), 

Karlsruhe Institute of Technology (KIT), Hertzstrasse 16, 76187, 

Karlsruhe, Germany, michael.hiete@kit.edu, Mirjam Merz, 

Frank Schultmann 

A hierarchical indicator framework to assess the vulnerability of industrial sectors 

against indirect disaster effects is developed for risk management. Fuzzy 

Decision-Making Trial and Evaluation Laboratory methodology (DEMATEL) 

— a method to analyze direct and indirect dependencies between variables — is 

enhanced for trapezoidal membership functions and used to analyze the dependencies 

between the sub-indicators. Correcting the elicited importance weights 

of the indicator model for the dependencies helps to avoid unwanted multiple 

counting of single dimensions within the indicator model. 

2 - An empirical causality procedure for objective early indicator 

identification 

Tobias Klatt, Department of Economic Studies, Göttingen 

University, Platz der Göttinger Sieben 3, 37073, Göttingen, 

Germany, Tobias.Klatt@wiwi.uni-goettingen.de, Klaus Moeller, 

Judith Huelle 

Early indications of environmental changes are crucial to the companies’ success 

under increasing complex and dynamic environments. To overcome subjective 

assessments used in early indicator selection, we propose an objective 

identification procedure to evaluate possible early warning indicators concerning 

their causal relationship with internal planning variables. An empirical application 

reveals that the procedures’ employment reduces the risk of selecting 

insignificant indicators and improves planning accuracy. 

EURO 24 - Lisbon 2010 WD-04 

Wednesday, 14:00 - 15:20 

� WD-02 


3.2.14 







1 - Computational Complexity of Stochastic Programming 

Alexander Shapiro, Industrial and Systems Engineering, Georgia 

Institute of Technology, Atlanta, GA30332, 30332, Atlanta, 

Georgia, United States, ashapiro@isye.gatech.edu 

Stochastic programming is a popular approach to optimization under uncertainty. 

Its origins are going back to pioneering papers of Dantzig(1955) and 

Beale (1955). The traditional approach to solving stochastic programming 

problems is to construct scenarios representing what may happen in the future. 

From a modelling point of view such scenarios can be considered as a 

discretization of the underline (true) stochastic data process. Consequently, 

computational complexity of the obtained optimization problem is determined 

by the number of generated scenarios. Unfortunately the number of scenarios 

needed to approximate the "true" distribution of the data process grows exponentially 

both with increase of the number of random parameters and number of 

stages. A way of dealing with this explosion of the number of scenarios is to use 

randomization approaches based on Monte Carlo sampling techniques. In this 

talk we discuss theoretical and computational aspects of the Monte Carlo sampling 

approach to solving two and multi-stage stochastic programming problems. 

� WD-04 


3.2.13 

Constructive and local search methods 



Chair: Philippe Lacomme, Laboratoire LIMOS, Université Blaise 

Pascal, BP 10125, 63173, Aubière Cedex, France, 

lacomme@sp.isima.fr 

Chair: Fernanda Raupp, DEI, PUC-Rio, Brazil, fraupp@puc-rio.br 

1 - Intensive Local Search Procedure: a new point-to-point 

metaheuristic for constrained continuous global optimization 

Fernanda Raupp, DEI, PUC-Rio, Brazil, fraupp@puc-rio.br, 

Wendel Melo, Marcia Fampa 

We present a new derivative-free point-to-point metaheuristic for constrained 

continuous global optimization called Intensive Local Search Procedure 

(ILSP). It employs classical strategies of metaheuristics for combinatorial optimization 

as well as combined strategies for approaching continuous spaces, 

which are applied in an exploration process in increasingly refined neighborhoods 

of current points. 

We show efficiency of ILSP on a standard set of 13 well-known test problems. 

Further, we compare its performance with some existing metaheuristics on the 

same set of test problems. 

2 - Implementing 2-opt based tabu search on sparse asymmetric 

TSPs 

Sumanta Basu, Operations Management, XLRI, Jamshedpur, 

Room No. 3, TMDC Building, XLRI Campus, C.H. Area (East), 

831001, Jamshedpur, Jharkhand, India, sumanta@xlri.ac.in, 

Diptesh Ghosh 

271

WD-05 EURO 24 - Lisbon 2010 

Traveling salesperson problem (TSP) instances encountered in real-world logistics 

operations are often sparse and asymmetric, and of large size. The literature 

on tabu search heuristics for the traveling salesman problem primarily 

deals with relatively small symmetric TSPs defined on complete graphs. In 

this paper, we provide data structures which speed up the performance of tabu 

search on large sparse asymmetric TSPs. We show that the speed up using our 

data strucures is significant, and allows us to use tabu search on much larger 

instances than those reported in the literature. 

3 - A new greedy approach to the Quadratic Assignment 

Problem (QAP) 

Theodoros Gevezes, Mathematical and Physical Sciences, 

Aristotle University of Thessaloniki, 51124, Thessaloniki, 

Greece, theogev@gen.auth.gr 

A new greedy approach is presented for the QAP, where instead of building a 

solution starting from an empty set, a feasible solution is reached gradually by 

deleting the worst elements. In order to implement this greedy approach, a set 

of necessary and sufficient conditions for a zero-one matrix to be a solution of 

the QAP is proved, while the verification of these conditions requires the solution 

of linear assignment problem with additional constraints. The new greedy 

approach has been incorporated in a Greedy Randomized Adaptive Search Procedure 

(GRASP) with favorable computational results. 

4 - A new heuristic for the minimization of tool switches 

problem 

Edson Senne, Mathematics, UNESP/FEG, Caixa Postal 205, 

12516-410, Guaratingueta, SP, Brazil, elfsenne@feg.unesp.br, 

Antonio Chaves, Horacio Yanasse 

The minimization of tool switches problem (MTSP) seeks a sequence to process 

a set of jobs so that the number of tool switches required is minimized. 

This work presents a new heuristic for the MTSP. This heuristic has a constructive 

phase, which is based on a graph where the vertices correspond the tools 

and exists an arc k=(i,j) binding vertices i and j if tools i and j are necessary 

for the execution of task k, and an improved phase based on the Iterated Local 

Search. Computational results show that the proposed heuristic has better 

performance than other methods from the literature. 

� WD-05 


3.2.16 

Electronics 



Chair: Marc Sevaux, EU/ME, Université de Bretagne Sud - UEB, 

Lab-STICC - CNRS UMR 3192, Centre de Recherche - BP 92116, 

56321, Lorient, France, marc.sevaux@univ-ubs.fr 




1 - Implementation of the UMTS Technology in the GSM existing 

Network: capacity/interference optimization 

Abdellah El Fallahi, logistics & computer sciences, ENSA 

Tetouan, ENSA Tetouan, ENSA, 93000, Tétouan tanger, 

Tétouan, Morocco, aelfallahi@gmail.com 

UMTS is the Third Generation mobile telecoms networks which will bring 

video, music and internet services to the cell phone and a range of electronic 

products. This paper deals with the implementation of the UMTS services in 

the existing GSM900 bandwidth. It focuses on the issue of spectrum and capacity 

reduction in UMTS900 deployment area and buffer area, which can be 

seen as an optimization problem, where the decision variable is the number of 

transmitters that assure an acceptable QoS of the GSM service. An efficient 

tabu search method is proposed to solve this problem. 

2 - Exact and metaheuristic approaches for memory cache 

management 

272 

Maria Soto, Lab-Sticc, Universitée de Bretagne Sud, Rue Saint 

Maude, 56100, Lorient, Brittany (Bretagne), France, 

maria.soto@univ-ubs.fr, André Rossi, Marc Sevaux 

Designing embedded systems is more and more challenging as technology empowers 

to integrate an increasing number of functionalities in a microchip. Because 

of this complexity increase, designers rely on CAD software which produce 

chips that often lack optimization. Consequently, the designed microchips 

are too power-consuming. Memory cache management has been shown to have 

a deep impact on performances, thus this work focuses on data allocation to 

memory banks. This issue is shown to be close to the k-weighted graph coloring 

problem and is addressed with graph coloring inspired metaheuristics 

3 - Point Coverage, Sink Location, and Data Routing Problems 

in Wireless Sensor Networks: MILP Formulations 

and Solutions 

Kuban Altinel, Dept. of Ind. Eng., Bogazici Univ, Bebek, 34342, 

Istanbul, Turkey, altinel@boun.edu.tr, Necati Aras, Evren 

Guney, Cem Ersoy 

The point coverage, sink location, and data routing problems are considered 

within a unified frame and two integrated mixed-integer linear programming 

formulations are developed. They are difficult to solve and a hybrid procedure 

is proposed. The best sensor locations are sought by tabu search in the upper 

level. For the fixed sensor locations, the remaining problem of determining 

sink locations and data routes are solved approximately in the lower level. According 

to the experimental results performed on a number of test instances, 

we can say that the new method is efficient and accurate. 

4 - Multi-objective optimization of Memory Built-In-Self- 

Test sharing 

Yann Kieffer, laboratoire G-SCOP, 38031 Cedex 1, Grenoble, 

France, yann.kieffer@g-scop.inpg.fr, Zaourar Lilia, Nadia 

Brauner 

Because of the growing integration of systems-on-chips, testing semiconductor 

on-chip memories is becoming more and more challenging. To 

improve testability and reduce costs, the additional testing circuitry has to be 

shared among memories. This sharing impacts the surface cost of these special 

elements, and also both the test time and test power necessary to test the individual 

chips after production. We present a model and solution for this sharing 

problem, where both sequential and parallel sharing are allowed, as specified 

by compatibility rules given as parameters of the optimization. 

� WD-06 


8.2.30 

DEA Application VII - Retailing and health 



Chair: Stella Sofianopoulou, Industrial Management & Technology, 

University of Piraeus, 80 Karaoli & Dimitriou street, 18534, Piraeus, 

Greece, sofianop@unipi.gr 

1 - Assessment for Teaching and Medical Integrated Activities 

inside Brazilian University Hospitals 

Angela Silva, Economia, Unoiversidade Estadualdo Rio de 

Janeiro, Rua Belisario Tavora 80 ap 506, 22245-070, Rio de 

Janeiro, Rio de Janeiro, angela@pep.ufrj.br, Marcos Estellita 

Lins, Maria Stella Castro Lobo, Roberto Fiszman 

This paper measures the performance of Brazilian Federal University Hospitals 

and the integration between medical and teaching activities. We propose a Network 

DEA model that considers the medical care and the teaching dimensions 

simultaneously, with equivalent weights. We consider an output-oriented VRS 

model. The chosen external input variables were: physicians, teachers (for 

both dimensions), beds and operational expenses (for the medical care dimension); 

the output variables were: medical grad students (teaching dimension) 

and adjusted admissions (medical care dimension) 

2 - The assessment of retailing efficiency using Network 

Data Envelopment Analysis 

Clara Vaz, Escola Superior de Tecnologia e de Gestão, Instituto 

Politécnico de Bragança, Campus de Santa Apolónia, Apartado 

134, 5301 - 857, Bragança, Portugal, clvaz@ipb.pt, Ana 

Camanho, Rui Guimarães

This paper describes a method for the assessment of retail store performance 

based on DEA. The assessment considers the stores aggregate several subunits, 

corresponding to sections with management autonomy. This motivated an analysis 

at the section level and the store level. The performance assessment of 

the sections involves a comparison among similar sections located in different 

stores, and evaluates efficiency spread. This is followed by an analysis at the 

store level to define targets for the sections by using a Network model that takes 

into account the sections share limited resources. 

3 - Analyses of investment efficiency using network DEA 

Hirofumi Amatatsu, Information Science of Graduate School of 

Engineering, Seikei University, 3-3-1 Kichijoji-Kitamachi, 

180-8633, Musasino-shi, Tokyo, Japan, 

amatatsu@mint.ocn.ne.jp, Tohru Ueda 

Enterprises have been investing to keep their positions in markets or to make 

new positions. Central and local governments have also been doing similar 

activities. In this paper we propose DEA algorithms to evaluate efficiency of 

these investments using financial reports and input-output tables for multiple 

periods. The algorithms include 1) Dynamic DEA algorithms for simple organizations 

2) Dynamic DEA algorithms for matrix type network organizations 

and 3) Malmquist productivity indexes. 

4 - Retail chain performance evaluation using Data Envelopment 

Analysis 

Vassilis Dedoussis, Industrial Management & Technology, 

University of Piraeus, 80 Karaoli & Dimitriou str., 185 34, 

Piraeus, Greece, vdedo@unipi.gr, Stella Sofianopoulou 

The efficiency of a retail chain is a major issue in the retailer’s competitiveness, 

since its profitability depends on the profitability of its parts. Data Envelopment 

Analysis is employed for resolving this problem in a fast food retail chain and 

assessing managerially useful measures of store-level retail productivity. The 

mathematical model created is multidimensional and accepts multiple inputs 

and outputs, both quantitative and qualitative for every outlet, which are then 

used as tools for measuring the technical efficiency of the stores. Computational 

results from a real-world test case are presented. 

� WD-07 


8.2.47 

Project Scheduling 

Stream: Project Management and Scheduling [c] 


Chair: Premysl Sucha, Department of Control Engineering, Czech 

Technical University, Faculty of Electrical Engineering, Karlovo 

namesti 13, Prague 2, 121 35, Prague, Czech Republic, 

suchap@fel.cvut.cz 

1 - Time-cost tradeoffs under time and cost chance constraints 

Zohar Laslo, Industrial Engineering and Management, SCE- 

Shamoon College of Engineering, Bialik/Bazel Strs, 84100, Beer 

Sheva, Israel, zohar@sce.ac.il, Gregory Gurevich 

Seeking for the improvement of the project planning, we analyzed two current 

procedures and a new developed procedure for crashing the project completion 

by additional budget. We consider a project with various types of activities 

where the randomness of their duration derives from external uncertainty, internal 

uncertainty or both of them and where correlation between their actual cost 

and random duration is known. The objective is to ’optimize’ the allocation of 

budget among project activities, seeking to minimize the budget, subject to any 

chance constrained contractual due date. 

2 - Social Network Analysis of project partnership 

Blazenka Divjak, Faculty of organization and informatics, 

University of Zagreb, Pavlinska 2, 42 000, Varazdin, 

blazenka.divjak@foi.hr, Nina Begicevic, Petra Peharda 

In this paper we present a research on the social network of project partnership 

in the EUREKA network. The main method used is the Social Network 

Analysis. Two hypotheses were set: H1 Countries from the same region cooperate 

more among themselves than with the countries from any other region. 

H2 Central countries in the social network of partnership in EUREKA projects 

are developed countries. We tested the hypotheses on countries from Northern, 

Central, Mediterranean, Western and South-Eastern and Eastern Europe. The 

results indicate that countries from the same region cooperate more with the 

countries outside the region and that the central countries in the social network 

are developed countries. 


3 - Flexible Human Resource Management through Decision 

Support Systems - A Case-Study in a Contact Center 

Outsourcer 

João Justino, HR Department, Teleperformance Portugal, R 

Alexandre Braga 25-B, 1150-003, Lisboa, Portugal, 

jj1165@acuity.pt, Joao Telhada 

Managing the relationship with employees, while keeping high levels of quality 

in customer satisfaction, has been a continuous challenge for business managers. 

Usually this relationship is strict, thus introducing flexibility is a breakthrough 

innovation that can only be achieved in a business intelligence framework. 

A case study is presented describing the design, adoption and implementation 

of a intelligence system in a contact center outsourcer for managing the 

usage of operators in a flexible way. Some indicators are given that illustrate 

the importance of such systems. 

4 - Take-give Resources in Project Scheduling with Time 

Windows 

Premysl Sucha, Department of Control Engineering, Czech 

Technical University, Faculty of Electrical Engineering, Karlovo 

namesti 13, Prague 2, 121 35, Prague, Czech Republic, 

suchap@fel.cvut.cz, Zdenek Hanzalek 

The problem that we address in this work is motivated by a real scheduling 

problem from a lacquer production which is seen as the project scheduling 

problem with general temporal and resource constraints. In addition, there are 

special resources called take-give resources that are needed from the beginning 

of an activity to the completion of another activity. In addition, we consider 

sequence dependent changeover time on take-give resources. We suggest two 

heuristic solutions to solve the problem. Performance of both heuristics is evaluated 

on a set of lacquer production benchmarks. 

� WD-08 


6.1.36 


Scheduling II 



Chair: Erwin Pesch, FB 5 - Institute of Information Systems, 

University of Siegen, Hoelderlinstr. 3, 57068, Siegen, Germany, 





Chair: Edmund Burke, School of Computer Science & IT, University 

of Nottingham, Jubilee Campus, Wollaton Road, NG8 1BB, 

Nottingham, United Kingdom, ekb@cs.nott.ac.uk 

1 - Evaluation and comparison of project management 

software 

Vassilis Kostoglou, Department of Informatics, Alexander TEI of 

Thessalon iki, P.O. Box 141, 57400, Thessaloniki, Greece, 

vkostogl@it.teithe.gr 

Most projects are characterized by complexity due to their size, the requirement 

for scheduling of tasks and tracking of progress, and the dire need for 

using their resources efficiently. A large number of relevant software is available 

to project managers. This work examines thoroughly and evaluates 12 selected 

project management programmes on six introduced main criteria, each 

consisting of several components. All programmes are tested and ranked for 

every criterion and aggregately according to their scores on a five grade scale. 

Software performances are commented and conclusions are drawn. 

2 - On Resource Complementarity in Activity Networks — 

Preliminary Results 

Helder Silva, IFAM - Instituto Federal de Educação, Ciência e 

Tecnologia do Amazonas, Rua Governador Danilo Areosa S/N, 

273


69075351, Manaus, Amazonas, Brazil, helder@ifam.edu.br, 

Anabela Pereira Tereso, José Oliveira 

The methodology of project management has been widespread in organizations 

of different functions and sizes. In this context, we address the issue of optimal 

resource allocation, and more specifically, the analysis of complementarity of 

resources (primary resource and supportive resource) in a project. We develop 

a conceptual system capable of determining the ideal mixture of resources allocated 

to the activities of a project, such that the project is completed on time 

with minimal cost. In this paper, we present the mathematical model, development 

details and the preliminary results obtained. 

3 - Forming a balanced team 

Stanislaw Walukiewicz, Systems Research Institute, Polish 

Academy of Sciences, ul. Newelska 6, 01-447, Warsaw, Poland, 

Stanislaw.Walukiewicz@ibspan.waw.pl 

We formulate the problem of formation an effective team, using the terms following 

Belbin. Next, we reformulate the approach using the concept of social 

capital and proximity. By social capital we understand formal and informal 

relations between at least two people. As they are not disjoint, we study proximity 

- specific interrelations linking people solving a given problem. We show 

that there are four forms of proximity and that they are mutually disjoint. We 

compare our approach to the existing balancing methods. 

4 - Modular Operational Support System (OSS) to Cutting 

Processes Parameters Optimization 

Elesandro Baptista, Industrial Engineering Post Graduation 

Program, UNINOVE-Nove de Julho University, Av. Francisco 

Matarazzo, 612, 05001-100, São Paulo, São Paulo, Brazil, 

elesandro@uninove.br, Nivaldo Coppini, Êmili Haguihara 

Cutting processes management has great value to industrial companies. Normally, 

currently used operational conditions are based on tool maker’s catalogs 

and on the operator experience. It is undisputed that each manufacturing scenario 

has its own characteristics and the optimized conditions can be achieved 

mainly with collected data from their own shop floor experience. The aim of 

this paper is to develop a project for a modular Operational Support System 

(OSS) that allows the user to optimize cutting process parameters in its own 

manufacturing scenario in real time with the process. 

� WD-09 


6.2.53 

Convex Duality in Mathematical 

Programming 



Chair: Sorin-Mihai Grad, Faculty of Mathematics, Chemnitz 

University of Technology, 09107, Chemnitz, Sachsen, Germany, 

grad@mathematik.tu-chemnitz.de 




1 - On a zero duality gap result in extended monotropic 

programming 

Radu Ioan Bot, Faculty of Mathematics, Chemnitz University of 

Technology, Reichenhainer Str. 39 Zi. 612, 09107, Chemnitz, 

Germany, bot@mathematik.tu-chemnitz.de, Ernö Robert 

Csetnek 

In this presentation we discuss and improve a zero duality gap statement given 

in [D.P. Bertsekas, Extended monotropic programming and duality, Journal of 

Optimization Theory and Applications 139 (2), pp. 209-225, 2008] for an extended 

monotropic programming problem. To this aim we use some convex 

analysis specific techniques based on subdifferential calculus, whereby a determinant 

role is played by a generalization of the Hiriart-Urruty - Phelps formula. 

2 - Duality for vector optimization problems via a general 

scalarization 

274 

Sorin-Mihai Grad, Faculty of Mathematics, Chemnitz University 

of Technology, 09107, Chemnitz, Sachsen, Germany, 

grad@mathematik.tu-chemnitz.de, Radu Ioan Bot 

Considering a vector optimization problem to which properly efficient solutions 

are defined by using convex cone-monotone scalarization functions, we attach 

to it, by means of perturbation theory, new vector duals. When the primal problem, 

the scalarization function and the perturbation function are particularized, 

different dual vector problems are obtained, some of them already known in the 

literature. Weak and strong duality statements are delivered in each case. Thus 

we extend to a more general framework the results presented in our contribution 

to the previous EURO Conference from Bonn. 

3 - Conjugate Duality in Multiobjective Optimization 

Gert Wanka, Faculty of Mathematics, Chemnitz University of 

Technology, Chemnitz University of Technology, Faculty of 

Mathematics, D-09107, Chemnitz, Germany, 

gert.wanka@mathematik.tu-chemnitz.de 

This paper is devoted to some concepts and results regarding duality for convex 

vector optimization problems. We consider problems with and without 

geometric and cone constraints and deal with Lagrange, Fenchel and Fenchel- 

Lagrange duality. We establish weak, strong and converse duality results accompanied 

by necessary and sufficient optimality conditions. 

References 

R. I. Bot, S.-M. Grad, G. Wanka, Duality in vector optimization. Springer- 

Verlag Berlin Heidelberg 2009 

4 - A new and condensed linearization algorithm for an important 

branch of Linear Fractional Programming Problem 

Reena Kapoor, mathematics, delhi university, j-5/4a krishna 

nagar, Delhi, 110051, Delhi, Delhi, India, 

reena_3105@rediffmail.com 

This paper deals with Linear Fractional Programming Problem with bounded 

variables under the restriction: absolute value of the denominator is greater than 

or equal to M0. An example is constructed to show importance of the problem 

in real world. One way to solve the problem is: divide it into two separate 

LFPP and apply Charnes-Cooper Method. The underlying contribution of this 

paper is: it provides a compact process that linearize the considered problem at 

one go. Comparative study of two techniques is also given. An example and 

the codes and data sets for the procedure are given in the end. 

� WD-11 


8.2.38 


Technology IV 









1 - Verification of the validity on the pattern of Royal ecustomer 

purchasing behavior 

Masashi Nakagawa, web business technology, The kyoto college 

of graduate studies for informatics, 606-8225, kyoto, Japan, 

nakagawa.3719@gmail.com, Hong Seung Ko, Shasha Liu, 

Megumi Hiramoto, Takeshi Teramoto 

The visualization of e-customer behavior is a very key issue in e-business. It is 

needed to understand the e-customer behavior process and the purchase pattern 

to visualize an e-customer behavior. In this paper, we will check up that the 

e-customer who must be retained and took the behavior pattern of following 

up the 7 steps of behavior process brings out the profitable sales to a company. 

Consequently, we verify that the profitable sales are increased by the 

most valuable e-customer taking the 7 steps through the correlation analysis 

and the causal analysis.

2 - Verification of the validity on the 7 steps of e-customer 

behavior 

Haruki Takada, The kyoto college of graduate studies for 

informatics, 606-8225, kyoto, Japan, htakada@websui.com, 

Hong Seung Ko 

In online way, the e-customer value to create a long term profits becomes a very 

important element. For this evaluation on the Internet / through e-customer behavior 

process, we propose the 7 steps of royal e-customer behavior process 

to be developed. We verify the effectiveness and the validity of the 7 steps 

by nomological validity and causality analysis. Furthermore, for measuring ecustomers’ 

value we consider motives which are appraisal criteria and factors 

concerning with following up the 7 steps. 

3 - Quantile Analysis for Multidimensional Acceptance Intention 

of E-reader 

Tzyy Jane Lai, Department of International Business„ Yuan Ze 

University, !35 Yuan Tung Rd. ChungLi, 320, TaoYuan, Taiwan, 

Taiwan, fntjlay@saturn.yzu.edu.tw, Hsien-Tung Tsai, Ya-Ling 

Chiu 

This study incorporates three different theories to build a multidimensional 

framework and use quantile regression to investigate the dynamic effects of 

determinants on users that have different level of acceptance intentions. Firstly, 

we extends the TAM and TPB model and then incorporate the social-identity 

expressiveness effects of technology adoption which includes cosmopolitanism 

and global identification. The study argues that the proposed framework can 

provide a more comprehensive picture in describing users’ acceptance intension 

of e-reader. 

� WD-12 


8.2.39 

AHP 08 


Processes 


Chair: Eizo Kinoshita, urban sicience, Meijo University, 



1 - Evaluation of Information Security and Efficient System 

Using Dominant AHP 

Norikumo Shunei, Management, General Management 

department, 12-5 Harayama, Okamachi, 444-0005, Okazaki, 

Aichi, Japan, norikumo@asu.ac.jp, Eizo Kinoshita 

In this study, I introduce a concept of the SA (the security architecture) that 

developed from a concept of EA( enterprise architecture) into the evaluation of 

information security service. Using the framework and the characteristic of the 

hierarchy model, it combines dominant AHP and suggests the technique of the 

new system optimization and promotion of efficiency. 

2 - Determining MDF & particleboard’s marketing effective 

criteria 

Majid Azizi, wood and paper sciences and technology, Faculty of 

natural resources, University of Tehran, karaj, tehran, Iran, 

Islamic Republic Of, mazizi@ut.ac.ir, Fatemeh Sarfi, Mehdi 

Faezipour, Amin Arian 

Goal of the research is determining MDF and particleboard effective criteria 

with respect to customer preferences. AHP has been applied to determine 

weighing values of the criteria. First step a decision tree is categorized in three 

major groups: qualitative, technical & technological and marketing criteria. In 

second step questionnaire prepared, distributed and gathered responses from 

the customers. Final results showed the highest priority of criteria for particleboard 

and MDF are nail and screw holding ability and machine ability of the 

panel, respectively. 

3 - A Super Pairwise Comparison Matrix in Dominant AHP 

Takao Ohya, School of Science and Engineering, Kokushikan 

University, 4-28-1 Setagaya, 154-8515, Setagaya-ku, Tokyo, 

Japan, takaohya@kokushikan.ac.jp, Eizo Kinoshita 


In this study, we propose the super pairwise comparison matrix (SPCM) which 

is a pairwise comparison matrix whose element is the relative ratio from a pairwise 

comparison of weights of the combination of alternatives and criteria. Furthermore, 

in the case of dominant AHP, we prove that the weights calculated by 

the logarithmic least square method to the SPCM equal the weights calculated 

by geometric mean method to the ordinary pairwise comparison matrices. 

4 - A Comparison of Dominant AHP/CCM and AHP/ANP 

Eizo Kinoshita, urban sicience, Meijo University, 



The Theory of Games is a conflict descriptive type model designed to minimize 

one’s loss. AHP, on the other hand, is a conflict solving type model and offers 

a method to describe which element in the conflict is more critical. This paper 

analyze the comparison of Dominant AHP/CCM (Concurrent Convergence 

Method), proposed by Kinoshita and Nakanishi, with AHP/ANP, proposed by 

Saaty and present the calculation methods and the mathematical structure of 

the former in the process.kkk 

� WD-13 


2.2.21 

Decision support for practical logistics 

problems 



Chair: Hokey Min, Management, Bowling Green State University, 

Dept of Management, 3008C, Bowling Green State University, 

43403, Bowling Green, Ohio, United States, hmin@bgsu.edu 

1 - An Intelligence Decision Support System for Paratransit 

Service Vehicle Routing and Scheduling 

Hokey Min, Management, Bowling Green State University, Dept 

of Management, 3008C, Bowling Green State University, 43403, 

Bowling Green, Ohio, United States, hmin@bgsu.edu, Emanuel 

Melachrinoudis 

In this paper, we propose an intelligent decision support system for solving 

multi-depot, multi-vehicle, and capacitated call-a-ride paratransit service problems 

with sliding time windows. The proposed system has been applied to real 

problems encoutering the Toledo Area Regional Paratransit Services (TARPS) 

in the United States. 

2 - Truck Routing and Driver Scheduling under Hours of 

Service Regulations 

Emanuel Melachrinoudis, Mechanical and Industrial 

Engineering, Northeastern University, 360 Huntington Avenue, 

2115, Boston, MA, United States, emelas@coe.neu.edu, Hokey 

Min 

Hours-of-service (HOS) regulations are intended to ensure truck drivers get 

adequate rest and perform safe operations. However, restricting the driver’s 

consecutive driving hours and expanding off-duty hours, combined with road 

traffic and other delays may lead to substantial transportation cost increases. A 

model that determines the route of a single truck and its driver’s schedule under 

HOS and time-dependent vehicle speed is developed that minimizes total trip 

time. The model is solved by a heuristic procedure and is tested 

3 - Locating a Semi-Obnoxious Facility with Mixed Distance 

Metrics 

Emre Yavuz, Mechanical and Industrial Engineering, 

Northeastern University, 235 Park Dr. Apt 20, 02215, Boston, 

MA, United States, emreyavuz999@gmail.com, Emanuel 

Melachrinoudis 

Semi-obnoxious facilities are those facilities that both provide a service to the 

community and have adverse effects on the people and the environment, such as 

power plants, chemical plants, airports, incinerators and waste dumps. These 

facilities should be located far away from population centers, but not too far 

away in order to contain transportation costs. A mixed distance metric bicriteria 

model is developed to locate a semi-obnoxious facility on a network by 

using network distances for transportation and Euclidean distances for undesirable 

effects and it is illustrated in an example involving the location of a 

semi-obnoxious facility in the Bursa Province of Turkey. 

275


4 - Theory of constraints for closed-loop supply chain 

management 

Hokey Min, Management, Bowling Green State University, Dept 

of Management, 3008C, Bowling Green State University, 43403, 

Bowling Green, Ohio, United States, hmin@bgsu.edu 

Close-loop supply chain operations often require dramatic changes in business 

paradigms that include the coordination of both forward and reverse supply 

chain flows, the substitution of disposal for salvage, and the tracing of products 

throughout their entire life cycles. These changes are often preceded by cultural 

changes across the supply chain. To embark on the successful change in 

closed-loop supply chain operations, this paper proposes a theory of constraints 

that can map necessary changes in the close-loop supply chain and guide those 

changes within various system constraints. 

� WD-14 


2.2.15 

Stochastic methods in actuarial sciences 



Chair: Griselda Deelstra, Mathematics, Actuarial Sciences, 

Université Libre de Bruxelles (U.L.B.), Boulevard du Triomphe, CP 

210, 1050, Brussels, Belgium, griselda.deelstra@ulb.ac.be 

1 - On a discrete time risk model with interest 

Maude Gathy, département de mathématique, service de 

Probabilité, Université Libre de Bruxelles, Boulevard du 

Trimphe, CP 210, 1050 , Bruxelles, Belgium, mgathy@ulb.ac.be, 

Claude Lefèvre 

This work is concerned with a discrete time model where an insurer receives 

fixed premiums and pays the claims amounts respectively at the beginning and 

at the end of the time periods. Non-null rates of interest are also incorporated. 

An exponential bound for the ruin probability over any finite time horizon is 

first provided. It depends on the length of the time horizon via an adjustment 

coefficient. The possibility for the insurer to use proportional reinsurance is 

then investigated. 

2 - Local risk minimization in case of stochastic volatility 

and jumps 

Michèle Vanmaele, Applied Mathematics and Computer Science, 

Ghent University, Krijgslaan 281, S9, 9000, Gent, Belgium, 

michele.vanmaele@ugent.be, Nele Vandaele 

The locally risk-minimizing hedging strategy for a risky asset whose price process 

is a continuous semimartingale is obtained from the risk-minimizing hedging 

strategy under the minimal martingale measure. However in the discontinuous 

case, this is no longer true. We will explain how one can determine the 

locally risk-minimizing hedging strategy in this latter case. As applications 

we will derive LRM hedging strategies for stochastic volatility models and for 

unit-linked life insurance contracts where the underlying asset is driven by a 

Lévy process. 

3 - Absolute ruin in the insurance risk model of Ornstein- 

Uhlenbeck type 

Ronnie Loeffen, Weierstrass Institute (WIAS), Mohrenstrasse 39, 

10117, Berlin, Germany, loeffen@wias-berlin.de, Pierre Patie 

We consider the classical risk process where the company earns interest on positive 

surplus and pays interest (at the same rate) when the surplus is negative. 

In this model, the company can get absolutely ruined, which is the event where 

the premium income can no longer compensate the interest payments. We derive 

simple expressions for the Laplace transform in space of both the finite 

and infinite time absolute ruin probability as well as the two-sided exit problem 

(related to the absolute ruin level). These Laplace transforms can be inverted 

for specific cases of the claim size distribution. 

4 - Backward stochastic differential equations and 

stochastic control: A direct approach and an application 

to portfolio optimization 

276 

Xavier De Scheemaekere, Solvay Brussels School of Economics 

and Management, Université Libre de Bruxelles (U.L.B.), Av. 

F.D. Roosevelt, 50, CP 145/1, 1050, Brussels, Belgium, 

xdeschee@ulb.ac.be 

This paper investigates the connection between backward stochastic differential 

equations (BSDEs) and Hamilton-Jacobi-Bellman partial differential equations 

(HJB PDEs) in stochastic control. Like HJB PDEs, BSDEs can be used to express 

the optimal solution of a stochastic control problem directly. Moreover, 

this stochastic approach naturally extends to the nonlinear case, i.e. when the 

payoff function is expressed in terms of a nonlinear conditional expectation. 

These results are applied in finance to solve linear and nonlinear dynamic portfolio 


� WD-15 


2.2.12 

Traveling salesman problems 



Chair: Ana Paias, DEIO/CIO, University of Lisbon, Portugal, 

ampaias@fc.ul.pt 

1 - Towards A Learning-based Heuristic Searching Reform 

Scheme 

Fan Xue, Department of Industrial and Systems Engineering, 

Hong Kong Polytechnic University, Honghom, N/A, Kowloon, 

Hong Kong, dewolf.xue@polyu.edu.hk, C.y. Chan, W.h. Ip, C.f. 

Cheung 

We investigate the issue of improving heuristic searching with supervised learning 

in large scale optimization. We noticed the "self-similarity’ in many problems 

and present an efficient model that can learn "patterns’ from a small subproblem 

and amend the heuristics in the given problem. Experiments were 

conducted on large-scale benchmark problems, including the Euclidean Traveling 

Salesman and the Staff Rostering. We find the results encouraging as 

we hope it unveils a promising direction of taking advantage of the power of 

machine leaning in large-scale optimization. 

2 - Sampling-based Heuristics for the Probabilistic Traveling 

Salesman Problem with Deadlines 

Dennis Weyland, IDSIA, Galleria 2, 6928, Manno - Lugano, 

Switzerland, dennis@idsia.ch 

The Probabilistic Traveling Salesman Problem with Deadlines is a Stochastic 

Vehicle Routing Problem with a computationally demanding objective function. 

We present an approximation for the objective function using Monte Carlo 

Sampling and use this approximation in some heuristics. Those heuristics are 

comparable with state-of-the-art approaches in terms of solution quality, while 

requiring significantly less computational time. 

3 - Solving the Multicolor TSP 

Emanuele Tresoldi, Università Statale di Milano, 20100, Milano, 

emanuele.tresoldi@unimi.it, Roberto Wolfler-Calvo, Sylvie 

Borne 

The problem adressed in this paper is a new interesting variant of the classical 

asymmetric traveling salseman problem. The set of nodes to visit is obtained 

by joinning clusters of nodes, each one characterized by a color. Two nodes 

of the same color must be separted in the optimal sequence by at least H and 

at most K nodes of the other colors. The problem consists in finding the optimal 

hamyltonian cycle, respecting all the different separation constraints. We 

present formulations, exacts and heuristics approaches. The computational results 

prove the efficacity of the proposed algorithms. 

4 - Integer Linear Programming and Dynamic Programming 

approaches for a Traveling Purchaser Problem 

with Additional Side-Constraints 

Ana Paias, DEIO/CIO, University of Lisbon, Portugal, 

ampaias@fc.ul.pt, Luis Gouveia, Stefan Voss 

We study the traveling purchaser problem with a limit on the maximum number 

of markets to be visited, a limit on the number of items bought per market and 

where only one copy per item needs to be bought. We present an ILP model 

and several variations of a Lagrangian relaxation combined with a subgradient 

optimization procedure. The relaxed problem is solved by dynamic programming 

(DP) as it can result from applying a state space relaxation technique to a 

DP formulation for the problem. Computational results show the effectiveness 

of the methods.

� WD-17 


1.3.14 

Data Mining Tools and Improvements 



Chair: Fatma Yerlikaya Ozkurt, Scientific Computing, Institute of 

Applied Mathematics, Industrial Engineering Department, Middle 

East Technical University, 06531, Ankara, Turkey, 

fatmayerlikaya@gmail.com 

Chair: Elcin Kartal Koc, Statistics, Middle East Technical University, 

Department of Statistics, No:234, 06531, Ankara, Turkey, 

kartalelcin@gmail.com 

1 - Parameter Estimation for Semiparametric Models with 

CMARS and Its Applications 

Fatma Yerlikaya Ozkurt, Scientific Computing, Institute of 

Applied Mathematics, Industrial Engineering Department, 

Middle East Technical University, 06531, Ankara, Turkey, 

fatmayerlikaya@gmail.com, Pakize Taylan, Gerhard-Wilhelm 

Weber 

The particular form of Generalized Linear Models are the semiparametric models 

in that the usual parametric terms are augmented by a single nonparametric 

component. The advantage of semiparametric models consists in some grouping 

(linear and nonlinear or parametric and nonparametric) that could be done 

for the features to assign appropriate submodels. We apply CMARS, constructed 

by conic quadratic programming, for the nonparametric part. For the 

parametric part, least square estimation with Tikhonov regularization is used. 

The applications of this study will be analyzed. 

2 - An ACO-based Clustering Algorithm 

Tulin Inkaya, Industrial Engineering, Middle East Technical 

University, METU Industrial Engineering Department, 06531, 

Ankara, Turkey, tulin@ie.metu.edu.tr, Sinan Kayaligil, Nur Evin 

Ozdemirel 

Aim of the proposed ACO-based clustering algorithm is to generate a set of 

non-dominated solutions which includes the target clusters. Ants form a network 

by placing edges between points that are in the same cluster. In the preprocessing 

step, neighborhoods of points and subclusters are formed in a graph 

theoretic context. In ACO, only the points on the subcluster boundaries are 

used in clustering to ensure computational efficiency and ACO focuses on outlier 

detection and merging operations. Performance of the proposed algorithm 

is tested on various data sets. 

3 - Designing control limits of average control chart using 

subgroups for non-normal processes 

Shih-Chou Kao, Graduate School of Operation and Management, 

Kao Yuan University, No.1821, Jhongshan Rd., Lujhu Township, 

821, Kaohsiung County, Taiwan, kaosc@cc.kyu.edu.tw 

Most articles determined the related constant values of an average control chart 

based on simple sizes, not subgroups for the skewed distributions. Rare researches 

discussed the influence of subgroups on monitoring for the skewed 

distributions and determined the suitable constant values of an average control 

chart. The study determined the constants of the average control chart by using 

a simulation method and fixing the probability of type I that is 0.0027 with lognormal 

distribution to construct the average control chart. Furthermore, compared 

the probabilities of type I and type II errors among the control charts, 

including the weighted variance (WV), skewed correction (SC) and traditional 

Shewhart control charts, the proposed control chart is superior to them, in terms 

of the two probabilities for a skewed process. 


� WD-18 


1.3.15 

Data Mining for Credit Scoring 

Stream: Applications in Business Intelligence and 

Knowledge discovery 


Chair: Hsin-Vonn Seow, Business School, University of Nottingham- 




1 - Recovery Rate Modelling for Unsecured Consumer 

Credit: Watching early repayment behaviour pays 

Jie Zhang, School of Management, University of Southampton, 

Quantitative financial risk management centre, Building 2, 

Highfield Campus, Southampton, SO17 1BJ, Southampton, 

England, United Kingdom, jz3g08@soton.ac.uk, Lyn Thomas 

Estimating Recovery Rate (RR) has become more important in consumer credit 

because of the new Basel Accord regulation and the increase in number of defaulters 

due to the recession. We examine whether short-term RR is helpful in 

modelling final RR. It is found that 12-month RR and 24-month RR are very 

significant variables in final RR prediction models. Thus, we look at two-stage 

models, where in stage one, short-term RR is predicted, and then the predicted 

short-term RR is used in the overall RR prediction models. 

2 - Transactors and Pricing 

Hsin-Vonn Seow, Business School, University of Nottingham- 



Hsin-Vonn.Seow@nottingham.edu.my, Lyn Thomas 

The saturated condition of the market for personal financial products requires 

financial institution to look at ways of increasing the acceptance rate of their 

products. We take the example of credit cards and varying the interest rates on 

them to make it more attractive to potential consumers. To achieve that, we 

advocate including a Transactor Score in the model to find a variant of interest 

rate on the card to offer that would have a high acceptance take from the 

potential consumer 

3 - Improving a Bank’s Quality of Products & Services and 

its Attractiveness by using systems & tools from manufacturing 

operations 

Maria Mavri, Business Administration, University of the 

Aegean, 8 Michalon Street, 82100, Chios, Greece, 

m.mavri@ba.aegean.gr, Vassilis Angelis, Katerina Dimaki, 

George Konstantas 

The ability of a bank to attract customers depends on its Image which has already 

been defined as a function of a multitude of factors. The present work 

assumes that a bank can effectively be represented as a manufacturing operation 

and uses tools and systems from this area to map out, assess and redesign 

the bank’s processes so as to improve its effectiveness and efficiency. Furthermore 

it uses the improvements achieved at process level, as inputs to the bank 

image model, thus, quantifying the effect of those improvements, on the quality 

of bank services and on its overall attractiveness. 

� WD-19 



Nonsmooth and Nonconvex Optimization 

Methods 



Chair: Marko M. Mäkelä, Department of Mathematics, University of 

Turku, FI-20014, Turku, Finland, makela@utu.fi 

Chair: Napsu Karmitsa, Department of Mathematics, University of 

Turku, FI-20014, Turku, Finland, napsu@karmitsa.fi 

277


1 - Truncated codifferential method for nonsmooth nonconvex 

optimization 

Adil Bagirov, School of Information Technology & 

Mathematical Sciences, University of Ballarat, University Drive, 

Mount Helen, P.O. Box 663, 3353, Ballarat, Victoria, Australia, 


We propose a new algorithm to locally minimize nonsmooth functions represented 

as a difference of two convex functions (DC functions). The algorithm is 

based on the concept of codifferential. It is assumed that DC decomposition of 

the objective function is known a priori. We develop an algorithm to compute 

descent directions using a few elements from codifferential. The convergence 

of the minimization algorithm is studied and its comparison with different versions 

of the bundle methods using results of numerical experiments is given. 

2 - A New Trust Region Algorithm Using Radial Basis 

Function Models 

Seppo Pulkkinen, Department of Mathematics, University of 

Turku, Vesilinnantie 5, Turun yliopisto, 20014, Turku, Finland, 

seppo.pulkkinen@utu.fi 

We present a new derivative-free algorithm using interpolation models and 

trust regions for black box optimization problems. The model function is constructed 

as a linear combination of radial basis functions augmented with a 

linear polynomial tail. A novel feature of the algorithm is that it solves the trust 

region subproblem by using a d.c. decomposition of the model function. The 

new algorithm is also tailored for exploiting possible structure of the problem. 

Numerical results illustrating the efficiency of the new algorithm compared to 

the present algorithms will also be presented. 

3 - Sparse regression via a trust-region proximal method 

Dongmin Kim, University of Texas at Austin, The University of 

Texas at Austin Department of Computer Sciences 1 University 

Station C0500 Taylor Hall 2.124, 78712-0233, Austin, TX, 

dmkim@cs.utexas.edu, Suvrit Sra, Inderjit Dhillon 

We present a method for sparse regression problems. Our method is based on 

the nonsmooth trust-region framework that minimizes a sum of smooth convex 

functions and a nonsmooth convex regularizer. By employing a separable 

quadratic approximation to the smooth part, the method enables the use of proximity 

operators, which in turn allow tackling the nonsmooth part efficiently. We 

illustrate our method by implementing it for three important sparse regression 

problems. In experiments with synthetic and real-world large-scale data, our 

method is seen to be competitive, robust, and scalable. 

4 - Comparing Nonsmooth Optimization Methods and Software 

Napsu Karmitsa, Department of Mathematics, University of 

Turku, FI-20014, Turku, Finland, napsu@karmitsa.fi, Adil 

Bagirov, Marko M. Mäkelä 

Nonsmooth optimization (NSO) methods may be divided in two main groups: 

subgradient and bundle methods. When developing new methods, testing is 

usually made between similar methods. We compare both bundle and subgradient 

methods as well as some methods that lie between them. The test set 

includes large amount of different NSO problems: e.g. convex and nonconvex 

problems, piecewise linear and quadratic problems and problems with different 

sizes. The aim of this paper is not to foreground some method over the others 

but to get an insight of which method to select for certain types of problems. 

� WD-21 


6.2.47 

Social Policy and Education 

Stream: Education, and Social Policy 


Chair: Hanife Akar, Department of Educational Sciences, Middle 

East Technical University, Orta Dogu Teknik Universitesi, Egitim 

Fakultesi EF 316, 06531, Ankara, Turkey, hanif@metu.edu.tr 

1 - Performance management and school achievement in 

Portuguese secondary schools 

278 

Cláudia S. Sarrico, Social, Juridical and Political Sciences, 

University of Aveiro, 3810-193, AVEIRO, Portugal, 

c.s.sarrico@ua.pt, Maria J. Rosa, Margarida F. Cardoso 

The purpose of this study is to better understand the relationship between performance 

management practices of Portuguese secondary schools and their 

achievement. Quantitative methods are used to measure school performance, 

taking a value-added approach. Schools are then positioned in a matrix of measured 

performance vs perceived performance (ranking by exam results). A set 

of schools with different levels of performance was selected for qualitative indepth 

case study analysis. The study provides a better understanding of which 

performance management practices are behind school success. 

2 - Efficacy beliefs of teachers to optimize learning opportunities: 

Incentives for teacher education policymaking 

Hanife Akar, Department of Educational Sciences, Middle East 

Technical University, Orta Dogu Teknik Universitesi, Egitim 

Fakultesi EF 316, 06531, Ankara, Turkey, hanif@metu.edu.tr, 

Feyza Erden 

Data based on a nationwide survey supported by TUBITAK show that overcrowding 

of class size has a number of dramatic disadvantages such as unequal 

participation of students, less active learning opportunities, and increased number 

of misbehaviors which may hinder optimum learning. This study helps 

understand the teacher efficacy levels in classroom management, especially 

in overcrowded public schools, which may in return provide an incentive for 

policy-makers to undertake teacher education policies, and equal school quality 

opportunities. 

� WD-22 



Maritime Vessel Routing and Deployment 



Chair: Jose Fernando Alvarez, Research & Innovation – Maritime 

Transport, Det Norske Veritas, Veritasveien 1, BRINO911, 1322, 

Høvik, Baerum, Norway, jose.fernando.alvarez@dnv.com 

1 - The linearized simultaneous string-design and cargorouting 

problem 

Christian Edinger Munk Plum, Operations Research - DTU 

Management, Network Advanced Solutions - Maersk Line, 

Esplanaden 50, 1098, Copenhagen K, Denmark, 

Christian.Edinger.Munk.Plum@maersk.com 

A global liner shipping company’s network consists of a billion-dollar investment 

in assets and similarly sized operational costs. The problem of designing 

this network, to minimize costs while considering many operational constraints 

is thus of great relevance. Empirical studies of the cost structure of a network’s 

strings (ship rotations), show a linear relation to the capacity and length of the 

string. This is used in the formulation of the problem, which will be solved 

with decomposition techniques such as column generation. 

2 - Improving the logistics of moving empty containers - 

Can new concepts avoid a collapse in container transportation? 

Stefan Voss, Wirtschaftsinformatik/Information Systems, 

University of Hamburg, Von-Melle-Park 5, 20146, Hamburg, 

Germany, stefan.voss@uni-hamburg.de, Robert Stahlbock 

Despite any crisis total seaborne trade has quadrupled over the past decades. 

However, billions of Euros are spent to deal with inefficiencies caused by repositioning 

of empty containers. Increased container population and global trade 

imbalance led to accumulation of empty containers in some major port areas 

and container shortage in others. We provide an analysis of data emphasizing 

the imbalances of container transportation. We discuss approaches for reducing 

empty container movements. Moreover, we discuss solution concepts to 

overcome parts of this situation. 

3 - Ship Scheduling 

Fotini Malliappi, School of Mathematics, UNIVERSITY OF 

SOUTHAMPTON, Highfield Road, SO17 1BJ ,

SOUTHAMPTON, United Kingdom, 

fmalliappi@googlemail.com, Chris Potts 

The presentation will focus on a classic ship scheduling problem, in which the 

routing and scheduling of a heterogeneous fleet of ships with time windows 

for pick-ups and deliveries at multiple ports is required in order to maximise 

profits. We will present an integer programming formulation of the problem. 

However, its complexity is such that heuristics methods are needed. A computational 

evaluation of a multi-start local search heuristic versus a variable 

neighbourhood search implementation will be presented as it indicates which 

method offers the most potential. 

4 - Simultaneous Deployment and Network Design in Liner 

Shipping: Formulation and solution method 

Shahin Gelareh, Production Planning - Operations Research, 

Technical University of Denmark, Lyngby 2800, Building 426, 

67663 , Copenhagen, Copenhagen, Germany, 

shahin.gelareh@gmail.com, David Pisinger 

We propose a linear mixed integer programming formulation for simultaneous 

design of network and deployment of containerships. We aim at overcoming 

the gap between the strategic planning problem of network design and tactical 

level problem of fleet deployment. By separating the string generation problem 

from the model, we propose a novel enumerative hierarchical decomposition 

approach based on an underlying primal decomposition method for solving instances 

of the problem either optimally or heuristically. Our computational 

results indicate superiority of our method to CPLEX on all instances. 

� WD-23 


6.2.49 

Data Mining in Portfolio Analysis 2 



Chair: Vadim Mottl, Intelligent Systems, Computing Center of the 


Russian Federation, vmottl@yandex.ru 

Chair: Michael Markov, Markov Processes International, 25 Maple 

Str, Suite 200, 07901, Summit, NJ, United States, 

michael.markov@markovprocesses.com 

1 - Portfolio Calibration Approach for Asset Allocation and 

Financial Optimizations 

Michael Markov, Markov Processes International, 25 Maple Str, 

Suite 200, 07901, Summit, NJ, United States, 

michael.markov@markovprocesses.com, Evgeny Bauman 

We introduce the Portfolio Calibration method in Financial Optimizations that 

results in optimal portfolios preserving their efficiency in different market scenarios. 

For each scenario the portfolio is projected onto the efficient portfolio 

set and then aggregated. Different types of projections are introduced. A robust 

efficient set Calibrated Efficient Frontier is defined. A measure of stability of 

portfolio efficiency is suggested. Using Markowitz MVO as an example, we 

show that the Calibration method produces more stable results than both the 

original MVO and other approaches. 

2 - Statistical analysis for style drift concept 

Ilya Muchnik, Comp. Science, Rutgers University, 96 

Frelinghuysen Road, 08854, Piscataway, NJ, United States, 

muchnikilya@yahoo.com, Megan Woods, Evgeny Bauman, 

Michael Markov 

Practitioners have a strong interest to get a relevant score associated with the 

concept to estimate "a quality degree of a manager style’. They take their attention 

on a variability of the manager’s style based on traditional econometric 

methods for regression analysis. Usually they use as the score a variance for a 

constant-based regression model. In this paper we follow the same approach. 

Our novelty includes two advantages: as a base linear regression model with a 

dependency portfolio from a time scale, and, we characterize a manager style 

by two criteria, passive and active. 


3 - Edge-preserving Kalman-Bucy filtration and smoothing 

of return time series: Detection of hidden events in the 

investment strategy 

Olga Krasotkina, Tula State University, 300600, Tula, Russian 

Federation, ko180177@yandex.ru, Vadim Mottl, Michael 

Markov 

The Bayesian approach to Dynamic Investment Style Analysis results in estimating 

a time-varying regression model of the publicly available time series of 

periodic returns on the portfolio being monitored with respect to the returns on 

asset classes the portfolio might be composed from. We consider an a priori 

Markov model of capital sharing which allows for sharp changes in the investment 

strategy. Having been applied to the return data of the Quantum hedge 

fund in 1992, the technique recovered the mechanism of the "Black Wednesday" 

(September 16) when George Soros broke the Bank of England. 

4 - Variable selection in model-based clustering using penalized 

mixtures of t-distributions and an application to 

financial market segmentation 

Alberto Cozzini, Statistics, Imperial College, 12 Kingswater 

Place, 1 Battersea Church Road, SW11 3BQ, London, England, 

United Kingdom, a.m.cozzini@ic.ac.uk, Giovanni Montana 

We propose a model-based clustering algorithm for segmenting financial markets 

into homogeneous groups. Several indicators describing the price dynamics 

of the markets have been extracted from historical time series of returns and 

are used as variables for clustering. The suggested model is a penalised finite 

mixture of multivariate t-distributions. By taking a penalised likelihood approach, 

we are able to discern variables that are informative for clustering from 

unimportant variables. Various penalty functions and experimental results will 

be presented. 

� WD-24 


6.2.50 

Practical Issues in Timetabling 



Chair: Geoffrey Forster, Director, Scientia Ltd., CPC1, Capital Park, 

Fulbourn, CB21 5XE, Cambridge, United Kingdom, 

geoffrey_forster@scientia.com 

1 - Determining Most Convenient and Optimal Timetable 

by Assigning Difficulty Index 

Harshita Agarwal, Computer Science and Engineering, PES 

Institute of Technology, Flat No:M-5, 1st floor, Sapthagiri 

Residency VII, #2/62, Kathriguppe Government School Road, 

off Kathriguppe main road„ 560085, Bangalore, Karnataka, 

India, h.agarwal2811@gmail.com, Nitish Shangari, Saurabh 

Shekhar 

This paper proposes a scheme to generate alternative timetables for a dynamic 

constraint satisfaction problem and to select the most optimal solution. A novel 

algorithm to compute scores for each solution’s popularity among students and 

faculty was innovated. Thus, the qualitative problem of obtaining timetables 

was successfully quantified. Real world constraints regarding faculty schedules, 

course requirements, laboratories and infrastructure were met using OR 

scheduling techniques. A generic tool ready for use in universities was developed. 

For case study, PESIT timetable was considered. 

2 - Individual Student Timetables: What Has Been Implemented 

in Syllabus Plus and Lessons Learnt. 

Geoffrey Forster, Director, Scientia Ltd., CPC1, Capital Park, 

Fulbourn, CB21 5XE, Cambridge, United Kingdom, 

geoffrey_forster@scientia.com 

We describe our experiences in incorporating individual student timetables into 

Syllabus Plus, a leading planning and timetabling system that is widely used 

throughout the World. Syllabus Plus uses constraint satisfaction techniques to 

construct optimised timetables that take into account the requirements of all resources 

including staff, students, rooms and equipment. The paper will explain 

how the needs of individual students can be considered at different stages in the 

timetabling lifecycle. The relevance of individual student timetables for those 

institutions that want to support the Bologna Process and its requirement for 

3-year Bachelor degrees will be described. 

279


3 - An integer programming approach for classroom infrastructure 

planning 

Pablo A. Rey, Industrial Engineering, Universidad Diego 

Portales, Vergara 432, Santiago, Chile, pablo.rey@udp.cl, Jaime 

Miranda 

The Faculty of Economics and Business at the Universidad de Chile is currently 

facing a rising increase in the number of enrollment causing serious infrastructure 

problems. For this reason it has embarked on a project that envisages the 

construction of a new building. This work presents an Integer Programming 

model that aims to determine the optimal number of classrooms in this building, 

additionally considering the schedules of the courses offered. It presents 

the current situation and a series of future scenarios that will guide the construction 

of this new building. 

4 - A procedure to determine work-team cross-training 

goals to cope with production mix variation and employee 

absence 

Jordi Olivella, Institute of Industrial and Control Engineering 

and Department of Management, Technical University of 

Catalonia, Avda Canal Olimpic, s/n, 08860, Castelldelfels, 

Barcelona, Spain, jorge.olivella@upc.edu, David Nembhard 

Often organizations introduce cross-training in their work teams to obtain benefits 

as flexibility and redundancy having neither a complete analysis of future 

production requirements nor operations schedule established in detail. Flexibility 

and redundancy effectiveness are difficult to measure in such a case and 

the procedures offered by the literature to define cross-training goals are not 

applicable. This situation is found in autonomous production teams and nurse 

services, between others. An intuitive system to establish the capacity to cope 

with production mix variation is proposed. A procedure is presented to determine 

the cross-training goals to cope with a certain production mix variation — 

defined with the aforementioned system— and with a certain level of employee 

absence, by minimizing the difficulty of the cross-training to be acquired. 

� WD-25 


6.2.48 




Chair: Eric Bourreau, COCONUT, LIRMM, 161 Rue Ada, 34000, 

Montpellier, France, eric.bourreau@lirmm.fr 

1 - A column generation approach for scheduling nuclear 

power plants refueling 

Roberto Wolfler-Calvo, LIPN, Université Paris Nord, 93430, 

Villetaneuse, France, roberto.wolfler@lipn.univ-paris13.fr, Lucas 

Létocart, Laurent Alfandari, Antoine Rozenkop, Daniel Chemla, 

Guillaume Turri 

A column generation approach is proposed for the problem of scheduling nuclear 

power plants production, outages and refueling. The columns represent 

feasible sequences of production and outages over the time horizon, for each of 

the nuclear power plants. This column generation scheme is started on a simplified 

and smaller version of the original problem. The columns are generated 

solving a shortest path subproblem for each nuclear power plant. Once chosen 

the optimal set of columns, an MIP model is solved including non-nuclear 

power plants and all demand scenarios. 

2 - CSP and greedy algorithm for a large-scale energy management 

problem with varied constraints 

Mirsad Buljubasic, Department of Mathematics, Faculty of 

Natural Sciences, Univ. Sarajevo, ul. Zmaja od Bosne 33, 71000, 

Sarajevo, Bosnia And Herzegovina, mirsad_bulj@yahoo.com, 

Haris Gavranovic, Faik Catibusic, Ibrahim Numanagic 

The proposed method is divided in two interdependent procedures, each dealing 

with one subset of constraints. The set of constraints is divided in two 

subsets: one is dealing exclusively with outages of plants of type 2 and the 

other subset consists of constraints imposed on the level of production and the 

fuel consumption in all power plants. Several solutions for the first set of constraints 

is found using a CSP solver. Then, the greedy procedure determines the 

feasible plan of production respecting given schedule of outages. The method 

finds feasible solutions for all given 6 instances. 

280 

3 - Hybrid constraint programming/local search for a largescale 

energy management problem with varied constraints 

Hadrien Cambazard, Cork Constraint Computation Centre, -, 

Cork, Ireland, h.cambazard@4c.ucc.ie, Emmanuel Hebrard, 

Barry O’Sullivan 

We describe our hybrid constraint programming/local search entry to the 

Roadef/Euro Challenge. Our approach involves three steps: First, we schedule 

the outages of the Type 2 plants with a constraint program. Second, we determine 

the level of refueling performed at each outage and return a constraint, or 

cut, to the master problem, if unfeasibility is detected. Third, we compute the 

daily production for each plant. The process is iterated, and at each iteration, 

the search is stochastically guided by the current best solution and through a 

relaxation of the objective function. 

4 - A functional programming approach for an energy planning 

problem 

Valentin Weber, G-SCOP, 46 avenue Félix Viallet, 38000, 

GRENOBLE, France, valentin.weber@g-scop.grenoble-inp.fr, 

Julien Darlay, Yann Kieffer, Louis Esperet, Guyslain Naves 

The ROADEF/EURO challenge 2010 is to give an actual program that solves a 

real-world energy management problem given by EDF, including discrete constraints, 

and uncertainty in the form of different, fixed, known scenarios. We 

tried what we think is an original angle of attack from the software engineering 

point of view: we use the functional programming language CAML to implement 

all of the input/output routines and ad-hoc combinatorial algorithms to 

solve the problem. We will report on the strengths and weaknesses of such an 

approach for a medium- or large-scale optimization problem. 

� WD-26 


3.1.11 

New Achievements in Game Theory II 

(cooperative and noncooperative) 






Chair: Sirma Zeynep Alparslan Gok, Mathematics, Faculty of Arts 

and Sciences, Suleyman Demirel University, Faculty of Arts and 






1 - Parametric Direct Mechanisms for Joint Replenishment 

Evren Korpeoglu, Department of Industrial Engineering, Bilkent 

University, Bilkent University Engineering Faculty, Industrial 

Engineering Dept., 06800, Ankara, Turkey, 

evrenko@bilkent.edu.tr, Alper Sen, Kemal Guler 

We consider the joint replenishment problem in a non-cooperative setting 

where players’ characteristics are not verifiable. Joint replenishment frequency 

and allocation of joint costs are governed by parametric direct mechanisms 

based on players’ reported characteristics. We obtain necessary and sufficient 

conditions for the existence and uniqueness of a Nash equilibrium in which all 

players participate in joint replenishment. We characterize the equilibrium behavior 

and analyze the equilibrium outcomes with respect to efficiency and the 

distribution of joint costs. 

2 - On Taxes for Atomic Network Games with Heterogeneous 

Users 

George Karakostas, Computing & Software, McMaster 

University, 1280 Main St. W., L8S4K1, Hamilton, Ontario,

Canada, karakos@mcmaster.ca, Dimitris Fotakis, Stavros 

Kolliopoulos 

In atomic network games with a finite number of non-cooperative users selecting 

paths, the users experience congestion-dependent latency on the network 

edges. Our aim is to mitigate the inefficiency caused by the selfish users by 

introducing taxes on the edges. These taxes are strongly (weakly)-optimal if 

all of (at least one of) the induced equilibria minimize(s) the total latency. We 

show the existence of weakly-optimal taxes for single-source network games 

with heterogeneous users, each with a different tax sensitivity (heterogeneous), 

and discuss the existence of strongly-optimal taxes. 

3 - Fair and Efficient Equilibrium in Collective Actions 

Dongbin Huang, Department of Environmental Sciences, ETH 

Zurich, Institute of Terrestrial Systems, Universitaestrasse 16, 

8092, Zurich, Switzerland, dongbin.huang@env.ethz.ch, Hans 

Rudorf Heinimann 

In collective actions, two or more persons strive for a solution that best satisfies 

multi-dimensional preferences of each. A good solution for all can be an equilibrium 

that is both fair and efficient. An adequately specified mechanism to 

achieve such an equilibrium will help to solve collective action problems. Here 

we define a fair and efficient solution as a contracting point on the trajectory of 

Nash solution where the potentials of increasing positions are maximally exploited, 

the common efficiency loss is within tolerance, and the exaggeration 

of positions is controlled. 

4 - Downsian competition with an arbitrary number of parties 

Tom Blockmans, MOSI, Vrije Universiteit Brussel, Pleinlaan 2, 

1050, Brussel, Belgium, tblockma@vub.ac.be, Mark Van 

Lokeren 

We consider a unidimensional model of spatial electoral competition with an 

arbitrary number of political parties. Voters are continuously distributed along 

[0,1] such that the density is strictly positive on (0,1). Downsian competition 

between the parties is modelled as a non-cooperative game with [0,1] as the 

common strategy set. Extending the result of Sofronidis (Math. Social Sci. 

50 (2005), no. 3, 331-335) with four parties, we determine the existence and 

value of the Nash equilibrium strategies as well as the necessary and sufficient 

conditions regarding voter distribution. 

� WD-27 


8.2.06 

TRAFFIC MANAGEMENT 



Chair: Riccardo Rossi, Department of Structural and Transportation 

Engineering, University of Padova, Via Marzolo, 9, I35131, Padova, 

-, Italy, riccardo.rossi@unipd.it 

1 - Analysis of gap-acceptance behavior at road junctions: 

experiments with a driving simulator 

Massimiliano Gastaldi, Department of Structural and 

Transportation Engineering, University of Padova, Via Marzolo, 

9, I35131, Padova, Padova, Italy, 

massimiliano.gastaldi@unipd.it, Riccardo Rossi, Claudio 

Meneguzzer 

In this paper the preliminary results of experiments on gap-acceptance behavior 

using a fixed-base driving simulator are described. The analysis focuses on 

the right turn maneuver from a minor street at a priority road junction between 

two-lane two-way rural roads. The primary objectives of the research are: to 

test the driving simulator ability to represent a real gap-acceptance situation, by 

comparing the results with those obtained from the analysis of the real junction, 

and to analyze the effect on gap-acceptance behavior of conditions that are not 

easily recorded in the real situation. 

2 - Roundabouts performance estimation: an experimental 

comparative analysis for different methodologies 

Roberto Camus, Civil and Environmental Engineering, 

University of Trieste, P.le Europa, 1, 34127, Trieste, Italy, 


camus@dica.units.it, Massimiliano Gastaldi, Giovanni Longo, 

Riccardo Rossi 

This paper presents the results carried out from a comparative analysis of 

roundabout capacity estimation models. Using data collected during experimental 

observations performed on an existing roundabout, different types of 

models (macro, micro and mesoscopic) have been compared with reference to 

their capability to estimate real capacity. A detailed analysis has interested in 

particular an mesoscopic approach which refers to the facility as a succession 

of merging areas and allows estimating capacity, traffic flows and delays. 

3 - The effect of information quality on travelers’ behavior 

in ATIS context of choice 

Roberta Di Pace, Department of Transpo.Engineering 

"L.Tocchetti", University of Naples Federico II, via Claudio, 21, 

80125, napoli, italy, Italy, roberta.dipace@unina.it, Gennaro 

Nicola Bifulco 

In order to model travelers’ compliance under ATIS (Advance Traveler Information 

Systems), the effect of information accuracy has been studied. The 

research has been carried out by considering data from two experiments (based 

on a Stated Preference approach), carried out by means of two different travel 

simulators. In the experiments respondents are asked to made repeatedly their 

route choices All observed choices have been analyzed by applying appropriate 

statistical test and the effect of information accuracy on compliance has been 

tested under different reliability scenarios. 

4 - An empirical analysis of vehicle time headways on rural 

two-lane two-way roads 

Riccardo Rossi, Department of Structural and Transportation 

Engineering, University of Padova, Via Marzolo, 9, I35131, 

Padova, -, Italy, riccardo.rossi@unipd.it, Massimiliano Gastaldi, 

Riccardo Maratini 

This paper presents the results carried out from an experimental analysis focused 

on vehicle time headways distributions with reference to two-lane twoway 

roads. Our attention has been focused on roads localized in Northern Italy 

and characterized by different levels of traffic; using data collected by inductive 

loops and radar sensors, a wide set of observations has been made available. 

The statistical analysis of these data has allowed to test a set of headway distribution 

models highlighting their goodness-of-fit with reference to the empirical 

distributions. 

� WD-28 



Scheduling in Production and 

Communication 



Chair: Celia Glass, Cass Business School, City University, 106 

Bunhill Row, EC1Y 8TZ, London, United Kingdom, 

c.a.glass@city.ac.uk 

1 - Experimental evaluation of the work function algorithm 

for solving the on-line k-server problem 

Robert Manger, Department of Mathematics, University of 

Zagreb, Bijenicka cesta 30, 10000, Zagreb, Croatia, 

manger@math.hr, Tomislav Rudec, Alfonzo Baumgartner 

We are concerned with practical applicability of the work function algorithm 

(WFA) for solving the on-line k-server problem. Our aim is to measure experimentally 

the actual performance of the WFA in terms of its serving costs. 

First, we describe a relatively efficient implementation of the WFA. Next we 

present some experiments based on our implementation, where the WFA has 

been tested on considerably large problem instances. By using the obtained 

experimental results, we finally compare the performance of the WFA with the 

corresponding theoretical estimates and with some other algorithms. 

2 - Scheduling with identical machines and servers 

Wafaa Labbi, Faculty of Mathematics, USTHB University, BP 

32, El-Alia, Bab-ezzouar, 16111, Algiers, fawalab@yahoo.fr, 

Mourad Boudhar 

281


In this talk, we address the problem of scheduling tasks on identical machines 

in the presence of servers. Before being processed on a machine, each task must 

undergo a special processing by servers. Machines as well as the servers can 

process only one task at the one and same time. We show that the problem is 

NP-hard and we give new polynomial subproblems. A new exact method based 

on mathematical modeling and heuristics are also presented with experimentation 

results. Based on randomly generated instances, these experimentations 

enable us to assess the efficiency of the suggested methods. 

3 - Optimization of Schedules for a Multi-Task Production 

Cell 

Karin Thornblad, Mathematical Sciences, Chalmers University 

of Technology, 412 96, Göteborg, Sweden, karint@chalmers.se, 

Ann-Brith Strömberg, Michael Patriksson, Torgny Almgren 

To produce optimal schedules for a real production cell in the aircraft engine 

industry working as a job-shop, three different optimization models are developed 

and tested on real production data at various levels of work load. The current 

production planning prerequisites are studied in order to construct relevant 

and functional objective functions. The evaluation of the results is performed 

using real instance data in a simulation model and compared with production 

schedules formed by the existing built-in scheduling algorithm together with a 

number of priority functions. 

4 - Periodic Scheduling for Wireless Mesh Networks 

Celia Glass, Cass Business School, City University, 106 Bunhill 

Row, EC1Y 8TZ, London, United Kingdom, 

c.a.glass@city.ac.uk, Eun-Seok Kim 

Wireless mesh networks (WMNs) provide flexible and low-cost Internet and 

Broadband access. Local access points (e.g., for WiFi access) connect to each 

other to facilitate a wide-area wireless mesh network. The star nature of access 

nodes disposes them to perfectly-periodic scheduling in which time is divided 

into time-slots, and each client gets a time slot at a predefined frequency. 

However, the problem of finding a feasible perfect-periodic schedule is NPhard. 

We develop feasibility conditions for perfect periodic scheduling, using 

number-theoretic techniques, and devise a polynomial time algorithm for 3 coprime 

frequencies. The results form a foundation for realistic scheduling of 

WMNs. 

� WD-29 


8.2.11 

Information and Ambiguity in Financial 

Modeling 



Chair: Georg Pflug, Department of Statistics and Decision Support 

Systems, University of Vienna, Universitaetsstrasse 5, A-1010, 

Vienna, Austria, georg.pflug@univie.ac.at 

1 - Multi-period acceptability functionals: The role of information 

monotonicity 

Raimund Kovacevic, Statistics and Decision Support Systems, 

University Vienna, Universitätsstr. 5, 1010, Wien, Wien, Austria, 

raimund.kovacevic@univie.ac.at 

Multi-period acceptability functionals valuate discrete-time stochastic processes 

(e.g. representing income streams). Typically such functionals can 

be used as objective functions in (multistage) stochastic optimization. Multiperiod 

functionals are defined in a generic way, such that the information available 

(expressed as filtration) enters explicitly the definition of the functional. 

Information monotonicity then demands that the value of a process increases, if 

information arises faster over time, which is expressed by a finer filtration. We 

analyze several functionals from literature with respect to information monotonicity 

and give a characterization of this property for certain class of functionals, 

related to the composition of conditional acceptability functionals. 

2 - Ambiguity and minimaxity in portfolio selection and 

282 

Georg Pflug, Department of Statistics and Decision Support 

Systems, University of Vienna, Universitaetsstrasse 5, A-1010, 

Vienna, Austria, georg.pflug@univie.ac.at, David Wozabal, Alois 

Pichler 

The typical procedure in portfolio optimization is to fix a probability model 

and to find an optimal portfolio for exactly this probability model. However, 

there is often much uncertainty about the correct model. We call the model 

uncertainty "ambiguity". Under ambiguity, the portfolio problem has the form 

of a maximin problem. We discuss saddlepoint solutions and demonstrate, that 

large ambiguity finally leads to optimal portfolios, which are equal-weight portfolios. 

3 - Minimum VaR and Minimum CVaR Optimal Portfolios: 

Estimation and Inference Procedure 

Taras Bodnar, Department of Statistics, European University 

Viadrina, Grosse Scharrnstrasse 59, 15230, Frankfurt (Oder), 

Germany, bodnar@euv-frankfurt-o.de, Wolfgang Schmid, Taras 

Zabolotskyy 

We derive the finite-sample distributions of the estimated characteristics of the 

minimum VaR and of the minimum CVaR portfolios. Very useful stochastic 

representations of these estimators are obtained and used for deriving their 

conditional and unconditional moments. The joint distribution of the estimated 

expected return and the estimated variance. This result is used for deriving 

exact tests for the corresponding portfolio characteristics and for constructing 

joint confidence regions for the minimum VaR and the minimum CVaR portfolios 

in the mean-variance space. 

4 - p-Order Conic Programming in Stochastic Optimization 

Pavlo Krokhmal, Mechanical and Industrial Engineering, 

University of Iowa, 3131 Seamans Center, 52242, Iowa city, IA, 

United States, krokhmal@engineering.uiowa.edu 

We discuss modeling of risk preferences in stochastic programs via p-order 

conic constraints. As an illustration, a portfolio optimization problem is considered. 

Several solution approaches are presented, including reduction to secondorder 

conic programming and polyhedral approximations of p-order cones, as 

well as a branch-and-bound scheme for mixed-integer problems with p-conic 

constraints. Computational results of portfolio optimization case studies are 

discussed. 

� WD-30 


8.2.13 

OR methods in portfolio management and 

asset allocation 


in Banking 


Chair: David Pla-Santamaria, Alcoy School, Technical University of 

Valencia, 03801, Alcoy, Spain, dplasan@esp.upv.es 

1 - Downside risk: How to select efficient portfolios from 

Dow Jones blue chips by the mean-semivariance efficient 

frontier model (E-SV) 

Mila Bravo, ALCOY SCHOOL, TECHNICAL UNIVERSITY 

OF VALENCIA, 03801, ALCOY, Spain, mibrasel@epsa.upv.es, 

David Pla-Santamaria 

This paper is characterized as follows. Scope and purpose: to apply E-SV 

(Ballestero, 2005) to an actual example of portfolio choice in which downside 

risk is measured by the semivariance. Opportunity set and observation period: 

Dow Jones stocks with daily prices observed over the period 2005-09. Returns: 

they are capital gains weekly computed. Validity of results is affected by 

these limitations. Diversification constraints: portfolio weight cannot exceed 

5%. Results:significant differences between E-SV and E-V portfolios of equal 

expected returns are found. Comparisons between them are made. 

2 - Portfolio selection with multiple time horizons: a 

stochastic goal programming approach 

Antonio Benito, Economics and Social Sciences, Technical 

University of Valencia, Plaza Ferrándiz y Carbonell s/n, EPSA, 

03801, Alcoy, Alicante, Spain, anbebe@esp.upv.es, Ana 

Garcia-bernabeu, Ignacio Gonzalez

Traditional approaches to portfolio selection require establishing a time horizon 

of historical returns over a period that the investor defines in a conventional 

way. To avoid arbitrary choice of this horizon we propose to use stochastic 

goal programming (SGP) with multiple criteria Cj (j = 1, 2, . . . .., n) such as 

C1 (observed returns over the last two years), C2 (last three years) and C3 

(last four years). As information provided by these horizons is of different 

quality/reliability, the criteria are weighted from the investor’s opinions on this 

reliability. 

3 - Asset allocation for standard and non-standard ethical 

investors 

Verónica Cañal, Applied Economics, University of Oviedo, 

Avda. del Cristo s/n, 33006, Oviedo, Spain, Spain, 

vcanal@uniovi.es, Celia Bilbao, Mar Arenas-Parra, Maria 

Victoria Rodriguez Uria 

In an asset allocation problem the ethical investor seeks the combination of securities 

that best suit his ethical profile. In this work we present models to asset 

allocation when the investor’s objectives include ethical features. We detect the 

standard ethical investors applying revealed preference techniques for estimating 

the implicit prices of each ethical attribute. A non-standard ethical investor 

disagrees with the assessment obtained by the revealed preference techniques. 

We have modeled both types of ethical investors. 

� WD-33 


8.2.19 

Decision making under model uncertainty 



Chair: Pauline Barrieu, Statistics, London School of Economics, 

Houghton street, WC2A 2AE, London, United Kingdom, 

p.m.barrieu@lse.ac.uk 

Chair: Max Fehr, Statistics department, LSE, WC2A2AE, London, 


1 - Assessing Cleanup Costs 

Bernard Sinclair-Desgagné, International Business, HEC 

Montréal, 3000 Chemin de la Côte Sainte-Catherine, H3T2A7, 

Montréal, Québec, bsd@hec.ca, Pauline Barrieu 

The remediation of contaminated sites is often subject to substantial cost overruns. 

This systematic discrepancy between estimated and realized costs is 

chiefly responsible for misguided land use and wasteful delays in the reconversion 

of former industrial sites. This paper derives relatively tractable and 

simple formulas for better assessing cleanup costs. These formulas are based 

on generic remediation methods, deal explicitly with incomplete information 

and uncertainty, and are robust to misspecication in key parameters such as the 

effectiveness of a prescribed treatment. 

2 - A Monte Carlo method for problems of optimal stochastic 

control with convex value functions 

Juri Hinz, Logistics, Zurich Universtity of Applied Sciences, 

IDP, Rosenstrasse 3, CH-8401 , Winterthur, Switzerland, 

hizr@zhaw.ch 

We present a method for the calculation of the optimal policy for infinite horizon 

optimal control problems whose value function is convex. Control problems 

of this type appear in many applications and encompass important examples 

arising in the area of partially observed Markov decision processes. We 

show that the calculation performance can be improved by a modification of 

the classical least-square approach. Our adaptation is based on the convexity 

property of conditional expectation, valid in our framework. 

� WD-35 


6.2.46 

MINLP - Problem-specific Approaches 



Chair: Sarah Drewes, Department of Mathematics, Technische 

Universität Darmstadt, Dolivostr. 15, 64293, Darmstadt, Germany, 

drewes@mathematik.tu-darmstadt.de 


1 - An MIP Approach to Gas Network Optimization 

Bjoern Geissler, Discrete Optimization, Technische Universität 

Darmstadt, Dolivostr. 15, 64293, Darmstadt, Germany, 

geissler@mathematik.tu-darmstadt.de, Alexander Martin, 

Antonio Morsi, Lars Schewe 

Many optimization problems over networks can be modeled most naturally as 

mixed-integer nonlinear (nonconvex) programs. Real-life problems often result 

in large-scale MINLPs, which are currently far away from being solvable by 

state-of-the-art solvers. Therefore, we present techniques to construct arbitrary 

tight mixed-integer linear programming relaxations of an underlying MINLP 

which satisfy an a priori error bound. We demonstrate the suitability of our 

approach by means of gas network optimization problems. 

2 - Decomposition Techniques for MINLPs on Loosely 

Coupled Networks 

Antonio Morsi, Mathematics, TU Darmstadt, Dolivostr. 15, 

64293, Darmstadt, Germany, 

morsi@mathematik.tu-darmstadt.de, Bjoern Geissler, Alexander 

Martin, Lars Schewe 

Decomposition techniques are an effective way of partitioning a large problem 

into smaller ones. We reduce the complexity of nonlinear loosely coupled 

network flow problems by decomposing the underlying network topologically. 

A relaxation of the constraints coupling adjacent blocks yields a partitioning. 

Based on this decomposition we present primal and dual solution methods in 

the context of our motivating applications from gas and water transport optimization.These 

methods are applied to produce a solution to the overall problem 

by calculating lower and upper bounds from the subproblems. 

3 - Mixed 0/1-Kelly Criterion: Geometric mean maximization 

revisited 

Sebastian Pokutta, Department of Mathematics, Technische 

Universität Darmstadt, Alois-Eckert-Str. 4, 60528, Frankfurt, 

Germany, sebastian.pokutta@mac.com, Sarah Drewes 

Second order cone programs occur in a natural way in financial risk management 

and portfolio optimization. While in the case of continuous variables 

these problems can be solved efficiently, in the presence of integral variables, 

representing e.g., indivisible assets, the situation becomes signficantly more 

complicated. We present an approach for maximizing the geometric mean 

(which is equivalent to the Kelly Criterion) of affine functions with binary variables. 

Exploiting the structure of this class of programs (weakly-coupled second 

order cone programs) these programs can be solved fast. 

4 - On Bound Computations for MINLP 

David M. Gay, AMPL Optimization LLC, 87108-3379, 

Albuquerque, NM, United States, dmg@ampl.com 

Various techniques, such as branching, bounding, cut-generation, and generation 

of outer approximations can all be useful in algorithms for solving 

mixed-integer nonlinear programming problems. When the nonlinear parts of 

a MINLP problem can be specified algebraically, slope computations (to which 

various European authors have contributed) can be useful in computing bounds, 

outer approximations, and domain reductions. This talk provides a review of 

these computations. 

� WD-36 


3.1.05 

Linear and Conic Programming I 

Stream: Linear and Conic Programming 





Chair: Sergei Chubanov, University of Siegen, Hoelderlinstr. 3, 

Siegen, Deutschland, 57076, Siegen, Germany, 


283


1 - An approach based on conic formulation and MC technique 

for the time-cost trade-off problem 

Mohammad Reza Peyghami, Mathematics, K.N. Toosi 

University of Technology, Math. Department, K.N. Toosi 

University of Technology„ P.O.Box 16315-1618, Tehran, Iran, 

16315-1618, Tehran, Tehran, Iran, Islamic Republic Of, 

peyghami@kntu.ac.ir 

A new approach based on conic formulation and Monte Carlo (MC) simulation 

technique is provided for the Time-Cost Trade-off Problem (TCTP) in PERT 

networks in which all activities are subjected to linear cost function and assumed 

to be exponentially distributed. The aim is to to maximize the project 

completion probability in a predefined due date. To do so, TCTP is constructed 

as a nonlinear optimization problem in a single path with decision variables of 

activity mean durations. We then develop an algorithm based on MC simulation 

technique and conic formulation to solve general TCTP. 

2 - Finding the Extrema of Continuous Piecewise Linear 

Functions 

Özge Arslan, Industrial Engineering, Koc University, Koc 

Universitesi Rumelifeneri Yolu, 34450 Sariyer, Istanbul, Turkey, 

odemiryapan@ku.edu.tr, Metin Turkay 

Development of accurate models and efficient solution algorithms for piecewise 

linear functions attracted a lot of attention because of its wide range of application 

areas. In this paper a novel linear programming formulation is presented to 

find the extrema of continuous piecewise linear functions in all shapes (convex, 

concave, non-convex). The simplex method moves among the extreme points 

of the feasible region while searching the optimal solution and our formulation 

constructs a feasible region which utilizes this property of the simplex method. 

3 - Divide and conquer: A polynomial algorithm for linear 

programming 

Sergei Chubanov, University of Siegen, Hoelderlinstr. 3, Siegen, 

Deutschland, 57076, Siegen, Germany, 


We present an algorithm for solving systems of linear inequalities. The algorithm 

uses a divide-and-conquer algorithm as an oracle. The algorithm is 

polynomial. Moreover, the algorithm can either find a solution of the system or 

decide that there is no 0,1-solution in strongly polynomial time. 

� WD-37 


3.1.09 

Application of outranking approach for 

sustainable development 



Chair: Pascal Oberti, University of Corsica, UFR Droit Economie„ 

Av. Jean-NICOLI, BP 52, 20250 CORTE, 20250, Corte, France, 

pascal.oberti@univ-corse.fr 

Chair: Jean-François NoËl, Université d’Angers, 78047, Angers, 

France, Jean-Francois.Noel@uvsq.fr 

1 - Sustainable development, economics and multicriteria 

evaluation: which links with the outranking approach? 

Jean-François Noël, CEMOTEV, University of Versailles - Saint 

Quentin en Yvelines, UVSQ - UFR SSH, 47 Boulevard Vauban, 

78047, Guyancourt cedex, France, Jean-Francois.Noel@uvsq.fr, 

Pascal Oberti 

In the ecological economics framework, evaluation of projects and policies for 

sustainable development is generally performed using multicriteria methods. 

However, the outranking approach is still little used. The purpose of this communication 

is to explore possible links. More particularly, will be considered: 

the hypothesis of weak comparability of the values, the substitutability or complementarity 

of natural capital to other forms of capital, the identification of 

sectoral situations and finding the best compromise, the rationality of actors, 

the distribution of winners and losers. 

284 

2 - Application of ELECTRE IS method to photovoltaic 

plant projects selection in Corsica 

Pierrick Haurant, Univeristé de Corse, 20250, Corte, France, 

haurant@univ-corse.fr, Pascal Oberti, Marc Muselli 

High solar potential and advantageous purchase tariffs of photovoltaic (PV) 

electricity make Corsica’s island very attractive for industrials of this sector. 

Confronted with an unprecedented offer of PV plant projects, regional public 

actors having to deliver an opinion resorted to multicriteria evaluations for aiding 

a file based selection. This communication presents the main results of a 

study which was ordered to us. Will be underlined project constraints, evaluation 

criteria and other associated parameters, as well as final recommendation 

resulting from outranking ELECTRE IS method. 

3 - Management effectiveness assessment in marine protected 

areas: practices and feasibility of an outranking 

approach 

Ange-Michel Poli, Université de Corse, 20250, Corte, France, 

poli.ange-michel@wanadoo.fr, Jean-François Noël, Pascal 

Oberti 

The multiple uses and objectives of marine protected areas (MPAs) involve a 

planned, adaptive and effective management. What practices, methodologies 

and operational tools to assess the effectiveness of this management? This paper 

analyzes in the literature the protocols for monitoring and evaluation, and 

study the feasibility of these methodologies within French, Mediterranean and 

African MPAs. We particularly highlight what can bring the outranking approach, 

restricted to a pseudo-criterion or ELECTRE-type in a multicriteria 

context. 

� WD-38 


6.2.44 

Recent Advances in Spare Parts Inventory 

Management 



Chair: Joern Meissner, Management Science, Lancaster University 


Kingdom, j.meissner@lancaster.ac.uk 

1 - On cross-correlation of demand for spare parts 

Nezih Altay, Management, DePaul University, 1 E. Jackson 

Blvd., Suite 7000, 60604, Chicago, Illinois, United States, 

naltay@depaul.edu 

Although the literature on the effects of correlation on stock control is plentiful, 

research mainly focuses on the correlation of demand of a single item (autocorrelation) 

or demand of multiple items (cross-correlation). A considerable 

portion of spare parts, on the other hand, display time-based cross-correlation, 

i.e. demand size is correlated with the time interval between demand occurrences. 

We explore the effects of time-based cross-correlation on stock control 

of spare parts. 

2 - Traditional and Non-Traditional Methods of Forecasting 

Lumpy Demand: Inventory Control Implications 

Adriano Solis, Management Science Area, School of 

Administrative Studies, York University, 4700 Keele St, M3J 

1P3, Toronto, Ontario, Canada, asolis@yorku.ca, Somnath 

Mukhopadhyay, Rafael S. Gutierrez 

Simple exponential smoothing, weighted moving average, Croston’s method, 

Syntetos-Boylan approximation (SBA), and neural network (NN) modeling are 

applied on a time series dataset of very lumpy demand. A simple NN model 

is found superior overall with respect to several scale-free forecast accuracy 

measures. However, demand forecasting performance may not translate into 

inventory systems efficiency. A periodic review inventory control system with 

forecast-based order-up-to levels is simulated. Holding and shortage costs and 

service levels are considered. Findings/insights will be presented. 

3 - Service Parts Inventory Control with Lateral Transshipment 

that Takes Time 

Guangyuan Yang, Erasmus University Rotterdam, Econometric 

Institute, Burg. Oudlaan 50, H09-21, 3062 PA, Rotterdam, 

Rotterdam, Netherlands, gyang@ese.eur.nl, Rommert Dekker

In equipment-intensive industries, the transshipment time for some slow moving 

service parts is not negligible. We assess the effect of non-negligible lateral 

transshipment time on various aspects of spare parts inventory control. Furthermore, 

we introduce customer-oriented service levels by taking the uncommitted 

pipeline stocks into account. A case study in the dredging industry shows that 

lateral transshipment may lead to lower system performance. Furthermore, we 

find that considerable savings can be obtained when we include the uncommitted 

pipeline stocks in inventory control decisions. 

4 - Demand Categorization for Safety Stock Planning of 

Spare Parts 

David Bucher, Management Science, Lancaster University, 

Lancaster University Management School, Department of 

Management Science, LA1 4YX, Lancaster, Lancashire, United 

Kingdom, d.bucher@lancaster.ac.uk, Joern Meissner 

We develop a demand categorization system for an automated selection of statistical 

(compound) distributions for safety stock planning for items with intermittent 

demand. We test our new technique on the data of the spare parts 

inventory of a major German car component manufacturer with 20,000 SKUs. 

The simulation shows significant cost reduction and an increased service level. 

Our results promote the application of categorization tools directly linked to 

the configuration of inventory systems. 

� WD-39 


6.2.45 

Scheduling and lotsizing under 

uncertainties I 



Chair: Alexandre Dolgui, IE & Computer Science, Ecole des Mines 

de Saint Etienne, 158, cours Fauriel, 42023, Saint Etienne, France, 

dolgui@emse.fr 

Chair: Mikhail Kovalyov, Faculty of Economics, Belarusian State 

University, Nezavisimosti 4, 220030, Minsk, Belarus, 

kovalyov_my@yahoo.co.uk 

1 - A lot-sizing and scheduling problem under uncertainties 

Alexandre Dolgui, IE & Computer Science, Ecole des Mines de 

Saint Etienne, 158, cours Fauriel, 42023, Saint Etienne, France, 

dolgui@emse.fr, Kseniya Schemeleva, Frédéric Grimaud 

We consider the lot-sizing and scheduling problem under uncertainties. It is 

assumed that all machines at the production line are imperfect, notably we 

concede that a deal of processed parts can be defective and each machine can 

breakdown. Thus we deal with random yields and random lead time uncertainties. 

The planning horizon and demand for parts of all products are given. 

We assume that a set-up time is necessary to change the type of manufacturing 

product. The goal is to optimize the order of lots and number of items of each 

product for a given planning horizon. 

2 - Batching Work and Rework Processes to Minimize the 

Makespan 

Frank Werner, Faculty of Mathematics, Otto-von-Guericke 

University, FMA,I, nstitute of Mathematical Optimization, PSF 

4120, 39016, Magdeburg, Germany, 

frank.werner@mathematik.uni-magdeburg.de, Irina 

Gribkovskaia, Sergey Kovalev 

A planning problem of an imperfect production is considered with a main facility 

and a facility dedicated to remanufacturing defective units. Units after 

processing in main facility are inspected for quality in batches. The quality 

inspection requires some time. Defective units are transported to the remanufacturing 

facility. The transportation also requires some time. The problem 

is to find a sequence of batch sizes such that the makespan is minimized. A 

linear programming formulation is suggested and an O(log K) time solution 

algorithm is developed. 

3 - Heuristics for an integrated production and maintenance 

planning problem 

Marouane Alaoui Selsouli, Ecole des Mines de Nantes, IRCCyN, 

La Chantrerie 4, rue Alfred Kastler, B.P. 20722, - F-44307, 


NANTES Cedex 3, France, malaou07@emn.fr, Najib. M. Najid, 

Abdelmoula Mohafid 

We consider a problem of integrating production and maintenance. Maintenance 

problem is the problem of determining the dates of preventive maintenance 

actions in time windows. Production problem treats the production 

planning problem known as multi items capacitated lot sizing problem. The 

resulting problem is modelled as a linear mixed-integer program. It takes into 

account demand shortage and the reliability of the production line. We will 

solve the problem with a Lagrangean heuristic and heuristic based on columns 

generation. The aim of those heuristics is to provide a lower bound and to 

construct a feasible solution for the integrated problem. Computation results 

show a comparison between the Lagrangean heuristic and heuristic based on 

columns generation. 

4 - Push and pull heuristics for remanufacturing with yield 

uncertainty 

Erwin van der Laan, RSM Erasmus University, P.O.Box 1738, 

3000DR, Rotterdam, Netherlands, elaan@rsm.nl 

In this paper we focus on the uncertainty in the quality of a recoverable product 

and its implications for managing a hybrid manufacturing/remanufacturing inventory 

system. We propose simple heuristics for controlling the inventories of 

both recoverable products and serviceable products (which is a mix of remanufactured 

and manufactured products) that are easily implemented in practice. 

An extensive numerical study quantifies the performance of the heuristics under 

various scenarios. 

� WD-40 


6.2.52 

Recent Advances in Industrial and 

Engineering Optimization I 






1 - Stream Mapping Value (MVS) Applied to Improve Productivity 

and Minimize Waste in a Tool for Stamping 

Maker 

Milton Vieira Junior, Industrial Engineering Post Graduation 

Program, Nove de Julho University - UNINOVE, Av. Francisco 


mvieirajr@uninove.br, Nivaldo Coppini, Marcelo Bonandi 

Many companies agree on improvement of manufacturing process using lean 

techniques. Value Stream Map can show where extra materials are piling up 

and it is often used to introduce lean manufacturing procedure. The aim of 

this work was to create a current state diagram that showed how the process 

currently works in tools for stamping maker. The second step was to identify 

the problems areas and create future state diagram that allows improving the 

process, reducing the waste and increase efficiency and quality 

2 - Cellular Manufacturing Application in Bosch Diesel 

Systems Injector Plant 

Zerin Turgut, Industrial Engineering, Bosch Diesel 

Systems/Bursa Unit, Bosch Sanayi ve Ticaret A.¸S., Organize 

Snayi Bölgesi Ye¸sil cad. no:27, 16159, Bursa, Turkey, 

zerin.turgut@tr.bosch.com, Ihsan Ozer, Burak Erismis, Halil 

Özbey, Cem Tangil 

Cellular manufacturing was implemented in body production with 377 employees, 

150 machines and 21 consequent sub-processes. Formerly, production had 

a complex structure with no scalability, lower productivity and higher WIP levels. 

Heuristic methods and lean principles were used to create scalable cells, 

discrete event simulation in EM-Plant software developed and a non-dominated 

solution found. Real life results are promising by higher productivity, less WIP 

and lead time. The most important consequence is a flexible, easy-to-manage, 

transparent, traceable and improvable system. 

285


3 - An efficient model for multiobjective cell formation in 

group technology 

Dmitry Krushinsky, Department of Operations, University of 

Groningen, Nettelbosje 2, 9747 AE , Groningen, Netherlands, 

d.krushinsky@rug.nl, Boris Goldengorin 

Cell formation is a key concept in group technology that provides a useful 

way to increasing the productivity and improving flexibility of a manufacturing 

system. Given approaches use sequential procedures and construct cells by 

heuristics, or formulate the problem as a mixed-integer linear program known 

to be computationally hard and use heuristics. We show that p-Median problem 

based cell formation model can be solved to optimality by general purpose 

solvers CPLEX, Xpress etc. on a standard PC within seconds. 

4 - Reliability and Sensitivity Analysis of a Repairable System 

with Imperfect Coverage and Service Pressure 

Kuo-Hsiung Wang, Department of Applied Mathematics, 

National Chung-Hsing University, Department of Applied 

Mathematics, National Chung-Hsing University, 402, Taichung, 

Taiwan, khwang@amath.nchu.edu.tw 

We study an M/M/R machine repair problem with variable servers in which 

failed machines balk. We derive analytic steady-state solutions through which 

several system performance measures can be obtained. A cost model is developed 

to determine the optimal values of the number of busy servers and 

balking rate. We use the direct search method and the Newton’s method to find 

the global minimum value until the balking rate constraint is satisfied. Numerical 

results are provided in which various system performance measures are 

evaluated under optimal operating conditions. 

� WD-41 


3.1.06 

Monte Carlo and Malliavin Calculus 

Stream: Simulation Methods in Finance 


Chair: Henry Schellhorn, mathematics, Claremont Graduate 

University, 710 N. College Ave, 91711, Claremont, CA, United 

States, Henry.Schellhorn@cgu.edu 

1 - An Algorithm for the Pricing of Path-Dependent American 

Options Using Malliavin Calculus 

Henry Schellhorn, mathematics, Claremont Graduate University, 

710 N. College Ave, 91711, Claremont, CA, United States, 

Henry.Schellhorn@cgu.edu 

We propose a recursive scheme to calculate backward the values of conditional 

expectations of functions of path values of Brownian motion. This scheme is 

based on the Clark-Ocone formula in discrete time. We construct an algorithm 

based on our scheme to efficiently calculate the price of American options on 

securities with path-dependent payoffs. Our algorithm can be combined with 

regression-based Monte Carlo methods, like the Tsitsiklis-Van Roy algorithm. 

In this case, our algorithm remedies the decrease of performance experienced 

by regression-based methods when the number of basis functions, or regressands, 

needs to be quite large, because of path-dependence. 

2 - A Table Method For Random Variate Generation 

˙Ismail Ba¸so˘glu, Industrial Engineering, Bogazici University, 

Bogazici University Industrial Engineering Department, 34342 

Bebek, Istanbul, 34342, ˙Istanbul, Turkey, 

ismail.basoglu@boun.edu.tr, Wolfgang Hörmann 

For generating random variates from non-standard distributions, we need universal 

algorithms. In this research, we come up with a universal algorithm 

"Polynomial Density Inversion (PDI)’. We try to see if it is competitive with 

existing methods with respect to simplicity, speed and other performance criteria. 

Mainly, the method approximates the density with piecewise polynomials. 

The algorithm is complicated, yet we can obtain outstanding approximations 

with small tables. Marginal execution time is also small which makes the PDI 

a preferable method for a large number of random variates. 

286 

� WD-42 


3.1.07 

Cops and Robber Games 



Chair: Nancy Clarke, Mathematics and Statistics, Acadia University, 

12 University Avenue, B4P2R6, Wolfville, Nova Scotia, Canada, 

nancy.clarke@acadiau.ca 

1 - Characterizations of k-copwin graphs 

Nancy Clarke, Mathematics and Statistics, Acadia University, 12 

University Avenue, B4P2R6, Wolfville, Nova Scotia, Canada, 

nancy.clarke@acadiau.ca, Gary MacGillivray 

We give two characterizations of the graphs on which k cops have a winning 

strategy in the game of Cops and Robber. These generalize the corresponding 

characterizations that are known in the one cop case. In particular, we give a 

relational characterization of k-copwin graphs, for all finite k, and then use this 

characterization to obtain a vertex elimination order characterization of such 

graphs. Instead of the elimination order being of the vertices of the given graph 

G as in the one cop case, it is an ordering of the vertices of the (k+1)-fold categorical 

product of G with itself. Most of our results hold for variations of the 

game and some of them extend to infinite graphs. 

2 - The Watchman Problem 

Bert Hartnell, Mathematics & Computing Science, Saint Mary’s 

University, 923 Robie St., B3H 3C3, Halifax, Nova Scotia, 

Canada, Hartnell@smu.ca 

Although a security firm may want to to have all nodes in a network monitored 

at all times (by a person or sensor either at the node itself or adjacent) this may 

be too expensive. This gives rise to the watchman problem where a person traverses 

the graph (returning to the starting point) in such a way that every node 

is either in this closed tour or adjacent to it. In general, given a network G 

and an integer t, we would like to determine the minimum number of watchmen 

needed, and their rounds, so that the maximum time that any node is not 

monitored is t. Limited progress will be described. 

3 - Graph searching games for the WDM reconfiguration 

problem 

David Coudert, Mascotte, CNRS/INRIA/UNSA, 2004 route des 

Lucioles, B.P. 93, 06902, Sophia Antipolis, France, 

David.Coudert@sophia.inria.fr 

The routing reconfiguration problem in WDM networks is to schedule the 

switching’s of a set of lightpaths from one routing to a new predetermined 

one. This problem is modeled as a digraph processing game, closely related to 

graph searching games, in which a team of agents is aiming at clearing, or processing, 

the vertices of a digraph. In this talk, we will survey the main results 

on digraph processing games, and in particular the complexity and hardness 

of optimizing tradeoffs between the total number of agents used and the total 

number of vertices occupied by an agent during the strategy. 

4 - Graph Searching and Graph Decompositions 

Nicolas Nisse, MASCOTTE, INRIA, I3S(CNRS/UNS), INRIA, 

2004 routes des Lucioles, 06902, Sophia Antipolis, France, 

nicolas.nisse@sophia.inria.fr 

In graph searching, a team of mobile agents must catch a fugitive hidden in a 

graph. Many versions of search problems have been considered that all look for 

a strategy allowing to catch the fugitive using the fewest number of agents. We 

briefly survey the numerous research directions in this field. Then, we focus 

on the relationship between search games and graph decompositions. Namely, 

we explain how search games provide an algorithmic interpretation of pathand 

tree-decompositions. This point of view allowed us to obtain new duality 

results generalyzing those in the litterature.

� WD-43 


8.2.02 

Life Insurance, Risk Management & OR 

Stream: Life Insurance, Risk Management & OR 


Chair: Angelika May, Institut für Mathematik, Carl von Ossietzky 

Universität Oldenburg, 26111, Oldenburg, Germany, 

angelika.may@uni-oldenburg.de 

1 - Credit Risk approach to value R&D projects in the pharmaceutical 

industry 

Nora Lisse, Institut für Mathematik, Universität Oldenburg, 

Carl-v.-Ossietzky-Str. 9-11, 26111, Oldenburg (Oldb), Germany, 

nora.lisse@uni-oldenburg.de, Angelika May 

R&D projects in the pharmaceutical industry are exposed to a high risk of default 

either due to the failure of a research phase or to a poorly performing 

commercialization of the produced drug. Instead of using a standard real option 

approach we derive the value of these risks with two different approaches 

from credit risk theory. In this presentation we focus on the valuation of the 

market risk with a first-passage time model and illustrate our adaption of this 

method through real data examples. 

2 - Intersections of ruin probabilities with respect to the 

initial surplus 

Tatjana Slijepcevic-Manger, Faculty of Civil Engineering, 

University of Zagreb, Fra Andrije Kacica-Miosica 26, 10000, 

Zagreb, Croatia, tmanger@grad.hr 

In this paper we study intersections of ruin probability functions with respect to 

the initial surplus for two risk models. The insurance company could use one 

model with smaller probability of ruin up to the intersection and then change 

to the other model in order to minimize the probability of ruin. 

3 - On Devising an Alarm System for Insurance Companies 

Shubhabrata Das, QMIS, Indian Institute of Management 

Bangalore, Faculty Block C 2nd Floor, Bannerghatta Road, 

560076, Bangalore, India, shubho@iimb.ernet.in, Marie Kratz 

One way of risk management for an insurance company is to develop an early 

and appropriate alarm system before the possible ruin, defined through the status 

of the aggregate risk process which in turn is determined by premium accumulation 

as well as claim settlement out-go for the insurance company. This 

paper designs an effective alarm system with a fair measure of effectiveness. 

We present comparisons of performances when the loss severity has Exponential, 

Pareto or discrete Logarithmic distribution. 

� WD-44 


8.2.03 

Dynamical Systems and Mathematical 

Modelling in OR II 













1 - Automated Traders in Commodities Electronic Markets 

Fodil Laib, Dépt. of Operational Research, University of Bejaia, 

Targa Ouzemour, Bejaia, 16000, Bejaia, fodil.laib@cevital.com 


Our purpose is to automatize trading of commodities by using automated agents 

instead of human traders. In this setup, each producer and consumer has a trading 

agent in the market arena trading on his behalf. The agent is fed with a 

stream of supply and demand forecasts, then it uses a parameterized trading 

strategy to build sell or buy orders. The performance of this trading system is 

measured by computing the distance between the generated price pattern compared 

to a benchmark pattern. We suggested a simulation approach to find the 

optimal parameters of the trading strategy. 

2 - A discussion on the Hamiltonian dynamical systems, 

integrable systems and dynamical system of diffusion 

of information 

Selma Belen, Mathematics and Computer, CAG University, 


Mersin, Turkey, selmabelen@cag.edu.tr, Gerhard-Wilhelm 

Weber, Erik Kropat 

In this paper, we discuss on the mechanism of diffusion which is not identical to 

the mechanism of diffusion observed in Hamiltonian systems. As the reversibility 

is certainly an important symmetry property in the context of Hamiltonian 

systems, it is discussed that if the reversibility is equivalent to a spatial symmetry. 

3 - A Dynamic Model for Evaluating the Effects of Piracy in 

Film Industries 

Sercan Oruc, Industrial Engineering, Middle East Technical 

University, Turgut Ozal Mah. 15. Cad., Utku Sit. 7/13 Batikent, 

06370, Ankara, -, Turkey, sercanoruc@gmail.com 

Film industry, as a type of creative industries, constitutes a dynamic environment 

where uncertainty is at high levels. This complexity renders the more 

traditional OR models somewhat ineffective, and thus, requires a dynamic analysis. 

In this study, a model showing the dynamics of film exhibition is given. 

The model shows the interactions within and between the theatrical and the 

DVD sales channels. The possible effects of piracy to the model are discussed, 

using the inferences obtained by the created model. The model is examined 

with scenario and sensitivity analysis. The model also can be extended for the 

whole film industry, or for some other creative industries like the publishing 

industry. 

4 - A System Dynamic Approach in Managing the Levels of 

Neurotransmitter for Patients with Chronic Depression 

Armagan Bayram, Industrial Engineering, Istanbul Kultur 

University, IKU Atakoy Campus Room 216 Bakirkoy, 34156, 

Istanbul, armagannbayram@yahoo.com, Dicle Cevizci, Canan 

Herdem 

A dynamic model is developed to analyze depression, caused by low levels of 

neurotransmitters between synapses. Neurotransmitter levels should be kept 

at a certain level. Antidepressants are used to increase the serotonin levels by 

closing these pumps. However, overdose of these antidepressants increases 

serotonin levels so much and triggers other diseases. Many unexpected cases 

can be prevented by seeing the behavior. In this model, behavior of disease and 

difference between healthy people are observed. 

� WD-45 


8.2.12 

Logistics and Promotions Management 

Stream: Demand and Supply in Consumer Goods and 

Retailing 


Chair: Rob Broekmeulen, OPAC, TU Eindhoven, P.O. Box 516, Pav. 

E10, 5600 MB, Eindhoven, -, Netherlands, 


1 - Modelling a logistics problem in retailing industry under 

uncertainty in an operational level 

Yousef Ghiami, Management, University of Southampton, 

Building 2, University Road„ Highfield, Southampton, SO17 

287


1BJ, Southampton, United Kingdom, yg5g09@soton.ac.uk, Yue 

Wu 

Uncertainties are the main sources of risk, making the planning for the future, 

challenging. In supply chain management field, researchers have been investigating 

uncertain factors (e.g. demand and raw material cost) and put effort 

to tackle these uncertainties and the consequent risks. In this research a threeechelon 

supply chain of a perishable product (e. g. dairy products) is considered, 

with stochastic demand at the stores. In this model the total logistics cost 

is considered, including transportation costs, production costs in plants and inventory 

costs in plants and stores. In order to deal with the demand uncertainty, 

the problem has been modelled with Robust Optimization approach. 

2 - Hedging Retail Promotions - Bidding Coordinating Contracts 

into a Pull Supply Chain with Retailer Power 

Arnd Huchzermeier, Production Management, WHU - Otto 

Beisheim School of Management, Burgplatz 2, 56179, Vallendar, 

Germany, ah@whu.edu, Andreas Breiter 

Demand risk from price promotions creates waste in the supply chain. We 

show that offering option contracts in addition to spot contracts can achieve 

coordination even in the presence of a dominant retailer. Employing a model 

of stockpiling consumers, we formulate a stochastic process of promotional demand. 

Based on point-of-sales data from a major German supermarket chain, 

we fit our model and quantify the additional profits achieved. A portfolio of 

supply contracts can simultaneously reduce i) out-of-stock situations and ii) 

end-of-period coverage, clearing the channel for future promotions. 

3 - Markdown Optimization in a Retail Chain under Demand 

Substitution 

Ozlem Cosgun, Industrial Engineering, Istanbul Kultur 

University, Istanbul Kultur Universitesi endustri muhendisligi 

bolumu atakoy kampus, 34156, Bakirkoy, Istanbul, Turkey, 

ozlem_ince@hotmail.com, Ufuk Kula 

We consider the markdown optimization problem faced by a leading apperal 

retail chain in Turkey. Markdown policies for product groups having a significant 

crossprice elasticity among eachother should be jointly determined,which 

makes finding optimal policies for products computationally intractable as the 

number of products increases. We first decompose sales into three compenents: 

marked-down price,substitution and time effects by using MNL model. Then 

we formulate the problem as an MDP and use approximate dynamic programming 

approach to solve it and provide insights on the markdown policy. 

4 - The Use of Generalized Additive Models for Demand 

Forecasting 

Amirhossein Sadoghi, Department of Management and 

Engineering, linkoping University, Rydsvägen 242 A LGH 17, 

584 34 , Linkoping, Sweden, amisa242@student.liu.se 

Forecasting using time series is often based on linear regression. In this paper, 

we use the Generalized Additive Models (sum of smooth functions) instead 

of the linear regression to analyze the behavior of the system in the demand 

forecasting of make-to-stock problem. Main focus is on the existence of the 

nonparametric analogue of multicollinearity, concurvity in sales data. This 

provides insight in to managing interdependent demands. We also investigate 

whether the specific temporal patterns are likely to affect the statistical forecasts 

of the sales history 

� WD-46 


8.2.14 

Approximation of Probabilities for Exotic 

and Compound Options 

Stream: Numerical Methods in Finance 





Chair: Susanne Griebsch, School of Finance and Economics, 

University of Technology Sydney, PO Box 123 Broadway, 2007, 

Sydney, NSW, Australia, susanne.griebsch@uts.edu.au 

288 

1 - The Evaluation of European Compound Option Prices 

under Stochastic Volatility using Fourier Transform 

Techniques 

Susanne Griebsch, School of Finance and Economics, University 

of Technology Sydney, PO Box 123 Broadway, 2007, Sydney, 

NSW, Australia, susanne.griebsch@uts.edu.au 

This study focuses on compound option pricing under Heston’s stochastic 

volatility dynamics. The compound option value is not only sensitive to 

changes of the underlying asset, but also to future changes of volatility. We 

develop a numerical pricing algorithm to solve this problem. It exploits that the 

representation of the compound option value can be divided into a difference of 

probabilities under two different probability measures. Approximations of the 

probabilities are obtained through a FFT-technique. The results are compared 

with other numerical methods, showing promising results. 

2 - The Evaluation of Swing Contracts with Regime Switching 

Boda Kang, School of Finance and Economics, University of 

Technology Sydney, PO Box 123, Broadway, 2007, Sydney, New 

South Wales, Australia, Boda.Kang@uts.edu.au, Carl Chiarella, 

Les Clewlow 

A typical gas swing contract is an agreement between a supplier and a purchaser 

for the delivery of variable daily quantities of gas, between daily limits, 

over a number of years at a set of contract prices. The main constraint of swing 

contracts are that in each gas year, there is a minimum volume of gas for which 

the buyer will be charged at the end of the year. We price such swing contracts 

for a regime-switching gas forward price curve. With the help of the recombing 

pentanonial tree, we are able to evaluate the prices, the optimal strategies and 

the hedge ratios of the swing contracts. 

3 - A new unscented Kalman filter with higher order 

moment-matching 

Ksenia Ponomareva, Brunel University, United Kingdom, 

ksenia.ponomareva@brunel.ac.uk, Paresh Date, Zidong Wang 

A new approximate Bayesian algorithm is proposed which generates sample 

points and probability weights that match the predicted values of marginal 

skewness and kurtosis of the unobserved state variables, in addition to their 

mean and covariance matrix. Its performance is illustrated by an empirical example 

of yield curve modeling with financial market data. This algorithm is 

a useful alternative to the extended Kalman filter (due to improved accuracy) 

and particle filter (due to significantly reduced computation). Possible practical 

applications include forecasting of macroeconomic variables. 

� WD-47 


8.2.16 

Data Mining in Early Warning Systems II 

Stream: Data Mining in Early Warning Systems 


Chair: Ozlem Ilk, Statistics, Middle East Technical University, 

Middle East Technical University, Dept. of Statistics, Office No: 

134„ Ankara, Turkey, oilk@metu.edu.tr 

1 - Risk analysis study of maritime traffic in the strait of 

istanbul 

Birnur Ozbas, Industrial Engineering Department, Bogazici 

University, Guney Kampus Muhendislik Binasi, Bebek, 34342, 

Istanbul, Turkey, birnur@ozbas.com.tr, Ilhan Or, Tayfur Altiok, 

Ozgecan S. Uluscu 

In this paper, results of a simulation-based risk analysis study for the Strait of 

Istanbul is presented. A simulation model for the transit traffic in the Strait 

(which considers traffic rules, vessel profiles, pilotage services, meteorological/geographical 

conditions) has been developed. Regarding risk assessment, 

two sets of factors are sought for during each transit: the probability of an 

accident and the potential consequences of this accident, both conditioned on 

various triggering events. Then, as a simulation run proceeds, the risks generated 

by each transit vessel are assessed.

2 - Trouble Condition Sign Mining for Hydroelectric Power 

Plants 

Takashi Onoda, System Engineering Lab., CRIEPI, 2-11-1, 

Iwado Kita, Komae-shi, 201-8511, Tokyo, Japan, 

onoda@criepi.denken.or.jp 

Kyushu Electric Power Co.,Inc. collects different sensor data (hereafter, operation 

data) to maintain the safety of hydroelectric power plants while the plants 

are running. It is very rare that trouble conditions occur in the power equipment. 

In this situation, we have to find trouble condition sign. In this paper, we 

consider that the rise inclination of special unusual condition data gives trouble 

condition sign. We propose an interactive trouble condition sign mining 

method for hydroelectric power plants by using one class and normal support 

vector machine. This paper also shows that the proposed method can find a 

trouble condition sign of bearing vibration from the real operation data. 

3 - Mahalanobis-Taguchi System for the Prediction of Pressure 

Ulcers Development in Surgical Patients 

Chao-Ton Su, Dept. of Industrial Engineering & Engineering 

Management, National Tsing Hua University, 101, Sec. 2, 

Kuang-Fu Rd.„ 30013, Hsinchu, Taiwan, ctsu@mx.nthu.edu.tw, 

Li-Fei Chen, Yan-Cheng Chen 

The Mahalanobis-Taguchi system (MTS) is a diagnostic/forecasting tool integrating 

Mahalanobis distance, orthogonal arrays and signal-to-noise ratio. This 

study aims to employ MTS to predict the pressure ulcers development in surgical 

patients and identify risk factors from data from patients during surgical 

procedures. Here, MTS obtained better values in index of sensitivity and gmeans 

than other analytical methods (Logistical regression, decision trees). We 

conclude that MTS is an effective approach for the diagnosis investigated. 

4 - Statistical early warning systems for company failure 

Ozlem Ilk, Statistics, Middle East Technical University, Middle 

East Technical University, Dept. of Statistics, Office No: 134„ 

Ankara, Turkey, oilk@metu.edu.tr, Deniz Akinc, Murat Cinko, 

Didem Pekkurnaz 

Foreseeing the early signs of financial failure is important for both the economical 

development of the country and for the self - evaluation of individual 

firms. In this study, publicly available panel data are collected from Istanbul 

Stock Exchange and they are investigated with the goal of detecting likely to 

fail companies. Annual data on 146 industry firms are considered between the 

years of 1999 and 2002. Both logistic regression and Marginalized Transition 

Random Effects Models (MTREM) are used as early warning systems. By 

these models, financial success probabilities for each company are calculated 

at time t by using the financial statements at time t-1; and the factors related 

to financial failure are determined. Depending on the year and response type, 

the correct classification rates in logistic regression models range between 62% 

and 72%. On the other hand, MTREM resulted in higher correct classification 

rates, which are between 79% and 100%. 

EURO 24 - Lisbon 2010 WE-05 

Wednesday, 15:40-17:00 

� WE-02 


3.2.14 

EURO Management Science Strategic 

Innovation Prize (MSSIP 2010) on the topic 

of Optimization in Telecommunications 

Stream: EURO Management Science Strategic 

Innovation Prize (MSSIP 2010) on the topic of 

Optimization in Telecommunications 


Chair: Eric Gourdin, CORE/TPN/TRM, Orange Labs, 38 rue du 

General Leclerc, 92794, Issy-les-Moulineaux, France, 

eric.gourdin@orange-ftgroup.com 

1 - Mssip 2010 

Mssip 2010 Winner, EURO, 0000, Brussels, 

MSSIP2010winner@gmail.com 

The MSSIP Prize is awarded for an outstanding innovative contribution to 

Management Science/OR each year when a EURO Conference takes place. 

This time the subject is Optimization in Telecommunications and the prize is 

awarded at the EURO 2010 Conference taking place in Lisbon (Portugal). The 

prize is intended to recognize the role of Operational Research/Management 

Science in the context of modern telecommunications. The winner is announced 

at this session. The prize is sponsored by SAP AG (12000 Euro). 

� WE-05 


3.2.16 

Tools for metaheuristics 



Chair: Lars Magnus Hvattum, Industrial Economics and Technology 


Alfred Getz veg 3, Sentralbygg 1, N-7491 Trondheim, Norway, 7491, 

Trondheim, Norway, lars.m.hvattum@iot.ntnu.no 

Chair: Frédéric Gardi, e-lab, Bouygues SA, 40 rue Washington, 

75008, PARIS, France, fgardi@bouygues.com 

1 - Toward Local Search Programming: LocalSolver 1.0 

Frédéric Gardi, e-lab, Bouygues SA, 40 rue Washington, 75008, 

PARIS, France, fgardi@bouygues.com, Thierry Benoist, 

Bertrand Estellon, Karim Nouioua 

This paper introduces Local Search Programming (LSP), as a paradigm allowing 

the practitioner to focus on the modeling of the problem using a simple 

formalism, and then to let its actual resolution to a solver based on efficient 

and reliable local-search algorithms. In other words, our goal is to offer a 

model-and-run approach to combinatorial optimization problems which are out 

of reach of existing Integer/Constraint Programming autonomous solvers. In 

this paper, LocalSolver 1.0 is presented, first software realization of our works 

on this subject. 

2 - A Theoretical and Empirical Study of Evolutionary 

Squeaky Wheel Optimisation (ESWO) 

Jingpeng Li, Computer Science, The University of Nottingham, 

Wollaton Road, NG8 1BB, Nottingham, United Kingdom, 

jpl@cs.nott.ac.uk, Andrew J. Parkes, Edmund Burke 

The SWO approach is a new metaheuristic based on the evolution of single 

solution. It undertakes search via iterative disruption, improvement and construction. 

Our experiments on various domains have demonstrated its efficiency 

and effectiveness. By building its Markov chain model and undertaking a matrix 

analysis, we prove its global optimality and derive its convergence rate. 

By studying the stationary distribution with an example, we gain insight into 

how to embed domain knowledge into the search, and reveal some interesting 

properties (e.g. a non-monotonic increase with the fitness). 

289

WE-06 EURO 24 - Lisbon 2010 

3 - Graphic Processor Unit Accelerated Simulated Annealing 

Framework 

Fatih Nar, Mathematics and Computer, Çankaya University, 

Ogretmenler Caddesi, No:14 Balgat, 06530, Ankara, Turkey, 

fatihnar@cankaya.edu.tr, Erdal Yilmaz 

In this study a generic framework for the parallel implementation of Simulated 

Annealing (SA) method is proposed. In proposed method search space 

is spanned by local SA search agents (SASAs) where optimum distribution of 

agents in search space is obtained using k-means clustering algorithm. Clusters 

are formed dynamically based on spatial position of SASAs and their fit values. 

Cost functions are categorized by GPU implementation issues and recommendations 

are given as a guideline. 

4 - Robust Optimization via Robust Local Search 

Kwong Meng Teo, National University of Singapore, Singapore, 

kwongmeng@alum.mit.edu, Dimitris Bertsimas 

A robust optimization method admissible to convex/nonconvex problems and 

problems not explicitly described with convex mathematical functions will be 

presented. The Robust Local Search algorithm operates directly on response 

surfaces of the cost/constraint functions and assumes only a black-box problem 

description; thus, it can be used in most real-world applications. We shall 

introduce the concept of robust local minima, discuss the convergence properties 

of the algorithm towards these minima, and report the results in engineering 

design and healthcare applications. 

� WE-06 


8.2.30 

DEA Application VIII — Software 



Chair: Ali Emrouznejad, Aston Business School, Aston University, 

B4 7ET, Birmingham, United Kingdom, a.emrouznejad@aston.ac.uk 

1 - An algorithm for DEA 

José Dulá, School of Business, Virginia Commonwealth 

University, Richmond, United States, jdula@vcu.edu 

This short tutorial will present a faster output-sensitive frame-based algorithm 

for DEA, which is based on a two-phase procedure. The first phase identifies 

the extreme efficient entities, the frame, of the production possibility set. The 

frame is then used in a second phase to score the rest of the entities. The new 

procedure applies to any of the four standard DEA returns to scale. It also 

imparts flexibility to a DEA study since it postpones the decision about orientation, 

benchmarking measurements, etc., to after the frame has been identified. 

2 - FURNAS RCM program’s validation and quantitative 

corrections using Data Envelopment Analysis 

Marcio Mariano Junior, Production Engineering, UFMG, Brazil, 

mj.marcio@gmail.com, Magno Silverio Campos, Joao Flavio F. 

Almeida 

This paper validate an Reliability Centered Maintenance program of a Brazilian 

state owned electrical utility using an operations research model of Data 

Envelopment Analysis to analyze and rank their 500kV circuit breaker failure’s 

mode, at that time using a simple FMEA and empirical engineering analysis. 

This model evaluates the relative performance and generate an pure quantitative 

priorities rank, without empirical inputs and corrections. Today a high number 

of engineers are retired and the model presented a correlation higher than 90% 

over the samples, been approved by the users. 

3 - New Model Designer of DEAOS 

Mohammad Reza Alirezaee, School of Mathematics, Iran 

University of Science and Technology, Hengam St., Resalat Sq., 

16846, Tehran, Iran, Islamic Republic Of, mralirez@yahoo.com, 

Ali Niknejad, Nassrin Alirezaee 

Wide variety of DEA models, which is growing day by day, creates a challenge 

to those who want to implement them. Complexities of the mathematical models 

make development of the models extremely hard. This problem is of great 

importance in DEAOS.Com which has a goal of delivering all the DEA models. 

To address this issue in DEAOS, a new model designer has been introduced. 

This engine allows the researchers to define models in an easy-to-use web interface. 

Doing so, will not just allow the researchers define their own models 

in the system but it will also simplify the verification process. 

290 

4 - Data Envelopment Analysis software for the advanced 

users 

Ali Emrouznejad, Aston Business School, Aston University, B4 

7ET, Birmingham, United Kingdom, 

a.emrouznejad@aston.ac.uk, Emmanuel Thanassoulis 

This paper presents software that takes its features closer to the latest developments 

in the DEA literature. The new software addresses a variety of issues 

such as: Assessments under a variety of possible assumptions of returns 

to scale including NIRS and NDRS; truly unlimited number of assessment 

units (DMUs); Analysis of groups of data by estimating automatically separate 

boundaries by group; Malmquist Index and its decompositions; Super 

efficiency; Automated removal of super-efficient outliers under user-specified 

criteria; Cross efficiency; allocative efficiency and Bootstrapping. 

� WE-08 


6.1.36 


Scheduling III 



Chair: Ayse Elvan Bayraktaroglu, Industrial Engineering, Istanbul 

Technical University, Isletme Fakultesi, Macka, 34367, Istanbul, 

bayraktaroglu@itu.edu.tr 

1 - Assembly Line Balancing with Outscoring 

Rifat Gürcan Özdemir, Industrial Engineering Department, 

Istanbul Kültür University, Atakoy Campus, Atakoy-Bakirkoy, 

34156, Istanbul, Turkey, rg.ozdemir@iku.edu.tr, Ufuk Kula 

Outsourcing has become a popular company strategy in recent years. In this paper, 

we consider a single product assembly line design and balancing problem, 

in which the manufacturer has an option of outsourcing a pre-determined set 

of product’s components. We develop a mathematical model that determines 

the optimal number of stations, solves the line-balancing problem. In addition, 

the model determines which components are to be outsourced. We apply the 

model in a home appliance product, and perform a numerical study to show the 

benefit of outsourcing. 

� WE-13 


2.2.21 

Hub Location 



Chair: Elena Fernandez, Statistics and Operations Research, 

Technical University of Catalonia, Campus Nord, C5-208, Jordi 

Girona, 1-3, 08034, Barcelona, Spain, e.fernandez@upc.edu 

1 - Models for a single-assignment hub location problem 

with capacity levels 

Isabel Correia, Departamento de Matemática- CMA, 

FCT-Universidade Nova de Lisboa, Monte da Caparica, 

2829-516, Caparica, Portugal, isc@fct.unl.pt, Stefan Nickel, 

Francisco Saldanha-da-Gama 

This paper considers an extension of the classical single-assignment hub location 

problem in which the capacity level of each potential hub must be chosen 

from a set of available capacities. For this problem different mixed integer linear 

programming models are presented. Some additional inequalities and preprocessing 

tests are proposed with the goal to enhance the models. We report 

the computational experience performed with a commercial solver in a battery 

of test problems in order to evaluate the performance of the different models 

discussed.

2 - Multimodal hub location and hub network design 

Sibel A. Alumur, Industrial Engineering Department, TOBB 

University of Economics and Technology, Sogutozu cad. No:43, 

Sogutozu, 06560, Ankara, Turkey, salumur@etu.edu.tr, Bahar 

Yetis Kara, Oya Ekin-Karasan 

In this new hub location problem, we include the possibility of using different 

hub links and allow for different transportation modes between hubs, and for 

different types of service time promises between origin-destination pairs while 

designing the hub network. In addition, we jointly consider transportation costs 

and travel times in this multimodal problem. We propose a linear mixed integer 

programming model together with some sets of effective valid inequalities 

and an efficient heuristic. Computational analysis is presented on the Turkish 

network data set. 

3 - The Dynamic Uncapacitated Hub Location Problem 

Ivan Contreras, Canada Chair in Distribution Management, HEC 

Montreal, 3000 chemin de la Cote-Sainte-Catherine, H3T 2A7, 

Montreal, Quebec, Canada, ivan.contreras@cirrelt.ca, 

Jean-François Cordeau, Gilbert Laporte 

This paper presents the Dynamic Uncapacitated Hub Location Problem which 

consists in selecting a set of hubs to be established and the routing of flow 

through the network, while minimizing the total cost over a finite time planning 

horizon. The costs include those for the location, operation and closing 

of hubs over time, and the costs of routing the flow. We propose a branch-andbound 

algorithm that uses a Lagrangean relaxation to obtain good bounds at 

the nodes of the tree. Numerical results on a battery of instances with up to 100 

nodes and 10 time periods are reported. 

4 - Hub location problems with role dependent objectives 

Justo Puerto, Estadistica e I.O., Universidad de Sevilla, Facultad 

de Matematicas, 41012, Sevilla, Spain, puerto@us.es, Antonio 

Manuel Rodríguez-Chía, Ana Bel Ramos-Gallego 

Although hub location models have been analyzed from the sum, maximum 

and coverage point of views, as far as we know, they have never been considered 

under an alternative unifying point of view. In this paper we consider new 

formulations, based on the ordered median objective function, for hub location 

problems with new distribution patterns induced by the different users’ roles 

within the Supply Chain network. 

� WE-14 


2.2.15 

Industrial Applications in Risk Management 

Stream: IBM Research Applications 


Chair: Eleni Pratsini, IBM Zurich Research Lab, Saeumerstrasse 4, 

8803, Zurich, Switzerland, pra@zurich.ibm.com 

Chair: Marco Laumanns, IBM Research Zurich, 8803, Rueschlikon, 

Switzerland, mlm@zurich.ibm.com 

1 - From Optimal Solutions to Smart Decisions 

Alain Chabrier, IBM España S. A., Santa Hortensia 26-28, 

28002 , Madrid, Spain, achabrier@es.ibm.com 

In this talk we describe the challenges faced when developing custom optimization 

based solutions. We will also summarize some keys to success that 

we have learnt from developing such applications for many years. Finally, we 

present how a platform like IBM ILOG ODM Enterprise may help answering 

these challenges. 

2 - Closing the gap between tactical ERP / MRP planning 

and operational execution in production industries — a 

simulation-based approach 

Ulrich Schimpel, Business Optimization, IBM Research, 

Saeumerstrasse 4, 8803, Rueschlikon, Switzerland, 

uschimpel@gmx.net, Satyadeep Vajjala, Manuel Parente 

Our case study looks at a multi-stage microchip production process and addresses 

possibilities to close the gap between commonly deterministic MRP 

schedules and the multi-constrained stochastic LEAN execution. Thereby, we 

give a deeper understanding of effects like variability in the demand, yield, and 

in the lead times. We highlight strategies to mitigate those risks and effectively 

narrow the aforementioned gap. Our approach includes operational aspects 

like running production sites at different TAKTs, co-product relationships, and 

various capacity constraints. 


3 - A Markov decison model for strategic resource management 

of power grid operators 

Michael Guarisco, Institute for Operations Research, ETH 

Zurich, Raemistrasse 101, 8092, Zurich, Switzerland, 

guarisco@ifor.math.ethz.ch, Marco Laumanns, Rico Zenklusen 

Based on regulatory requirements in electricity markets, power grid operators 

try to balance costs and quality of supply. The restoration time after incidents 

influences the quality of supply and depends on the availability of resources. 

We present a Markov decision model of a power supply system with endogenous 

restoration times. For each state of the power grid, the available resources 

are assigned to the failed components such that the expected average energy not 

supplied is minimized. The results may support strategic decisions of power 

grid operators in resource management. 

4 - Web-based Expert Elicitation in Bayesian Networks 

Lea Deleris, IBM, Dublin 15, Mulhuddart, Ireland, 

lea.deleris@ie.ibm.com, Debarun Bhattacharjya, Bonnie Ray 

Bayesian networks are increasingly popular for representing causal relationships 

in risk analysis applications. One of the challenges they present is the 

elicitation of the parameters of the underlying conditional probability distributions. 

We present a web-based tool that facilitates this elicitation process, in 

particular for distributed teams. We also discuss analytical methods for determining 

the order in which parameters are to be elicited, in order to maximize 

the information salvaged in case of early drop-off from experts, a problem particularly 

relevant to web-based elicitation. 

� WE-15 


2.2.12 

Vehicle Routing and Set Covering Models 



Chair: Rita Macedo, Departamento de Produção e Sistemas, 

Universidade do Minho, Universidade do Minho, Escola de 

Engenharia, Departamento de Produção e Sistemas, Campus de 

Gualtar, 4710-057, Braga, Portugal, rita@dps.uminho.pt 

1 - Consumer Dispersion and Logistics Costs in Various 

Distribution Systems 

Marcel Turkensteen, Business Studies, ASB, Fuglesangs Alle 4, 

8210, Aarhus V, Denmark, matu@asb.dk 

We address the relationship between the geographical dispersion of a set of demand 

points and the expected logistics costs. This is relevant in the strategic 

marketing decision which groups of consumers to target. We devise quickly 

computable measures for the logistics costs. In our experiments, dispersed sets 

of demand points are created. For various types of distribution systems, expected 

logistics costs are computed using continuous approximation, location 

and routing methodologies. We find that the average distance between locations 

is an effective estimate of the logistics costs. 

2 - On Generation of Routes for Set-Covering-Based Approach 

for Solving Capacitated VRP 

Milan Stanojevic, Faculty of Organizational Sciences, University 

of Belgrade, Jove Ilica 154, 11000, Belgrade, Serbia, 

milans@fon.rs, Gordana Savic 

Real-world routing problems are very hard to be solved. One possible approach 

is to solve corresponding set-covering problem with limited number of predefined 

routes. With a controllable route generation algorithm one can influence 

the execution time and the quality of solution. We propose a route generation 

algorithm based on a savings heuristic similar to Clarke-Wright algorithm. 

Some computational results will be presented. 

291


3 - Column generation based heuristic with heuristic pricing 

for capacitated vehicle routing problem 

Cristián Cortés, Civil Engineering Department, Universidad de 

Chile, Blanco Encalada 2002, 5th floor, Santiago, Chile, 

ccortes@ing.uchile.cl, Pablo A. Rey 

We show a heuristic algorithm for a distribution problem faced by a beer producer. 

The problem comprises trucks’ fixed costs, delivery benefits, time windows 

and routes with more than 40 clients. The relaxed MP is solved to optimality 

by CG and a set-covering problem is built with selected columns. The 

IP is solved for routing. A GRASP is used to price out new columns; when 

it fails, an Elementary SP is used to either checking optimality or finding new 

routes. Computational experiments are shown 

4 - Relaxation induced methods based on column generation 

for vehicle routing problems 

Rita Macedo, Departamento de Produção e Sistemas, 

Universidade do Minho, Portugal, rita@dps.uminho.pt, Cláudio 

Alves, J. M. Valério de Carvalho, Saïd Hanafi 

We address the resolution of vehicle routing problems through hybrid procedures 

that combine column generation, branch-and-bound and relaxation based 

heuristics. In particular, we combine the hybrid linear programming based algorithm 

proposed by Hanafi and Wilbaut (2009) for mixed integer programming 

problems with branch-and-price. These heuristics are convergent. They 

consist in solving iteratively the linear relaxation of the problem, and in deriving 

upper bounds. The linear relaxation corresponds to the well-known column 

generation model. Computational results are reported. 

� WE-17 


1.3.14 

Recent OR Advances by Statistics, 

Probability and Performance Measures 



Chair: Pakize Taylan, Mathematics, Dicle University, 21280, 

Diyarbakır, Diyarbakir, Turkey, ptaylan@dicle.edu.tr 




1 - New empirical likelihood methods with applications 

Gregory Gurevich, Industrial Engineering and Management, 

SCE - Shamoon College of Engineering, Balik Bazel, 84100, 

Beer Sheva, Israel, gregoryg@sce.ac.il, Yossi Hadad 

The objective of this research is to propose and examine various nonparametric 

statistical procedures based on empirical likelihood ratio methodologies. If 

certain key assumptions are met, one can show parametric likelihood methods 

are very powerful and efficient statistical tools. However, when these key 

assumptions are not met the parametric approach may be not optimal. We develop 

various ’artificial’ or ’approximate’ likelihoods’ techniques, which are 

both robust and efficient. These methods have a wide variety of applications in 

engineering and social science experiments. 

2 - Methods based on mathematical optimization for semisupervised 

data classification 

Burak Ordin, Mathematics, Ege University, Ege University 

Science Faculty, Department of Mathematics, 232, izmir, 

bornova, Turkey, burak.ordin@ege.edu.tr, Nur Uylas 

There has been an increase of interest for semi-supervised learning recently, 

because of the many datasets with large amounts of unlabeled examples and 

only a few labeled ones. In this paper we have compared two semi-supervised 

algorithms based on codifferential method and quasi-secant method for solving 

data classification problems. The algorithms compute hyperplane(s) to seperate 

two sets with respect to some tolerance. An error function is formulated and 

an algorithm for its minimization is expressed. We present results of numerical 

experiments using several UCI test data sets. 

292 

3 - Lottery Data Analysis Using Hierarchical Decision- 

Making Model 

Ulas Beldek, Eletronic and Communication Engineering 

Department, Çankaya University, Cankaya Universitesi, 

Ogretmenler caddesi no:14, Yuzuncuyi, Balgat, Cankaya, 

Ankara, 06530, Ankara, Turkey, u.beldek@cankaya.edu.tr, 

Kemal Leblebicio˘glu 

Lottery Sales forecasting has been addressed by using a single agent, previously. 

However, we propose a hierarchical decision-making model for the development 

of more than one agent where a lower level agent helps the development 

of higher level agent: The prediction of lower level agent is transferred 

to higher level by a performance measure. An intuitively suitable fusion equation 

combines the lower level decision with the decisions of a new developing 

structure to form the second level agent. Finally, the strength of the agents 

developed using the hierarchical model is demonstrated. 

4 - The influence of self-attribution bias on managerial incentives 

and the choice of performance measures 

Bjoern Anton, Chair of Accounting and Control, Technical 

University Munich - Germany, Arcisstraße 21, 80333, München, 

BY, Germany, bjoern.anton@wi.tum.de 

We examine the effects of self-attribution bias on a multiperiod moral hazard 

model in which an agent does not know his own ability but infers it from the 

output of his work. We compare the behavior of an agent who takes too much 

credit from high outputs to a purely rational one. Our results are: 1. The biased 

agent’s expected effort level is higher than that of a rational agent. 2. Low 

outputs may induce underconfidence / lower effort levels. 3. The level of overconfidence 

declines over time. 4. Aggregate performance measures may be 

better for biased agents. 

� WE-18 


1.3.15 

Data Mining Applications in Business 

Intelligence 

Stream: Applications in Business Intelligence and 

Knowledge discovery 


Chair: Richard Weber, Department of Industrial Engineering, 

University of Chile, Republica 701, 2777, Santiago, Chile, 

rweber@dii.uchile.cl 

1 - Business Intelligence in Electronic Commerce — Case 

Study 

Dragana Becejski-Vujaklija, Faculty of Organizational Sciences, 

University of Belgrade, Jove Ilica 154, 11000, Belgrade, Serbia, 

draganab@fon.rs 

In this work possibilities that come from the application of business intelligence 

in electronic commerce and network environment will be presented. Processing 

a large number of data required specific data structures because daily operations 

do not allow analytical processes to operate unimpeded. Relational data could 

not be used in the analytical processes, therefore this work displays a structure 

of a Data Warehouse model used in small domestic electronic commerce firm 

to record all needed data for analysis. 

2 - Forecasting and Management of Hospital Service Demand 

Richard Weber, Department of Industrial Engineering, 

University of Chile, Republica 701, 2777, Santiago, Chile, 

rweber@dii.uchile.cl, Oscar Barros, Eduardo Ferro, Carlos 

Reveco 

An efficient planning of hospital resources requires a reliable demand forecast. 

We are developing integrated management tools for several Chilean hospitals 

where based on predicted demand for several pathology types capacity 

planning for different resources is performed automatically. We compare traditional 

techniques such as linear regression and moving averages with more 

sophisticated tool such as neural networks. An out-of-sample error of about 

7% (MAPE) convinced the decision makers at the hospitals and allows us to 

install the developed systems for daily operation.

3 - Improving Execution Time and Accuracy for IP Classification 

Problems in Large Data Sets 

Jaime Miranda, Department of Management Control and 

Information Systems, Universidad de Chile, Diagonal Paraguay 

257, Chile, jmirandap@fen.uchile.cl, Richard Weber, Daniel 

Espinoza 

Many data mining applications require the analysis and classification of large 

data sets. Several methods exist for this task, being of particular interest Integer 

Programming (IP) models. However, these models’ weakness is the required 

computational time limiting their applicability to small data sets. We present 

a heuristic that uses cluster analysis as preprocessing for a reduced IP model 

achieving both, significantly lower computational time and less classification 

errors. 

4 - A 3PL providers classification model considering categorical 

variables on the use of information and communication 

technologies 

Mônica M. M. Luna, Department of Production and Systems 

Engineering, Federal University of Santa Catarina, Campus 

Universitário, Trindade, 88040-900, Florianópolis, SC, Brazil, 

monica@deps.ufsc.br, Carlos Ernani Fries 

ICT have greatly benefited the logistics industry, allowing high levels of connectivity 

between partners, promoting its differentiation and specialization. To 

characterize the service offer, a statistics and data mining based third-party logistics 

providers classification model which exclusively considers the presence 

of technological solutions through Yes/No statements is suggested. The results 

identified 3PL homogeneous clusters in the Brazilian market, equivalent 

to those models that make use of quantitative variables, usually associated with 

unreliability and difficult acquisition. 

� WE-19 



Nonsmooth Global Optimization 



Chair: Alexander Kruger, Graduate School of Information 

Technology & Mathematical Sciences, University of Ballarat, 

University Drive, Mount Helen, P.O. Box 663, 3353, Ballarat, 

Victoria, Australia, a.kruger@ballarat.edu.au 

1 - Asymptotic stability in optimal control problems with 

time delay 

Musa Mammadov, Graduate School of Information Technology 

and Mathematical Sciences, University of Ballarat, University 

Drive, Mount Helen, P.O. Box 663, 3353, Ballarat, Victoria, 

Australia, m.mammadov@ballarat.edu.au 

The problem of qualitative analysis of optimal trajectories for a special class of 

optimal control problems described by differential delay equations is considered. 

This kind of equations has attracted a significant interest in recent years 

due to their frequent appearance in a wide range of applications. They serve 

as mathematical models describing various real life phenomena in mathematical 

biology, population dynamics and physiology, electrical circuits and laser 

optics, economics, life sciences and others. 

2 - Using extended cutting angle and penalty methods for 

solving semi-infinite programming problems 

Albert Ferrer, Dpt. of Applied Mathematics I, Technological 

University of Catalonia, Av. Doctor Marañon, 44-50, 08028, 

Barcelona, Catalunya, Spain, alberto.ferrer@upc.edu 

Recently a unified framework concerning to Remez-type algorithms and integral 

methods coupled with penalty and smoothing methods has been introduced 

for solving convex semi-infinite programming. The framework is theoretical 

and no computational results are reported. Nevertheless, it suggests new methods 

with interesting computational properties. We propose an specific implementation 

that use the Extended Cutting Angle Method as an auxiliary method 

of the main procedure. Computational results are reported. 

EURO 24 - Lisbon 2010 WE-20 

3 - Direct Search Filter Methods 

Aldina Correia, Mathematics, ESTGF-IPP, Edifício do Salto, 

n o 4„ blc 6, 5 o esq, 4600-281, Amarante, Portugal, 

aic@estgf.ipp.pt, João Matias, Pedro Mestre, Carlos Serôdio 

Filter methods have been widely used in several areas of Constrained Nonlinear 

Optimization. These methods treat optimization problems as bi-objective 

attempts to minimize the objective function and a continuous function that aggregates 

the constraint violation functions. But, when the involved functions 

are non smooth, Unconstrained Derivative-free Methods must be used. This 

work presents results obtained by combining Filter method with other direct 

search methods and are proposed some alternatives to aggregate the constraint 

violation functions. 

� WE-20 


1.3.33A 

Social Networks 

Stream: Knowledge Discovery and Data Mining 


Chair: Armando Mendes, Mathematics, Azores University, Rua da 

Mãe de Deus, 9501-801, Ponta Delgada, Azores, Portugal, 

amendes@uac.pt 

Chair: Matthias Funk, Mathematic, University of the Azores, Rua 

Gonçalo, 9500, Ponta Delgada, mfunk@uac.pt 

1 - Large Social Networks Visualization Using the Algorithm 

of the Spanning Tree with Maximum Number of 

Leaves 

Luís Cavique, DCeT, Universidade Aberta, Rua da Escola 

Politécnica 147, 1269-01, Lisboa, Portugal, 

lcavique@univ-ab.pt, Armando Mendes 

In the web 2.0, social networks easily reach of thousands or millions of actors. 

A clear view of a small number of vertexes is easy to obtain. However, when 

the number of vertexes and edges increases, the view becomes incomprehensible. 

In this work, we intend to find the skeleton of the social network, by 

transforming the graph into a tree with the largest possible number of leaves, 

using the spanning tree algorithm with additional constraints. 

2 - Combining Data Mining Algorithms for Web Recommendation 

A. Jorge Morais, Department of Science and Technology, 

Universidade Aberta, Universidade Aberta - Delegação do Porto, 

Rua do Amial, 752, 4200-055, Porto, Portugal, 

ajorgemorais@gmail.com 

Data mining algorithms are used for recommendation of pages that might be 

useful for the user according to past behavior (of a given user or a group of 

users). Combining several algorithms to optimize user satisfaction within a 

multi-agent environment can be done in two ways: a competitive approach, 

where each agent fights for grabbing user’s attention, or a cooperative approach, 

where all agents play for the same side. In this work, both approaches were 

tested and a comparison of both against a single algorithm approach is presented. 

3 - Integration of different Cliques of Proverbial Knowledge 

Matthias Funk, Mathematic, University of the Azores, Rua 

Gonçalo, 9500, Ponta Delgada, mfunk@uac.pt, Luís Cavique 

By using 14 distinct inquiries we were able to analyze the knowledge of a 

huge number of proverbs inside the cultural space of Azores. At Euro 2009, 

we developed a pattern matching algorithm by using a incidence matrix resulting 

from the pair wise common knowledge on the best-known proverbs. By 

randomly picked an inquiry and it was possible to identify an intrinsic correlation 

between the paremiological competence and the person’s provenance. But 

these results must be validated with more data. Therefore, we now analyze all 

14 inquiries with the same method in order to compare results. 

4 - Hierarchical Clique Analysis in Social Networks Due to 

Common Knowledge of Proverbs 

Armando Mendes, Mathematics, Azores University, Rua da Mãe 

de Deus, 9501-801, Ponta Delgada, Azores, Portugal, 

amendes@uac.pt, Matthias Funk 

293


We present the Hierarchical Clique Analysis, a new algorithm for social networks 

analysis. The algorithm is exemplified with data about the recognition 

of proverbs collected in interviews in all Azorean islands and also in three 

Azorean emigration locations in the USA. Interpreting the set of this data as 

an incidence matrix of a graph, we obtain 8 oriented and isolated sub-graphs 

which distinguish the society in a kind of different families of proverbial users. 

The Hierarchical Clique Analysis finds distinct clusters with a high inner homogeneity. 

� WE-21 


6.2.47 

Education and Sustainable Development 

Stream: Education, and Social Policy 


Chair: Hanife Akar, Department of Educational Sciences, Middle 

East Technical University, Orta Dogu Teknik Universitesi, Egitim 

Fakultesi EF 316, 06531, Ankara, Turkey, hanif@metu.edu.tr 

1 - Mind the gap: modelling learning in a professional curriculum 

Jo Smedley, Newport Business School, University of Wales, 

Newport, Allt-yr-yn Avenue, NP20 5DA, Newport, United 

Kingdom, jo.smedley@newport.ac.uk 

In an increasingly litigious society, a legal understanding is often required to 

support learning in non-related areas. Accessibility can be assured through a 

range of technological and non-technological learning approaches with information 

appropriately "packaged’ to engage with a widened range of learners. 

This presentation will reflect on lessons learned during project development 

and implementation and present a model for action learning involving a professional 

curriculum. Outcomes reflect the differences in user expectations between 

academic and professional subjects. 

2 - Unemployment Similarities among Portuguese Regions 

— a Cluster Analysis Approach 

Elisa Barros, Escola Superior de Tecnologia e Gestão, Instituto 

Politécnico de Bragança, Campus de Sta Apolónia, Apartado 

1134, 5301-857, Bragança, Bragança, Portugal, ebarros@ipb.pt, 

Alcina Nunes 

The regional distribution of the unemployed individual characteristics is of core 

importance for the development of public policies that can fight the unemployment 

phenomenon, especially in times of crises. The data mining cluster 

methodology allows finding groups of regional areas that share the same 

characteristics for the register unemployed and, therefore, helps in a better understanding 

of the problem and possible solutions. Preliminary results for the 

Portuguese regions show a clear division of the territory among four regions 

— north and south and urban and rural areas of the country — concerning individual 

characteristics such as the gender, age, education or unemployment 

duration. These results have policy consequences. 

3 - Need for educational policy-making for the sustainable 

development of children living in poverty 

Hanife Akar, Department of Educational Sciences, Middle East 

Technical University, Orta Dogu Teknik Universitesi, Egitim 

Fakultesi EF 316, 06531, Ankara, Turkey, hanif@metu.edu.tr, 

Aysegul Ozsoy 

Improving educational opportunities for children of poverty may have a positive 

impact on their lives, especially, it may lead them to an upward social 

mobility to enhance a sustainable future. This talk is based on data drawn from 

a nationwide study whose participants are parents and children from squatter 

neighborhoods. Findings rate financial issues at the top of needs, and urge 

schools to provide children with poor households better school quality facilities 

to receive equality of opportunity in education. Also, social adaptation to 

urban live emerges as a need for social policy-making. 

294 

� WE-22 



Maritime Logistics: Theory and Practice 



Chair: Heng-Soon Gan, Mathematics and Statistics, University of 

Melbourne, Australia, Department of Mathematics and Statistics, 

University of Melbourne, 3010, Parkville, VIC, Australia, 

hsg@unimelb.edu.au 

1 - Robust short-sea ship routing and scheduling 

Cristina Requejo, DMat-CIDMA, University of Aveiro, 

3810-193, Aveiro, Portugal, crequejo@ua.pt, Agostinho Agra, 

Marielle Christiansen, Rosa Maria Figueiredo, Lars Magnus 

Hvattum 

A fleet of ships must service a given set of cargoes. Several ports are closed for 

service during nights and weekends, the loading or discharging may take several 

days, and it is important to avoid ships waiting in ports during the weekend 

before finishing the service. Maritime transportation is associated with a high 

degree of uncertainty, mainly due to bad weather and unpredictable service 

times. We discuss how to design robust ship routes and schedules and present 

mathematical models. 

2 - Optimizing Schedules for Cooperative Engagements 

from a United States Navy Sea Base 

Javier Salmeron, Operations Research, Naval Postgraduate 

School, 1411 Cunninham Rd, 93943, Monterey, CA, United 

States, jsalmero@nps.edu, Jeffrey Kline, Greta S. Densham 

This work presents Global Fleet Station Mission Planner (GFSMP), an optimization 

tool to aid in planning and scheduling of humanitarian-assistance 

and other theater-security cooperation missions for the U.S. Navy. GFSMP 

helps fleet staffs to examine how one naval ship deployed for six months with 

embarked teams can best meet its mission and logistical requirements. We illustrate 

the application of GFSMP in the U.S. Second Fleet’s Trident Warrior 

2009 exercise. Solutions significantly improve total mission value achieved and 

reduce costs compared to manual planning. 

3 - Discrete time models for an Inventory Ship Routing 

Problem 

Agostinho Agra, Matemática, Universidade de Aveiro, campus 

universitário de santiago, 3810-193, Aveiro, Portugal, 

aagra@ua.pt, Marielle Christiansen, Henrik Andersson 

We consider an Inventory Ship Routing Problem that combines routing and 

inventory management at all ports of a single product. The product is produced 

and stored at production ports and transported by a heterogeneous fleet 

of ships to the consumption ports. Inventory capacities are considered on the 

production and consumption ports. We present a mathematical formulation of 

the problem where the time is discretized to easily take the varying production 

and consumption rates into account. Then we discuss different approaches to 

strengthen that formulation and report computational results. 

4 - A Multi-Product Inventory Routing Problem with Varying 

Consumption Rates 

Heng-Soon Gan, Mathematics and Statistics, University of 

Melbourne, Australia, Department of Mathematics and Statistics, 

University of Melbourne, 3010, Parkville, VIC, Australia, 

hsg@unimelb.edu.au, Henrik Andersson, Marielle Christiansen 

We consider here a maritime inventory routing problem with varying consumption 

rates minimising total shipment, inventory and purchasing costs. There are 

draft limitations on ships entering production and consumption ports. More 

than one product can be loaded onto a ship. We will present an arc-based formulation 

for this problem and report on some preliminary results, including a 

decomposition attempt.

� WE-24 


6.2.50 

Educational Timetabling 



Chair: Tiago Pais, School of Computer Science, University of 

Nottingham, School of Computer Science, Jubilee Campus, NG8 


txp@cs.nott.ac.uk 

1 - University Course Timetabling using Graph Colouring 

Heuristics 

Khodakaram Salimifard, Industrial Management Department, 

Persian Gulf University, Mahini Road, 7516913798, Bushehr, 

Iran, Islamic Republic Of, salimifard@pgu.ac.ir, Salman Babaei 

Timetabling problems have been widely investigated in the operational research 

and artificial intelligence research communities for more than four decades. 

Graph colouring and generalizations are useful tools in modelling a wide variety 

of scheduling and assignment problems. This paper concentrates on university 

course timetabling. The primary purpose of our work is to solve a real 

timetabling problem at the Persian Gulf University of Iran. The paper is focused 

on weekly scheduling for the faculty of humanities. The problem includes 

40 courses, 264 course groups, and 66 teachers that have to be assigned 

to 24 classrooms within some 25 predefined time slots. To find a solution for 

such a problem, we have applied a heuristic approach based on graph colouring. 

We have used three graph colouring constructive heuristics including (1) 

Largest Degree, (2) Largest Weighted Degree, and (3) Saturation Degree. Four 

types of hard constraint and four types of soft constraints are considered. Hard 

constraints have to be satisfied under any circumstances. Timetables with no violations 

of hard constraints are called feasible solutions. Soft constraints need 

to be satisfied as much as possible. Soft constraints are used to evaluate the 

goodness of the solutions. The results show that the timetable generated by the 

algorithm is much better than the one manually created by department staff. 

2 - University course scheduling with expanding campus 

Loo Hay Lee, Industrial and Systems Engineering, National 

University of Singapore, Singapore, iseleelh@nus.edu.sg, Hung 

Hui-Chih, Kien-Ming Ng, Ek Peng Chew 

We consider a course scheduling problem for a university that is expanding its 

campus with a new location. The two parts of campus will be connected by 

shuttle service. Our problem is to reallocate courses to new part of campus and 

to minimize traffic impact. We first analyze student enrollment data and cluster 

courses according to their correlations. Based on existing course schedule, we 

predict the student movement and build a mixed integer programming model 

for course scheduling. We solve the model and then evaluate its traffic impact. 

Both analytic and numerical results will be presented. 

3 - Using Choquet Integral to combine different heuristic 

values for the exam timetabling problem 

Tiago Pais, School of Computer Science, University of 

Nottingham, School of Computer Science, Jubilee Campus, NG8 


txp@cs.nott.ac.uk, Edmund Burke 

In this work we present a constructive heuristic approach based on the Choquet 

integral. We use this method to combine the information given by different 

basic heuristics. A fuzzy measure is used to model the importance of each 

heuristic in addition to the interaction between them. We test our approach on 

2 different testbeds and compare its performance against the individual heuristics. 

Moreover, we also compare the results against the best results reported in 

the literature. 

� WE-25 


6.2.48 




Chair: Christian Artigues, LAAS, CNRS, 7 avenue du Colonel 

Roche, 31077, Toulouse Cedex 4, artigues@laas.fr 


1 - ROADEF/EURO Challenge 2010: Final result announcement 

Christian Artigues, LAAS-CNRS, 7 avenue du Colonel ROCHE, 

31077, Toulouse Cedex 4, challenge@roadef.org, Eric Bourreau, 

H. Murat Afsar, Ender Ozcan, Guillaume Dereu 

We present the results of the ROADEF/EURO Challenge 2010, an international 

optimization contest proposed jointly by EURO, the French OR society 

(ROADEF) and an industrial partner (EDF). Many prizes are offered. EDF 

proposes a 4000 euros for the junior category, 4000 euros for the senior category 

and 2000 euros for the multi-thread category. Intermediate qualification 

results (available since February 2010 on http://challenge.roadef.org/) have already 

shown that the competition is very tight, but after this presentation, the 

suspense will be over as the winners will be revealed. 

� WE-26 


3.1.11 

Game Theory and Social Choice 

Stream: Cooperative Game Theory [c] 


Chair: Dorota Marciniak, Polish Academy of Sciences, Warsaw, 

Poland, Dorofia@gmail.com 

1 - Cooperation in a group pursuit game 

Yaroslavna Pankratova, Mathematics of economic research„ 

International Banking Institute, Nevsky pr. 60, Saint-Petersburg, 

Russian Federation, yasyap@gmail.com 

A nonzero-sum group pursuit game with one pursuer and m evaders is considered. 

Choosing control variables at current time moment each player knows 

the moment t and his own and all other players’ positions. The evaders are discriminated 

what means that the pursuer knows their velocity vectors at the same 

moment t. With every nonzero-sum pursuit game we associate a corresponding 

cooperative game. We prove that in this game there exists the nonempty 

core. There is an interconnection between existence conditions of the Nash 

equilibrium of the pursuit game and nonemptiness of the core. 

2 - Probabilistic power indices for games with abstention 

Josep Freixas, Applied Mathematics 3, Technical University of 

Catalonia, Av. Bases de Manresa, 61-73, E-08242 MANRESA. 

Spain, 08242, Manresa, Spain, josep.freixas@upc.edu, Daniel 

Palacios 

In this paper we introduce several power indices that admit a probabilistic interpretation 

for games with abstention or with three levels of approval in the input 

level. We analyze the analogies and discrepancies between standard known indices 

for simple games and these extensions for this more general context. We 

conclude by proposing procedures to easily compute them. 

3 - A power analysis for voting games with consensus 

Dorota Marciniak, Polish Academy of Sciences, Warsaw, 

Poland, Dorofia@gmail.com, Josep Freixas 

In this presentation we introduce and examine the egalitarian property for the 

most established power indices on the class of simple games. We prove that 

Shapley—Shubik index, Banzhaf and Johnston scores also satisfy this property. 

We also give counterexamples for Holler, Deegan—Packel, normalised 

Banzhaf and normalised Johnston indices. We show that egalitarian property is 

a stronger condition for effcient power indices than the Lorentz domination. 

� WE-28 



Scheduling under Resource Constraints 



Chair: Can Akkan, Faculty of Management, Sabanci University, 

Orhanli, Tuzla, 34956, Istanbul, Turkey, canakkan@sabanciuniv.edu 

295


1 - New models for the Multi-Skill Project Scheduling Problem 

Carlos Montoya, Département Automatique-Productique, Ecole 

des Mines de Nantes, La chantrerie,4 rue Alfred Kastler, 44307, 

Nantes, Pays de la Loire, France, cmonto10@emn.fr, Odile 

Bellenguez-Morineau, David Rivreau 

This work introduces different approaches to solve the Multi-Skill Project 

Scheduling Problem (MSPSP).The aim is to find a schedule that minimizes 

makespan, considering precedence relationship and resources constraints.In 

this problem, the performance of each activity requires a given number of workers,with 

predefined skills.Practical applications of this problem includes buildings 

construction,production and software development planning.This work is 

based on the implementation of different MIP models. Also, we discuss their 

efficiency and limits. 

2 - Minimizing Ripple Effect in Single-Machine Rescheduling 

due to New Operation Insertion – A Branch-and- 

Bound Approach 

Can Akkan, Faculty of Management, Sabanci University, 


canakkan@sabanciuniv.edu 

We assume there are n-1 operations already scheduled (with a release-time, 

a “hard” and a “soft” due-time). The pre-schedule is feasible w.r.t the “soft” 

due-times. There is a new operation with a given release-time and a “hard” 

due-time. The objective is to insert the new operation into the schedule so that 

total tardiness of the operations in the pre-schedule with respect to the “soft” 

due-times is minimized, such that they do not violate their release-time and 

“hard” due-time costraints. A branch-and-bound algorithm is designed that 

makes use of interval algebra. 

� WE-29 


8.2.11 




Chair: Michael Zabarankin, Mathematical Sciences, Stevens Institute 

of Technology, Castle Point on Hudson, 07030, Hoboken, NJ, United 

States, mzabaran@stevens.edu 

1 - Does Model Framework Matter? 

Andrea Roncoroni, Finance, ESSEC, Avenue Bernard Hirsch, 

BP 105, 95021, Cergy-Pontoise, France, roncoroni@essec.fr 

Models are unable to appropriately represent the features owned by the underlying 

variables. We unveil a more general and ubiquitous source of risk stemming 

from the discrepancy in performance among alternative frameworks in which 

a given model can be cast. We consider a market with both spot and futures 

prices quoted and show that any reasonable pricing model can be equivalently 

cast by assigning dynamics to the pair or the whole term structure of futures 

prices. Framework risk can be defined as any assessment of the relative model 

performance in either framwork. 

2 - Risk-return portfolio optimization with disutility function 

Cristinca Fulga, Department of Mathematics, Academy of 

Economic Studies, Piata Romana 6, sector 1, 010374, Bucharest, 

Romania, fulga@csie.ase.ro 

We propose a multi-objective model for portfolio selection in which the risk is 

taken into account by considering first, an utility function which captures the 

attitude towards risk of the decision maker and second, the Conditional Value at 

Risk which is minimized such that the risk of high losses is reduced. Practical 

issues, such as transaction costs, are incorporated in the decision model. Computational 

results based on real data drawn from the Bucharest Stock Exchange 

are given. 

296 

� WE-30 


8.2.13 

Risk measurement and control in volatile 

financial markets 


in Banking 


Chair: Giorgio Consigli, Mathematics, Statistics and Computer 

Sciences, University of Bergamo, Via dei Caniana 2, 24127, 

Bergamo, Italy, giorgio.consigli@unibg.it 

1 - Long term financial scenarios, sampling and optimal 

strategic planning 

Giorgio Consigli, Mathematics, Statistics and Computer 

Sciences, University of Bergamo, Via dei Caniana 2, 24127, 

Bergamo, Italy, giorgio.consigli@unibg.it 

We present a statistical model for long term financial planning under assumptions 

of random market risk premia and inflation risk adjustment to yield an 

optimal dynamic control strategy over a 10 year planning horizon. The growing 

adoption of decision tools, combining the Monte Carlo method for risk 

assessment purposes, sampling techniques for the derivation of a relevant set of 

economic scenarios and linear or nonlinear objective function for the solution 

of complex decision problems shows the practical relevance of stochastic programming 

approaches in financial management problems. We discuss the key 

elements of this methodological framework within alternative formulations of 

real world, practical ALM problems. 

2 - A sequential learning method for tracking stochastic 

volatility 

Enza Messina, DISCo - Department of Informatics, Systems and 

Communication, Università degli Studi di Milano Bicocca, viale 

Sarca 336, 20126, Milano, Italy, messina@disco.unimib.it, 

Giorgio Consigli, Cristina Manfredotti 

In this talk we present a sequential learning approach for estimating risk indicator 

parameters within a stochastic volatility model. The approach, based on 

a particle filtering method, is aimed at tracking and forecasting the stochastic 

volatility through the Bayesian estimation of the parameters of a stochastic process 

with jumps. In this way we estimate the probability of a jump occurrence 

and its size distribution on the base of the discrete observation of prices. This 

technique can be embedded in a portfolio optimization model. 

3 - Optimal DC-type pension fund management 

Francesco Sandrini, Institutional investments, Pioneer 

Investments, 16 Appianstrasse, Unterforhing, bei Munchen, 

81675, Munchen, Germany, 

francesco.sandrini@pioneerinvestments.com 

Between Defined Benefit and Defined Contribution, the need for safeguarding 

minimum returns related to inflation and individual liabilities is emerging as 

a consensus solution across several European Countries, currently involved in 

changes within their regulatory framework: is this a third way forward for European 

Pension Fund Schemes? How quantitative techniques such as Dynamic 

Stochastic Programming, which have been broadly employed in the past to 

solve problems related to Asset and Liability Management (ALM) can be now 

employed to structure solutions both in terms of advisory and product development 

for Defined Contribution Pensionplans? We exploit a scenario -based pension 

advisory framework simulating Individual Liabilities by using Montecarlo 

techniques and perform strategic allocations coherent with the achievement of 

minimum level substitution-rates for the plan members. We solve an dynamic 

optimization problem where employees’ contributions are proportional to their 

salaries and the final target functions relates to variables traditionally the domain 

of defined contribution pension funds such as substitution rated (ratio 

between annuities and last salaries) .This framework aims to become over time 

a new planning and advisory benchmark for the new generation of hybrid DC 

plans. 

4 - Credit Risk Management via Stochastic Programming 

Patrizia Beraldi, Department of Electronics, Informatics and 

Systems, University of Calabria, Via P. Bucci - CUBO 41/C, 

87036, Rende (CS), ITALY, Italy, beraldi@deis.unical.it, 

Antonio Violi, Giorgio Consigli

In this work we propose an integrated approach to manage bond portfolios. 

With respect to other ALM problems the main difference is the particular nature 

of financial instruments considered. Bonds are subject to the price volatility 

caused by market dynamics but also to the default probability of counterparts. 

Moreover, the presence of (various types of) coupons introduces a more 

complex cash flow management, since future investment decisions are strongly 

linked with coupon incomes. The proposed approach integrates two correlated 

risk sources, market and credit risk, under very general statistical assumptions, 

by means of a special-purpose scenario generator and a multistage stochastic 

programming model. Two key issues (consistent risk factors statistical characterization 

and effective dynamic optimization) have been jointly implemented 

and tested during the crisis period, in order to provide an integrated decision 

tool with practical relevance. 

� WE-33 


8.2.19 

Realistic Production Scheduling I 

Stream: Realistic Production Scheduling 


Chair: Ruben Ruiz, Departamento de Estadistica e Investigación 

Operativa Aplicadas y Calidad, Universidad Politecnica de Valencia, 

Camino de Vera S/N, 46021, Valencia, Spain, rruiz@eio.upv.es 

1 - Estimation of distribution algorithm flow shop scheduling 

with lot-streaming and setup times 

Ruben Ruiz, Departamento de Estadistica e Investigación 

Operativa Aplicadas y Calidad, Universidad Politecnica de 

Valencia, Camino de Vera S/N, 46021, Valencia, Spain, 

rruiz@eio.upv.es, Quan-Ke Pan 

We consider a n-job m-machine lot-streaming flow shop scheduling problem 

with sequence-dependent setup times under both the idling and no-idling production 

cases, and the objective is to minimize the makespan. A novel estimation 

of distribution algorithm (EDA) is proposed with a job permutation 

based representation. A simple but effective local search is fused to enhance 

the intensification capability. A speed-up method is presented to reduce the 

computational effort needed for the local search technique and the NEH-based 

heuristics. 

2 - MILP models and solution approaches for scheduling a 

chemical batch process 

Laura Hege, Laboratoire Génie Industriel, Ecole Centrale Paris, 

92290, Chatenay-Malabry, France, laura.hege@gmail.com, 

Céline Gicquel, Michel Minoux 

We study a production scheduling problem encountered in a chemical plant 

producing enzymes. The chemical process is a sequential, 4-stage process with 

parallel production units at each stage. We propose a first MILP formulation 

based on a discrete time representation and derive a family of valid inequalities 

to strengthen it. We then investigate a second MILP formulation based on 

a continuous time representation and use it to devise a neighbourhood-search 

heuristic algorithm. Finally, we provide computational results showing the efficiency 

of both approaches. 

3 - A Production Scheduling Implementation Using Critical 

Chain Method 

Zeynep Gergin, INDUSTRIAL ENGINEERING 

DEPARTMENT, ISTANBUL UNIVERSITY, Avcilar Kampusu, 

Avcilar, 34320, ISTANBUL, Turkey, zgergin@goldenminds.com 

In this study, using Critical Chain Project Management (CCPM) Method, a 

production scheduling is implemented in a print house. Firstly single, secondly 

concurrent orders are scheduled like a project using Critical Path Method 

(CPM). In order to apply CCPM, the constraint resource (drum) is determined, 

all resources are leveled, buffers are calculated manually and the expected delivery 

time is forecasted. Then, results are confirmed with CC Pulse software. 

The results support the related literature regarding the consistency of scheduling 

projects with CCPM instead of CPM. 


� WE-35 


6.2.46 

Advances in Mixed-Integer Linear and 

Nonlinear Programming 



Chair: Leo Liberti, LIX, Ecole Polytechnique, LIX, Ecole 

Polytechnique, 91128, Palaiseau, France, leoliberti@gmail.com 

Chair: Andrea Lodi, D.E.I.S., University of Bologna, Viale 

Risorgimento 2, 40136, Bologna, Italy, andrea.lodi@unibo.it 

1 - Sampling issues for intensity modulated radiation therapy 

(IMRT) treatment planning 

Humberto Rocha, Inesc - Coimbra, Portugal, hrocha@mat.uc.pt, 

Joana Matos Dias, Brígida da Costa Ferreira, Maria do Carmo 

Lopes 

Operations research has made significant contributions to the improvement of 

IMRT optimization. Many mathematical optimization models have been proposed 

for the intensity problem including mixed integer linear models. Regardless 

the formulation used, problem size is always the biggest challenge 

to overcome. The most common strategy to address this problem is sampling 

which enhances gains in computational time at a cost: the quality of the solution 

deteriorates. A clinical example of a head and neck cancer case is used to 

discuss both the need and the influence of sampling in IMRT optimization. 

2 - Binet graphs and rank 1 closure 

Gautam Appa, Operational Research, London School of 

Economics, Houghton Street, WC2A 2AE, London, United 

Kingdom, g.appa@lse.ac.uk, Konstantinos Papalamprou, 

Leonidas Pitsoulis 

Network matrices arising out of the incidence matrices of directed graphs are 

the building blocks of totally unimodular matrices. We try to generalise these 

by starting with incidence matrices of mixed graphs, ie, graphs with directed 

and undirected edges. They give rise to bi-directed networks or binet graphs 

and lead to binet matrices B. These are shown to have rank 1 closure, ie, rank 

one cuts give the convex hull of integer points in the polyhedron Bx=0. 

Theory, applications and generalisations of binet graphs and binet matrices are 

presented in this talk. 

3 - De-convexification tightens CHR bounds on convex 0-1 

quadratic programming problems. 

Monique Guignard-Spielberg, OPIM, University of 

Pennsylvania, 5th floor, JMHH, 3730 Walnut Street, 191046340, 

Philadelphia, PA, United States, guignard_monique@yahoo.fr, 

Aykut Ahlatcioglu, Michael Bussieck, Mustafa Esen, Alex 

Meeraus 

Bounds on the optimal value of a convex 0-1 quadratic programming problem 

with linear constraints can be improved by a preprocessing step that adds to 

the quadratic objective function terms which are equal to 0 for all 0-1 feasible 

solutions yet increase its continuous minimum. Using Plateau’s QCR method 

(2005), or one of its predecessors, the methods of Hammer and Rubin (1970) 

or that of Billionnet and Elloumi (2008), strengthens the CHR as well as the 

continuous bounds. We present results for convex GQAP problems. 

4 - Request Routing and Client Assignment in content Distribution 

Network 

Chris Potts, School of Mathematics, University of Southampton, 

Highfiled, SO17 1BJ, Southampton, Hampshire, United 

Kingdom, C.N.Potts@soton.ac.uk, Narges Haghi, Tolga Bektas 

A Content Distribution Network (CDN) is a system of servers containing selected 

objects in the form of data that are placed at selected nodes of a network. 

We propose an integer programming model to solve the joint problem of request 

routing and client assignment which explicitly considers delays in transmitting 

objects. The resulting model is a nonlinear mixed integer programming 

formulation. We propose a method that is based on Lagrangian relaxation, decomposition, 

and subgradient optimization to quickly obtain lower bounds for 

the problem. 

297


� WE-36 


3.1.05 

Linear and Conic Programming II 

Stream: Linear and Conic Programming 


Chair: Mohammad Reza Peyghami, Mathematics, K.N. Toosi 

University of Technology, Math. Department, K.N. Toosi University 

of Technology„ P.O.Box 16315-1618, Tehran, Iran, 16315-1618, 

Tehran, Tehran, Iran, Islamic Republic Of, peyghami@kntu.ac.ir 




1 - An algorithm for the multiparametric 0-1-integer linear 

programming problem with interval data 

Alejandro Crema, Escuela de Computación, Facultad de 

Ciencias, Universidad Central de Venezuela, Apartado 47002, 

1041-A, Caracas, Venezuela, Caracas, Venezuela, 

alejandro.crema@ciens.ucv.ve, Edgar Peraza 

The multiparametric 0-1-Integer Linear Programming (0-1-ILP) problem relative 

to the objective function is a family of 0-1-ILP problems in which the 

problems are related by having identical constraint matrix and right-hand-side 

vector. We present an algorithm to perform multiparametric analysis in the case 

of an objective function with interval data. 

2 - A new complexity result on multiobjective linear integer 

programming 

Victor Blanco, Departamento de Algebra, Universidad de 

Granada, Facultad de Ciencias, Campus Fuentenueva, 18071, 

Granada, Granada, Spain, vblanco@ugr.es, Justo Puerto 

We present algorithms for solving multiobjective integer programming problems. 

The algorithms uses the Barvinok’s rational functions of the polytope 

that defines the feasible region and provides the entire set of nondominated solutions 

for the problem. Theoretical complexity results on the algorithm are 

presented. We prove that encoding the entire set of solutions of the problem in 

a generating function is polynomially doable, when the dimension is fixed. We 

also provide polynomial delay algorithms for enumerating this set. 

3 - Advertisement Reservation Problem on a Relationship- 

Based Environment 

Ilkay Gultas, Industrial Engineering, Istanbul Kultur University, 

Atakoy-Istanbul-Turkiye, 34156, Istanbul, Turkey, 

i.gultas@iku.edu.tr, Nihan Karaca 

TV networks provide TV programs free of charge but they acquire revenue by 

telecasting advertisements during programs. A key problem is to determine 

which reserved advertisements will be accepted, it is complicated by limited 

time inventory, different GRPs for different demographics, competition avoidance 

and the relationship between TV networks and clients. A mixed integer 

linear programming approach is proposed, and an implementation is made to a 

case from one of the biggest TV networks of Turkey. 

4 - A MIP for potential conflicts minimization by speed regulations 

David Rey, INRETS, Bron, France, david.rey@inrets.fr, 

Christophe Rapine, Rémy Fondacci, Sophie Constans 

We address the speed regulation problem in air traffic management. Increasing 

the airspace capacity has become a priority in order to deal with future air traffic 

demand. We investigate automatic speed modulations for en-route flights. 

Focusing on aircraft crossing times at intersection points to ensure separation, 

speed regulation is turned into travel time control. We propose a mixed integer 

linear model to minimize potential conflicts, i.e. when two or more aircraft are 

below the separation norm. The promising results pave the way for reducing 

the air traffic controllers’ workload. 

298 

� WE-37 


3.1.09 

Robustness concerns and multiple criteria 

decision aid 






1 - On Designing a Robust Electricity Distribution Network 

André Rossi, Lab-STICC - UMR 3192, Université de 


France, andre.rossi@univ-ubs.fr, Alexis Aubry, Mireille 

Jacomino 

Power distribution networks are seriously challenged as the gap between production 

and demand is rapidly decreasing over years. The consequences of 

demand locally exceeding electricity supply are very likely to have a global 

impact as most of the networks are now interconnected together. To prevent 

a blackout to happen, distribution networks must be endowed with robustness 

features so as to ensure a balanced power distribution despite demand variations. 

This is achieved by maximizing the minimum supply margin over power 

suppliers. A cutting plane algorithm is proposed to address this problem. 

2 - A robustness indicator in blending problems 

Jorge Aguilera, LGI2P, EMA, Site EERIE, Parc Scientifique G. 

Besse, 30035, Nimes, France, jorge.a.aguilera.c@gmail.com, 

Stefan Janaqi, Mériam Chébre 

In this paper we explore the Blending Problem feasibility with uncertain data 

from a geometric point of view. The BP feasibility can be naturally defined as 

a non empty intersection F of two polytopes: the possible blend’s polytope and 

the target polytope. We define two robustness indicators on the feasibility of a 

blend based on its depth in F and on an approximation of F’s width. Finally, we 

explain how to produce a blend with a given depth in F. From a practitioner’s 

point of view, these measures have to be computed efficiently, so linear models 

are used. 

3 - Robust Ordinal Regression of Value Functions Handling 

Interacting Criteria 

Salvatore Greco, Deapartment of Economics and Quantitative 

Methods, University of Catania, Corso Italia 55, 95129, Catania, 

Italy, salgreco@unict.it, Vincent Mousseau, Roman Slowinski 

Multiple attribute additive utility does not permit to represent interactions between 

the considered criteria and for this non-additive integrals, such as Choquet 

integral and Sugeno integral, have been proposed. The non-additive integrals, 

however, need that the evaluations on all criteria are expressed on the 

same scale. In the context of the Robust Ordinal Regression, we propose an 

aggregation model, called UTAGMS—INT, which modifies the usual additive 

value model so as to handle interactions between criteria without the necessity 

of expressing all the evaluations on the same scale. 

4 - On the multi-agent job shop scheduling problem 

Cyril Briand, LAAS - CNRS, 7, Av. Colonel ROCHE, 31077, 

Toulouse Cedex 4, France, briand@laas.fr, Thomas Lehaux 

We focus on job shop scheduling problems wherein each machine is associated 

with an agent having its own decisional autonomy. An agent only knows the set 

of tasks it has to deal with, as well as the set of completion time intervals associated 

with the preceding tasks. The objective is to determine a global schedule 

satisfying both the agent’s objectives and some global (social) objectives collectively 

shared by the agents. The problem is modeled in such a way that the 

agent’s self organization is relatively robust to the uncertainties arising from 

the other agents.

� WE-38 


6.2.44 

Issues in Inventory Management 

Applications II 

Stream: Inventory Management [c] 


Chair: Kelvin T Chirenje, Applied Maths, National University Of 

Science and Technology, 10 Rukumbati Rd, Zengeza 3, +263, 

Chitungwiza, Harare, Zimbabwe, kchirenje@gmail.com 

1 - A Study of the Inventory Models for Perishable Products 

in Aerospace Industry 

Ying-Chyi Chou, Department of Business Administration, 

Tunghai University, 181 Taichung-kang Rd., Sec. 3, Taichung, 

Taiwan, R.O.C., 407, Taichung„ Taiwan, 

rosechyi@yahoo.com.tw, Hsiao-Wen Chu 

The research takes Aerospace Industrial Development Corporation (AIDC) as 

an example to calculate and verify the models, and look for the minimum expected 

average cost per unit for each situation. Our study is based on industry’s 

characteristics and products’ attributes to develop four models. The objectives 

of these models are to get optimal order quantity and reorder point by calculating 

the minimum expected average cost per unit time. We provide a suggestion 

for confronting these situations for managers not only in aerospace industry but 

also any company related to perishable items. 

2 - Improving the Service Level in Grocery Stores: Are You 

the Right Candidate for Item-Level RFID Tagging 

Esma Nur Cinicioglu, School of Business, Quantitative Methods 

Department, Istanbul University, Istanbul Universitesi, Isletme 

Fakultesi, Sayisal Yontemler Anabilim Dali, 34320, Istanbul, 

Turkey, esmanurc@istanbul.edu.tr 

In this research to evaluate the suitability of RFID implementation in a grocery 

store we developed a factor called beta which is defined as the rate of the unresponded 

demand compared to the error source.Three different scenarios of a 

grocery store model is simulated which illustrate the stepwise implementation 

of the RFID technology.The results indicate that a grocery store with beta values 

close to 1 is an ideal candidate for RFID implementation.With beta values 

going down to 0 the service level improvement that is achieved is decreasing. 

� WE-39 


6.2.45 

Scheduling and lot sizing under 

uncertainties II 



Chair: Mikhail Kovalyov, Belarus Academy of Sciences, Belarus, 

koval@newman.bas-net.by 

Chair: Alexandre Dolgui, IE & Computer Science, Ecole des Mines 

de Saint Etienne, 158, cours Fauriel, 42023, Saint Etienne, France, 

dolgui@emse.fr 

1 - Dynamic lot sizing problem with stochastic production 

rates and demands 

El-Houssaine Aghezzaf, Industrial Management, Ghent 

University, Technologiepark 903 -, Campus Ardoyen, 9052, 

Zwijnaarde, Belgium, ElHoussaine.Aghezzaf@UGent.be 

We consider the dynamic lot-sizing problem in which production rates as well 

as demands are subject to uncertainty. We assume that these production rates 

can be influenced through preventively acting on the production system. We 

investigate equivalent deterministic models as well as appropriate solution approaches 

for the problem when multi-items are produced on a production system 

having a specific configuration. 


2 - Two-supplier inventory systems with resource constraints 

Ekaterina Bulinskaya, Dept. Mathematics and Mechanics, 

Moscow State University, Leninskie gory 1, 119991, Moscow, 

Russian Federation, ebulinsk@mech.math.msu.su, Larisa 

Afanasyeva, Elena Yarovaya 

We consider a discrete-time inventory model with periodic ordering. Demand 

is described by a sequence of iid rv’s with a known (or unknown) distribution. 

There are two suppliers. The first one delivers an order immediately, the second 

one has a lag. The aim is to minimize n-period expected costs. In contrast with 

previous papers there are orders size constraints. Optimal policy is proved to 

be determined by critical levels satisfying Bellman equations. Model stability 

is established. The research is supported by RFBR grant 10-01-00266. 

3 - Selecting Capacity at the Outset and Price after Uncertainty 

has been Realized 

Yigal Gerchak, Dept of Industrial Engineering, Tel-Aviv 

University, 69978, Tel-Aviv, Israel, ygerchak@eng.tau.ac.il 

The standard lot sizing/pricing model has both selected at the outset (before 

demand curve has become known). We delay the pricing decision to after demand 

curve becomes known, which is relevant to industries where pricing need 

no be committed early. We deal with both the Linear-Additive and iso-elasticmultiplicative 

model. Coordination issues are also discussed. 

� WE-40 


6.2.52 

Recent Advances in Industrial and 

Engineering Optimization II 






1 - Stream Mapping Value (MVS) in a Manufacturer of Gearboxes: 

Procedure Facilitator to Deployment of Lean 

Production System. 

Nivaldo Coppini, Industrial Engineering Post Graduation 

Program, UNINOVE-Nove de Julho University, Av. Francisco 


ncoppini@uninove.br, Luiz Bekesas, Milton Vieira Júnior 

A powerful tool for production optimization is the Mapping Value Stream. It 

makes it possible to identify steps covered by a group of products. This allows 

planning and changes introducing benefits in productivity and costs: displaying 

cases for product, identifying waste, directing the information to introduce 

techniques of lean production, identifying the path to MVS prior to practical 

application on the basis of performance indicators. We deal with the steps of 

MVS for a gearboxes manufacturer to make profit or to go towards deciding 

about a Lean Production System. 

2 - Bi-Objective Assembly Line Balancing with Flexible Machines 

S.m.t. Fatemi Ghomi, Industrial Engineering, Amirkabir 

University of Technology, Hafez avenue, Tehran, Tehran, Iran, 

Islamic Republic Of, fatemi@aut.ac.ir, U. Bahalke, M. 

Karimi-Nasab 

In most real assembly lines, minimizing the cycle time and the total costs are 

important issues. But every manager should pay attention to total costs. A new 

assembly line balancing problem is considered based on 11 assumptions. The 

problem is formulated by a mixed integer model. For special characteristics, a 

new simple heuristic method is proposed for obtaining the most of the global 

Pareto-optimal solutions. At each iteration a floating objective bound is added. 

Computational experiences illustrate the efficiency and efficacy of the method 

in comparison with previous solution approaches. 

299


� WE-41 


3.1.06 

Financial Mathematics and Simulation 

Stream: Simulation Methods in Finance 





Chair: Henry Schellhorn, mathematics, Claremont Graduate 

University, 710 N. College Ave, 91711, Claremont, CA, United 

States, Henry.Schellhorn@cgu.edu 

1 - Bayesian and Bootstrapped Maximum Likelihood Estimation 

Approaches for ARCH and AR-ARCH Models of 

Brazilian Financial Time Series: A Performance Comparison 

Sandra Cristina Oliveira, Business Administration, UNESP - 

Univ. Estadual Paulista, Av. Domingos da Costa Lopes, 780, 

17602-496, Tupã, São Paulo, Brazil, sandra@tupa.unesp.br, 

Marinho Gomes Andrade 

This work is concerned with the parameter estimation for autoregressive conditional 

heteroscedasticity models as well as for autoregressive models with 

ARCH errors. The estimates are computed by using both Maximum Likelihood 

and Bayesian approaches. For the ML case reliable confidence intervals 

are computed via the Bootstrap Resampling Method, while for the Bayesian 

case we introduce a novel reparametrization strategy to the parameters. The 

proposed methodology is evaluated by considering two Brazilian financial time 

series. 

2 - Sensitivity Analysis in the Binomial Lattice Approach 

for Real Options 

Babak Jafarizadeh, Dept of Petroleum Engineeirng, University of 

Stavanger, Norway, babak.jafarizadeh@uis.no 

Binomial lattice methods are an intuitive way of modeling real options and can 

provide great decision insights. In this approach, the underlying Marketed Asset 

Disclaimer assumption provides an easy way of consolidating all sources 

of uncertainty into a single stochastic process used for evaluating the options. 

It is hard to trace back the effect of each source of uncertainty on option value 

once the optimal decisions are identified. We explore ways to run sensitivity 

analysis within this evaluation framework using off-the-shelf spreadsheet and 

simulation programs. 

3 - Studying Value at Risk (VaR) Models in forecasting the 

Emerging market indices volatilities 

Milad Nozari, Graduate school of management and economics 

(GSME), Sharif University of Technology, Tehran, Iran, Islamic 

Republic Of, milad_nozari@yahoo.com, Pedram Jahangiry, 

Sepideh Mohamad Raee Naeeni, Mohsen Bahramgiri 

We aim to compare Value at Risk models on emerging market indices using 

parametric and semi-parametric approaches and figure out the one which predict 

market risk in its best way. Among many parametric models, GARCH was 

used and enriched through applying Extreme Value Theory. Filtered Historical 

Simulation was selected among semi-parametric models. Complicated calculations 

and negligible nuances of our results wipe out the preference of using 

parametric models to semi-parametric ones. 

� WE-42 


3.1.07 

Graph Guarding 



Chair: Anthony Bonato, Mathematics, Ryerson University, 350 

Victoria St, M5B2K3, Toronto, ON, Canada, abonato@ryerson.ca 


1 - Meyniel’s Conjecture 

Anthony Bonato, Mathematics, Ryerson Univesity, 250 Victoria 

St, M6R1T5, Toronto, Ontario, Canada, abonato@ryerson.ca 

The game of Cops and Robbers and the cop number of a graph are topics 

of increasing interest in graph theory and computer science. The minimum 

number of cops needed to capture the robber is the cop number of a graph 

G, and is written c(G). Relatively few bounds are known on the cop number. 

Meyniel’s conjecture—which is possibly the deepest problem surrounding the 

cop number—states that for a connected graph G with n vertices, c(G) = sqrt(n). 

We give an introduction to Cops and Robbers, the cop number, and describe 

some recent approaches to settling the conjecture. 

2 - Guarding grids and related graph problems 

Pawel Zylinski, Institute of Informatics, University of Gdansk, 

Wita Stwosza 57, 80-952, Gdansk, Poland, 

zylinski@inf.ug.edu.pl 

The problem of guarding grids was formulated by Ntafos in 1986. A grid P is 

a connected union of vertical and horizontal line segments, and a point x in P 

can see a point y in P if the line segment xy is a subset of P. A set of points S, 

being a subset of P, is a guard set for grid P if any point of P is seen by at least 

one guard in S. We shall present several variants of the problem, including cooperative 

guards, fault-tolerant guards, mobile guards, and the pursuit evasion 

problem, and discuss their relation to the well-known graph theory problems, 

i.e., matching, coloring, domination. 

3 - Graphs with average degree smaller than 30/11 are 

burning slowly 

Pawel Pralat, Department of Mathematics, West Virginia 

University, 26505, Morgantown, WV, United States, 

pralat@math.wvu.edu 

Suppose that a fire breaks out at a given vertex v of G. In each subsequent time 

unit, a firefighter protects one vertex which is not yet on fire, and then the fire 

spreads to all unprotected neighbours of the vertices on fire. The objective of 

the firefighter is to save as many vertices as possible. 

The surviving rate rho(G) of G is defined as the expected percentage of vertices 

that can be saved when a fire breaks out at a random vertex of G. Let eps 

>0. We show that graphs with average degree smaller than 30/11 are burning 

slowly (the constant 30/11 cannot be improved). 

4 - Modeling Firefighting with Graphs 

Stephen Finbow, St Francis Xavier University, Canada, 

sfinbow@stfx.ca 

Let G be a connected graph with at least two vertices. Suppose that a fire breaks 

out at a vertex v of G and a firefighter then protects a vertex not yet on fire. Afterwards, 

the fire spreads to all its unprotected neighbours in each time interval. 

The fire and firefighter take turns until the fire can no longer spread. This is the 

basic model of the spread of a fire on a graph. 

We will discuss a variety of results, including computational complexity, optimal 

strategies on grid graphs and the surviving rate for various classes of 

graphs. 

� WE-43 


8.2.02 

Research Aspects Related to Life - Risk and 

Insurance 

Stream: Life Insurance, Risk Management & OR 


Chair: Irini Dimitriyadis, Mathematics and Computer Sciences, 

Bahcesehir University, Bahcesehir Univ. Dept of Mathematics, and 

Computer Sciences, 34353, Istanbul, Besiktas, 

dimitri@bahcesehir.edu.tr 




1 - Interactive Selection on Multidimensional Interval Data 

Kurt Nielsen, Institute of Food and Resource Economics, 

University of Copenhagen, Rolighedsvej 25 building C, 1958,

Frederiksberg C., Denmark, kun@life.ku.dk, Jens Leth 

Hougaard 

We present an outranking procedure that supports selection of alternatives represented 

by multiple attributes with interval valued data. The procedure is interactive 

in the sense that the decision maker direct the search for preferred 

alternatives by providing weights of the different attributes as well as parameters 

related to risk attitude and weighted dominance. The outranking relation 

builds on comparisons between optimistic and pessimistic weighted values as 

well as weighted dominance relations supported by volume based measures. 

2 - Cumulative Prospect Theory for the interval of reference 

points 

Renata Dudzinska-Baryla, Operations Research, University of 

Economics in Katowice, ul. 1 Maja 50, 40-287, Katowice, 

Poland, dudzinska@ae.katowice.pl, Donata Kopanska-Brodka 

The behavior of individuals choosing among risky alternatives is described by 

the Cumulative Prospect Theory. The basic assumption of this theory is that 

decision maker understands outcomes as gains and losses relative to the reference 

point. That is why the valuations of risky alternatives most of all depend 

on this point. In our approach we state that the individuals attitude towards risk 

in the sense of the EUT is the essential cause which outcomes are considered 

as gains or losses. We analyze some properties of the valuation of prospects 

when the interval of reference points changes. 

3 - Knowledge discovery vs. domain knowledge: the case 

of wrong signs 

Reza Shahi, School of Management, University of Southampton, 

Highfield, SO17 1BJ, Southampton, United Kingdom, 

mrams@soton.ac.uk 

Robustness and comprehensibility are two sides of the same coin in model development. 

Being able to incorporate domain knowledge is one of success 

elements for any data mining application. The conformity of the sign of model 

coefficients with the domain knowledge is one aspect of this incorporation; applying 

data mining techniques without investigating the rational between variables 

might be misleading even if the technique is robust. This paper studies 

wrong sign problem in corporate default prediction models and tries to find 

reasons and solutions. 

� WE-44 


8.2.03 

Systems and Mathematical Modelling in OR 

III 














1 - Using Absorbing Markov Chain to Evaluate the Slot 

Customer Experiences 

Xiaoming Liu, FBA, University of Macau, University of Macau, 

Macau, NA, MACAU, China, xmliu@umac.mo, Zhaotong Lian 

This study derives the conditional sojourn times of a special Absorbing Markov 

Chain which models the dynamic of slot machines, and uses the theoretical 

results to obtain the formula for important slot customer experiences: hit frequency, 

the customer losing probability, the mean and standard deviation of 

time-on-device per customer, time-on-device per losing customer etc. The results 

allow slot managers and designers to evaluate the effects of payout table 

to customer experiences and then help to design the slot machine that gives the 

best trade-off of customer experience and casino gain. 

2 - The analysis of capital stock dynamics by the optimal 

transformation 

Jan Kodera, Center for Basic Research in Dynamic Economics 

and Econometrics, University of Economics Prague, nam. 

Winstona Churchilla 4, 130 67 , Praha 3, Czech Republic, 

kodera@vse.cz, Quang Van Tran, Miloslav Vosvrda 

Firstly, we introduce a modified Kalecki model with a time delayed investment 

function which takes into account the diminishing marginal efficiency of capital 

and its installation costs. The model is a difference-differential equation. 

Secondly, we analyze the data generated by the model using the method of optimal 

transformation in order to reconstruct the original structure. This approach 

helps us to determine the effectiveness of this method and we find that the optimal 

transformation seems to be a suitable tool. Then we analyze the empirical 

data from the time series of capital stock. 

� WE-46 


8.2.14 

AIMMS User Meeting 

Stream: AIMMS User Meeting [c] 


Chair: Frans de Rooij, AIMMS, Paragon Decision Technology B.V., 



1 - AIMMS User Meeting 

Frans de Rooij, AIMMS, Paragon Decision Technology B.V., 


frans.de.rooij@aimms.com, Ovidiu Listes, Gertjan de Lange 

We invite all current and prospective users of the AIMMS mathematical modeling 

system to participate in this AIMMS User Meeting at the EURO conference. 

Please come along to hear about the newest AIMMS features and 

development plans, for example in Robust Optimization and in Server-Based 

deployment of optimization applications. Of course, there will also be time to 

meet other AIMMS Users and to share your experience and feedback. 

301

STREAMS 

Actuarial Sciences and 

Stochastic Calculus 

Invited 

Ricardo Josa-Fombellida 

Universidad de Valladolid 

ricar@eio.uva.es 

Juan Pablo Rincon-Zapatero 

Universidad Carlos III de Madrid 

jrincon@eco.uc3m.es 

Track(s): 14 

3 sessions 

Agent-Based Modeling 

Invited 

Massimo Genoese 

University of Karlsruhe 


Track(s): 46 

2 sessions 

Agent-Based Modeling [c] 

Contributed 

Massimo Genoese 

University of Karlsruhe 


Track(s): 46 

1 session 

AIMMS User Meeting [c] 

Contributed 

Frans de Rooij 

Aimms 


Track(s): 46 

1 session 

Algorithmic Decision Theory 

Invited 

Alexis Tsoukiàs 

Cnrs - Lamsade 


Ulrich Junker 

ILOG, An IBM Company 

uli.junker@free.fr 

Track(s): 43 

2 sessions 

302 

Algorithmic Decision Theory [c] 

Contributed 

Alexis Tsoukiàs 

Cnrs - Lamsade 


Ulrich Junker 

ILOG, An IBM Company 

uli.junker@free.fr 

Track(s): 43 

2 sessions 

Analytic Hierarchy Processes, 

Analytic Network Processes 

Invited 

Josef Jablonsky 

University of Economics Prague 

jablon@vse.cz 

Y. Ilker Topcu 

Istanbul Technical University 

ilker.topcu@itu.edu.tr 

Track(s): 12 

14 sessions 

Applications in Business 

Intelligence and Knowledge 

discovery 

Invited 

Richard Weber 

University of Chile 

rweber@dii.uchile.cl 

Track(s): 18 

2 sessions 

Boolean Programming 

Invited 

Endre Boros 

Rutgers University 

Endre.Boros@rutcor.rutgers.edu 

Track(s): 29 

7 sessions 

Combinatorial Optimization 

Invited 

Paolo Toth 

University of Bologna 

paolo.toth@unibo.it 

Track(s): 2 

8 sessions 

Combinatorial Optimization [c] 

Contributed 

Paolo Toth 


paolo.toth@unibo.it 

Track(s): 2 

1 session 

Computational Biology, 

Bioinformatics and Medicine 

Invited 


Politechnika Poznanska 


Metin Turkay 

Koc University 

mturkay@ku.edu.tr 

Giovanni Felici 

Cnr 

giovanni.felici@iasi.cnr.it 

Track(s): 24 

9 sessions 

Computational Statistics 

Invited 

Pakize Taylan 

Dicle University 

ptaylan@dicle.edu.tr 


Middle East Technical University 


Track(s): 17 

4 sessions 

Container Terminal Operations 

Invited 

Ceyda Oguz 

Koc University 

coguz@ku.edu.tr 

Track(s): 40 

3 sessions 

Convex Optimization 

Invited 

Attila Gilanyi 

University of Debrecen 

gilanyi@math.klte.hu 

Michel Baes 

Eth 

michel.baes@ifor.math.ethz.ch 

Track(s): 34 

6 sessions 

Cooperative Game Theory 

Invited 

Mariana Rodica Branzei 

"Alexandru Ioan Cuza” University 

branzeir@info.uaic.ro 

Sirma Zeynep Alparslan Gok 

Faculty of Arts and Sciences, Suleyman 

Demirel University 


Track(s): 26 

9 sessions

Cooperative Game Theory [c] 

Contributed 

Mariana Rodica Branzei 

"Alexandru Ioan Cuza” University 

branzeir@info.uaic.ro 

Sirma Zeynep Alparslan Gok 

Faculty of Arts and Sciences, Suleyman 

Demirel University 


Track(s): 26 

1 session 

Cutting and Packing 

Invited 

Jose Fernando Oliveira 

Universidade do Porto 

jfo@fe.up.pt 

A. Miguel Gomes 

University of Porto 

agomes@fe.up.pt 

Track(s): 20 

9 sessions 

Data Mining and Applications [c] 

Contributed 

Vadim Strijov 

Computing Center of the Russian 

Academy of Sciences 

strijov@ccas.ru 

Track(s): 42 

2 sessions 

Data Mining and Decision Making 

Invited 

Lai-Soon Lee 

UPM Serdang 

lslee@math.upm.edu.my 

Hsin-Vonn Seow 

University of Nottingham- Malaysia 

Campus 



2 sessions 

Data Mining in Early Warning 

Systems 

Invited 

Gulser Koksal 



Inci Batmaz 


ibatmaz@metu.edu.tr 

Track(s): 47 

2 sessions 

EURO 24 - Lisbon 2010 STREAMS 

Data Mining in the Financial 

Sector 

Invited 

Vadim Strijov 

Computing Center of the Russian 

Academy of Sciences 

strijov@ccas.ru 

Track(s): 23 

4 sessions 

DEA and Performance 

Measurement 

Invited 

Dimitris Despotis 

University of Piraeus 


Ozren Despic 

Aston University 

o.despic@aston.ac.uk 

Meryem Duygun Fethi 

University of Leicester 

m.fethi@le.ac.uk 

Ana Camanho 

Universidade do Porto 

acamanho@fe.up.pt 

Vania Sena 

Aston University 

v.sena@aston.ac.uk 

Track(s): 6 7 

19 sessions 

Decision Making 

Contributed 

Track(s): 42 

4 sessions 

Decision Support Systems 

Invited 

Pascale Zaraté 

Toulouse University 

zarate@irit.fr 

Fatima Dargam 

SimTech Simulation Technology 

F.Dargam@SimTechnology.com 

Track(s): 37 

2 sessions 

Demand and Supply in Consumer 

Goods and Retailing 

Invited 

Heinrich Kuhn 

Catholic University of 

Eichstaett-Ingolstadt 

heinrich.kuhn@ku-eichstaett.de 

Winfried Steiner 

Clausthal University of Technology, 

Institute of Management and 

Economics 

winfried.steiner@tu-clausthal.de 

Rob Broekmeulen 

TU Eindhoven 


Track(s): 45 

2 sessions 

Demand, Pricing and Revenue 

Management 

Invited 

Alf Kimms 

University of Duisburg-Essen 

alf.kimms@uni-due.de 

Robert Klein 

Universität Augsburg 

robert.klein@wiwi.uni-augsburg.de 

Track(s): 43 

3 sessions 

Discrete and Global Optimization 

Invited 

Xiaoling Sun 

Fudan University 

xls@fudan.edu.cn 


1 session 

Discrete Optimal Control 

Invited 

Dmitrii Lozovanu 

Academy of Sciences of Moldova 





Track(s): 47 

2 sessions 

303

STREAMS EURO 24 - Lisbon 2010 

Dynamic Programming 

Invited 

Lidija Zadnik Stirn 

University of Ljubljana 


Moshe Sniedovich 

University of Melbourne 

m.sniedovich@ms.unimelb.edu.au 

Track(s): 45 

2 sessions 

Dynamical Systems and Game 

Theory 

Invited 

Alberto A. Pinto 

University of Minho 

aapinto1@gmail.com 

Track(s): 19 

9 sessions 

Dynamical Systems and 

Mathematical Modelling in OR 

Invited 

Selma Belen 

CAG University 

selmabelen@cag.edu.tr 

Track(s): 44 

3 sessions 

Education, and Social Policy 

Invited 

Hanife Akar 


hanif@metu.edu.tr 

Track(s): 21 

2 sessions 

Ejor 

Invited 

Roman Slowinski 

Poznan University of Technology 


Track(s): 5 

1 session 

Emerging Applications of OR 

Invited 




Erwin Pesch 

University of Siegen 


Track(s): 10 11 

15 sessions 

304 

Energy, Environment and Climate 

Invited 

Hans-Jakob Lüthi 

Ethz 

luethi@ifor.math.ethz.ch 

Wolf Fichtner 

Kit 

wolf.fichtner@wiwi.uni-karlsruhe.de 

Track(s): 33 

7 sessions 

Energy, Environment and Climate 

[c] 

Contributed 

Hans-Jakob Lüthi 

Ethz 

luethi@ifor.math.ethz.ch 

Wolf Fichtner 

Kit 

wolf.fichtner@wiwi.uni-karlsruhe.de 

Track(s): 33 

7 sessions 

Engineering Optimization 

Invited 

Moritz Diehl 

K. U. Leuven 

moritz.diehl@esat.kuleuven.be 

Volker Schulz 

University of Trier 

Volker.Schulz@uni-trier.de 

Track(s): 40 

4 sessions 

EURO Doctoral Dissertation 

Award 

Invited 

Mikael Rönnqvist 

Department of Finance and 

Management Science 

mikael.ronnqvist@nhh.no 

Track(s): 5 

1 session 

EURO Excellence in Practice 

Award 2010 

Invited 

M. Grazia Speranza 

University of Brescia 


Track(s): 5 

2 sessions 

EURO Management Science 

Strategic Innovation Prize (MSSIP 

2010) on the topic of Optimization 

in Telecommunications 

Invited 

Eric Gourdin 

Orange Labs 

eric.gourdin@orange-ftgroup.com 

Track(s): 2 

1 session 

Experimental Economics and 

Game Theory 

Invited 

Ulrike Leopold-Wildburger 

Karl-Franzens-University 

ulrike.leopold@uni-graz.at 

Stefan Pickl 

Universität der Bundeswehr München 

stefan.pickl@unibw.de 

Track(s): 38 

4 sessions 

Facilitated Modelling in OR 

Invited 

L. Alberto Franco 

University of Warwick 

alberto.franco@warwick.ac.uk 

Gilberto Montibeller 

London School of Economics 

g.montibeller@lse.ac.uk 

Track(s): 35 

5 sessions 

Facility Logistics 

Invited 

René de Koster 

Erasmus University Rotterdam 

rkoster@rsm.nl 

Track(s): 45 

2 sessions 

Financial Mathematics and OR 

Invited 

Mustafa Pinar 

Bilkent University 

mustafap@bilkent.edu.tr 




Track(s): 25 

9 sessions


Invited 

Georg Pflug 

University of Vienna 

georg.pflug@univie.ac.at 

Rita D’Ecclesia 

Università di Roma 

rita.decclesia@uniroma1.it 

Ronald Hochreiter 

WU Vienna University of Economics 

and Business 

ronald.hochreiter@wu.ac.at 

Track(s): 29 

6 sessions 

Financial Optimization 

Invited 

Gautam Mitra 

Brunel University 

gautam.mitra@brunel.ac.uk 

J. E. Beasley 

Brunel University 

john.beasley@brunel.ac.uk 

Track(s): 27 

4 sessions 

Fuzzy Systems, Neural Networks 

& Artificial Intelligence 

Invited 


J. W. Goethe University 

rommel@wiwi.uni-frankfurt.de 

Rasmus Fonseca 

University of Copenhagen 

rfonseca@diku.dk 

Track(s): 36 

6 sessions 

Fuzzy Systems, Neural Networks 

& Artificial Intelligence [c] 

Contributed 


J. W. Goethe University 

rommel@wiwi.uni-frankfurt.de 

Rasmus Fonseca 

University of Copenhagen 

rfonseca@diku.dk 

Track(s): 36 

3 sessions 


Geometric Clustering 

Invited 

Andreas Brieden 


andreas.brieden@unibw.de 

Peter Gritzmann 

TU München 

gritzman@ma.tum.de 

Track(s): 45 

1 session 

Global Optimization 

Invited 

Mirjam Duer 

Rijksuniversiteit Groningen 

M.E.Dur@rug.nl 

Eligius M.T. Hendrix 

Universidad de Málaga 

eligius.hendrix@wur.nl 

Janos D. Pinter 

PCS Inc. & Dalhousie University 

jdpinter@hfx.eastlink.ca 

Track(s): 43 

3 sessions 

Graph Searching and Guarding 

Invited 

Boting Yang 

University of Regina 

boting@cs.uregina.ca 

Track(s): 42 

3 sessions 

Graphs and Networks 

Invited 

Dominique de Werra 

Epfl 

dominique.dewerra@epfl.ch 


8 sessions 

Health Care Management 

Invited 

Marion Rauner 


marion.rauner@univie.ac.at 

Stefan Nickel 

Universitaet Karlsruhe 

Stefan.Nickel@kit.edu 

Teresa Melo 

University of Applied Sciences 

teresa.melo@htw-saarland.de 

Sally Brailsford 

University of Southampton 

s.c.brailsford@soton.ac.uk 

Track(s): 22 

3 sessions 

Health Care Management [c] 

Contributed 

Marion Rauner 


marion.rauner@univie.ac.at 

Stefan Nickel 



Teresa Melo 

University of Applied Sciences 

teresa.melo@htw-saarland.de 

Sally Brailsford 


s.c.brailsford@soton.ac.uk 

Track(s): 22 

3 sessions 

IBM Research Applications 

Invited 

Eleni Pratsini 

IBM Zurich Research Lab 

pra@zurich.ibm.com 

Track(s): 14 

1 session 

Ill-posed Variational Problems - 

Theory, Methods and 

Applications 

Invited 

Xiaoqi Yang 

The Hong Kong Polytechnic University 

mayangxq@polyu.edu.hk 

Track(s): 48 

2 sessions 

305


Inventory Management 

Invited 

Dolores Romero Morales 

University of Oxford 

dolores.romero-morales@sbs.ox.ac.uk 

Track(s): 38 

4 sessions 

Inventory Management [c] 

Contributed 

Dolores Romero Morales 

University of Oxford 

dolores.romero-morales@sbs.ox.ac.uk 

Track(s): 38 

1 session 

Iterative Methods for Economic 

Models 

Invited 

Adriana Gnudi 

University of Bergamo 

adriana.gnudi@unibg.it 

Elisabetta Allevi 

University of Brescia 

allevi@eco.unibs.it 

Igor Konnov 

University of Kazan 

Igor.Konnov@ksu.ru 

Track(s): 48 

2 sessions 

Keynote Speakers 

Invited 

José Paixão 

Faculty of Sciences - University of 

Lisbon 

jpaixao@fc.ul.pt 

Track(s): 1 2 

14 sessions 

Knowledge Discovery and Data 

Mining 

Invited 

Julie Greensmith 

University of Nottingham 

jqg@cs.nott.ac.uk 

Luís Cavique 

Universidade Aberta 

lcavique@univ-ab.pt 


1 session 

306 

Life Insurance, Risk Management 

& OR 

Invited 

Angelika May 

Carl von Ossietzky Universität 

Oldenburg 

angelika.may@uni-oldenburg.de 

Irini Dimitriyadis 

Bahcesehir University 

dimitri@bahcesehir.edu.tr 

Track(s): 43 

2 sessions 

Linear and Conic Programming 

Invited 

François Glineur 

Université catholique de Louvain 

(UCLouvain) 

Francois.Glineur@uclouvain.be 

Tibor Illes 

University of Strathclyde 

tibor.illes@strath.ac.uk 

Track(s): 36 

2 sessions 

Location Analysis 

Invited 

Stefan Nickel 



Francisco Saldanha-da-Gama 

University of Lisbon 

fsgama@fc.ul.pt 

Alfredo Marín 

University of Murcia 

amarin@um.es 

Track(s): 5 13 

16 sessions 

Long Term Financial Decisions 

Invited 

Thomas Burkhardt 

Universitaet Koblenz-Landau 

tburkha@uni-koblenz.de 

Track(s): 41 

3 sessions 

Long Term Planning in Energy, 

Environment and Climate 

Invited 

Nadia Maïzi 

MINES ParisTech 

nadia.maizi@mines-paristech.fr 

Track(s): 32 37 

4 sessions 

Lot-sizing and Scheduling, 

Economic Order Quantity 

Invited 

Anders Segerstedt 

Luleå University of Technology 

anders.segerstedt@ltu.se 

Christian Almeder 

Vienna University of Economics and 

Business 

christian.almeder@wu.ac.at 

Bernardo Almada-Lobo 

Faculty of Engineering of Porto 

University 


Track(s): 34 

7 sessions 

Machine Learning and Its 

Applications 

Invited 

Sureyya Ozogur-Akyuz 

Bahcesehir University 

sureyya.akyuz@bahcesehir.edu.tr 

Zakria Hussain 

Unvertsity College London 

Z.Hussain@cs.ucl.ac.uk 

Track(s): 26 

5 sessions 

Maritime Logistics 

Invited 

Marielle Christiansen 

Norwegian University of Science and 

Technology 


Track(s): 22 

4 sessions 


Invited 

Sandor Zoltan Nemeth 

The University of Birmingham 

nemeths@for.mat.bham.ac.uk 

Armin Fügenschuh 

Zuse Institut Berlin 

fuegenschuh@zib.de 




Tamas Terlaky 

Lehigh University 

terlaky@lehigh.edu 

Track(s): 9 

14 sessions

Mathematical Programming [c] 

Contributed 




Armin Fügenschuh 

Zuse Institut Berlin 

fuegenschuh@zib.de 

Sandor Zoltan Nemeth 

The University of Birmingham 

nemeths@for.mat.bham.ac.uk 

Tamas Terlaky 

Lehigh University 

terlaky@lehigh.edu 

Track(s): 7 

3 sessions 

MCDA I: New Approaches and 

Applications 

Invited 

Salvatore Greco 

University of Catania 

salgreco@unict.it 

Bernard Roy 

Université Paris-Dauphine 

roy@lamsade.dauphine.fr 

Track(s): 37 

6 sessions 

MCDA II: Axiomatic Basis, 

Meaningfulness, and other 

Issues 

Invited 

José Rui Figueira 

Technical University of Lisbon 

figueira@ist.utl.pt 

Denis Bouyssou 

Cnrs-lamsade 

bouyssou@lamsade.dauphine.fr 

Track(s): 30 

8 sessions 

MCDA II: Axiomatic Basis, 

Meaningfulness, and other 

Issues [c] 

Contributed 




Denis Bouyssou 

Cnrs-lamsade 

bouyssou@lamsade.dauphine.fr 

Track(s): 30 

3 sessions 


Metaheuristics 

Invited 

Marc Sevaux 

Université de Bretagne Sud - UEB 

marc.sevaux@univ-ubs.fr 

Kenneth Sörensen 

Universiteit Antwerpen 

kenneth.sorensen@ua.ac.be 

Andreas Reinholz 

University Dortmund 

andreas.reinholz@gmx.de 

Track(s): 3 4 5 

33 sessions 

Methodology of Societal 

Complexity 

Invited 

Dorien DeTombe 

Chair Euro Working Group 

detombe@nosmo.nl 

Track(s): 31 

4 sessions 

Mixed-Integer Non Linear 

Programming 

Invited 

Andrea Lodi 


andrea.lodi@unibo.it 

Leo Liberti 

Ecole Polytechnique 

leoliberti@gmail.com 

Track(s): 35 

6 sessions 

Multi-Objective Optimization 

Invited 




Jacques Teghem 

Faculté Polytechnique de Mons 


Kathrin Klamroth 

University of Wuppertal 

klamroth@math.uni-wuppertal.de 

Track(s): 23 

10 sessions 

Network Optimization 

Invited 

Bernard Fortz 

Université Libre de Bruxelles 

bfortz@euro-online.org 

Luis Gouveia 


legouveia@fc.ul.pt 

Track(s): 40 

4 sessions 

Network Optimization [c] 

Contributed 

Bernard Fortz 

Université Libre de Bruxelles 

bfortz@euro-online.org 

Luis Gouveia 


legouveia@fc.ul.pt 

Track(s): 40 

2 sessions 

Nonconvex Programming: Local 

and Global Approaches 

Invited 

Tao Pham Dinh 

INSA Rouen 


Hoai An Le Thi 

University Paul Verlaine - Metz 

lethi@univ-metz.fr 

Track(s): 45 

3 sessions 

Nonlinear Programming 

Invited 

Edite M.G.P. Fernandes 


emgpf@dps.uminho.pt 

Track(s): 48 

3 sessions 

Nonlinear Programming [c] 

Contributed 

Edite M.G.P. Fernandes 


emgpf@dps.uminho.pt 

Track(s): 48 

2 sessions 

307


Nonsmooth Optimization 

Invited 

Adil Bagirov 

University of Ballarat 


Albert Ferrer 

Technological University of Catalonia 

alberto.ferrer@upc.edu 

Antonio Fuduli 

Universita’ della Calabria 

antonio.fuduli@unical.it 

Track(s): 19 

4 sessions 

Numerical Methods in Finance 

Invited 

Ömür Ugur 



Susanne Kruse 

Hochschule der 

Sparkassen-Finanzgruppe 

susanne.kruse@dsgv.de 

Track(s): 46 

2 sessions 

Operational Research and 

Quantitative Models in Banking 

Invited 

Constantin Zopounidis 

Technical University of Crete 

kostas@dpem.tuc.gr 

Track(s): 30 

4 sessions 

308 

Optimal Control 

Invited 

Erik Kropat 


erik.kropat@unibw.de 

Gustav Feichtinger 

Vienna University of Technology 

gustav@eos.tuwien.ac.at 




Ursula Felgenhauer 

Brandenburg University of Technology 

Cottbus (Germany) 

felgenh@tu-cottbus.de 

Vladimir Veliov 

Vienna University of technology 

veliov@tuwien.ac.at 

Sabine Pickenhain 

BTU Cottbus 

sabine.pickenhain@tu-cottbus.de 

Track(s): 39 

9 sessions 

Optimization and Data Mining 

Invited 

Emilio Carrizosa 

Universidad de Sevilla 

ecarrizosa@us.es 

Theodore Trafalis 

University of Oklahoma 

ttrafalis@ou.edu 

Renato De Leone 

Università di Camerino 

renato.deleone@unicam.it 

Track(s): 42 

2 sessions 

Optimization for Sustainable 

Development 

Invited 

Nora Touati Moungla 

LIX, Ecole Polytechnique 

touati@lix.polytechnique.fr 

Vincent Jost 

CNRS - Ecole Polytechnique 

vjost@lix.polytechnique.fr 

Track(s): 48 

2 sessions 

OR and Ethics 

Invited 

Fred Wenstøp 

BI Norwegian School of Management 

fred.wenstop@bi.no 

Track(s): 31 

2 sessions 

OR and Real Implementations 

Invited 

Hans-Jürgen Zimmermann 

Inform 

zi@or.rwth-aachen.de 

Istvan Maros 

Imperial College London 

i.maros@imperial.ac.uk 

Track(s): 36 

4 sessions 

OR Applications in Industry 

Invited 

Jens Wollenweber 

Fraunhofer SCS 

jens.wollenweber@scs.fraunhofer.de 

Geir Hasle 

Sintef Ict 

Geir.Hasle@sintef.no 

Track(s): 31 

6 sessions 

OR for Development and 

Developing Countries 

Invited 




Honora Smith 


honora.smith@soton.ac.uk 

Leroy White 

University of Bristol 


Hans W. Ittmann 

CSIR Built Environment 

hittmann@csir.co.za 

Track(s): 37 

5 sessions 

OR for Madeira (and related 

challenges) 

Invited 




Track(s): 46 

3 sessions

OR for Sustainable Development 

Invited 

Leonidas Sakalauskas 

Institute of Mathematics&Informatics 

sakal@ktl.mii.lt 

Edmundas Zavadskas 

Vilnius Gediminas Technical University 

edmundas.zavadskas@adm.vtu.lt 

Tatjana Vilutiene 

Viulnius Gediminas Technical 

University 

tatjana.vilutiene@vgtu.lt 

Vida Maliene 

School of the Built Environment, 

Liverpool John Moores University 

v.maliene@ljmu.ac.uk 

Track(s): 43 

3 sessions 

OR in Agriculture and Forest 

Management 

Invited 

LluisM Pla 

University of Lleida 

lmpla@matematica.udl.es 

Manfred Gronalt 

University of Natural Resources and 

Applied Life Sciences 


Track(s): 32 

8 sessions 

OR in Agriculture and Forest 

Management [c] 

Contributed 

LluisM Pla 

University of Lleida 

lmpla@matematica.udl.es 

Manfred Gronalt 

University of Natural Resources and 

Applied Life Sciences 


Track(s): 32 

3 sessions 

OR in Fisheries, Maritime 

Sciences and Related Aspects 

Invited 

Pall Jensson 

University of Iceland 

pall@hi.is 

Track(s): 46 

1 session 


OR in Military 

Invited 

Ana Isabel Barros 

TNO - Defense, Security and Safety 

ana.barros@tno.nl 

Track(s): 45 

2 sessions 

OR in Oil Sector 

Invited 

Irina Dolgopolova 



Track(s): 47 

2 sessions 

OR in Sports 

Invited 

Michael Trick 

Carnegie Mellon University 

trick@cmu.edu 

Sigrid Knust 

TU Clausthal 

knust@math.tu-clausthal.de 


2 sessions 

OR in Water Management 

Invited 

Halil Önder 


onde@metu.edu.tr 

Track(s): 47 

2 sessions 

OR/MS: Beyond Mathematics 

Invited 

Heiner Müller-Merbach 

Universität Kaiserslautern 

hmm@bior.de 

Track(s): 48 

1 session 

Portfolio Decision Analysis 

Invited 

Ahti Salo 

Aalto University School of Science and 

Technology 

ahti.salo@tkk.fi 

Alec Morton 

London School of Economics 

a.morton@lse.ac.uk 

Jeffrey Keisler 

University of Massachusetts Boston 

jeff.keisler@umb.edu 

Track(s): 44 

3 sessions 

Preference Learning 

Invited 

Roman Slowinski 

Poznan University of Technology 


Track(s): 29 

1 session 

Project Management and 

Scheduling 

Invited 

Erwin Pesch 


erwin.pesch@uni-siegen.de 

Track(s): 7 8 

19 sessions 

Project Management and 

Scheduling [c] 

Contributed 

Erwin Pesch 


erwin.pesch@uni-siegen.de 

Track(s): 7 

2 sessions 

Public Transport 

Invited 

Leo Kroon 


lkroon@rsm.nl 


Georg-August Universität Göttingen 

schoebel@math.uni-goettingen.de 

Track(s): 16 

11 sessions 

Public Transport [c] 

Contributed 

Leo Kroon 


lkroon@rsm.nl 


Georg-August Universität Göttingen 

schoebel@math.uni-goettingen.de 

Track(s): 16 

2 sessions 

Realistic Production Scheduling 

Invited 

Ruben Ruiz 

Universidad Politecnica de Valencia 

rruiz@eio.upv.es 

Track(s): 33 

1 session 

309


Revenue Management 

Invited 

Houyuan Jiang 

University of Cambridge 

h.jiang@jbs.cam.ac.uk 

Ayse Kocabiyikoglu 

Bilkent University 

aysekoca@bilkent.edu.tr 

Track(s): 41 

5 sessions 

ROADEF/EURO challenge 

Invited 

Christian Artigues 

Cnrs 

artigues@laas.fr 

Track(s): 25 

6 sessions 

Scheduling 

Invited 

Vitaly Strusevich 

University of Greenwich 

sv02@gre.ac.uk 

Jean-Charles Billaut 

University of Tours 

jean-charles.billaut@univ-tours.fr 

Track(s): 28 

7 sessions 

Scheduling under Resource 

Constraints 

Invited 

Jan Weglarz 

Politechnika Poznanska 

Jan.Weglarz@put.poznan.pl 

Joanna Jozefowska 

Poznañ University of Technology 

jjozefowska@cs.put.poznan.pl 

Track(s): 39 

5 sessions 

SD Modeling in Sustainable 

Development 

Invited 

Pierre Kunsch 

Vrije Universiteit Brussel 

pkunsch@vub.ac.be 

Erik Pruyt 

Delft University of Technology 

E.Pruyt@tudelft.nl 

Track(s): 44 

3 sessions 


Semi-Infinite Optimization 

Invited 

Jan-J Ruckmann 

University of Birmingham 

J.Ruckmann@bham.ac.uk 

Oliver Stein 

Karlsruhe Institute of Technology 

stein@kit.edu 

Track(s): 46 

2 sessions 

Simulation Based Decision 

Support 

Invited 

Miroljub Kljajic 

University of Maribor 

miroljub.kljajic@fov.uni-mb.si 

Stig C Holmberg 

Mid Sweden University 

shbg@ieee.org 

Track(s): 44 

3 sessions 

Simulation Methods in Finance 

Invited 

Giray Okten 

Florida State University 

okten@math.fsu.edu 

Track(s): 41 

2 sessions 

Soft OR and Problem Structuring 

Methods 

Invited 

John Mingers 

Kent University 

j.mingers@kent.ac.uk 

Track(s): 35 

4 sessions 

Software for OR/MS 

Invited 

Robert Fourer 

Northwestern University 

4er@iems.northwestern.edu 

Bjarni Kristjansson 

Maximal Software, Ltd. 


Track(s): 21 

13 sessions 

Stochastic Modeling and 

Simulation 

Invited 

Erik Kropat 


erik.kropat@unibw.de 

Zeev (Vladimir) Volkovich 

Ort Braude Academic College 

zeev@actcom.co.il 

Track(s): 18 

13 sessions 

Stochastic Models for Service 

Operations 

Invited 

Ger Koole 

VU University Amsterdam 

koole@few.vu.nl 

Track(s): 47 

1 session 

Stochastic Programming 1 

Invited 

Rüdiger Schultz 

University of Duisburg-Essen 

schultz@math.uni-duisburg.de 

Track(s): 28 

7 sessions 

Stochastic Programming 2 

Invited 

Andras Prekopa 


prekopa@rutcor.rutgers.edu 

Tamas Szantai 

Budapest University of Technology and 

Economics 

szantai@math.bme.hu 

Track(s): 46 

2 sessions 

Stochastic Programming 2 [c] 

Contributed 

Andras Prekopa 


prekopa@rutcor.rutgers.edu 

Tamas Szantai 

Budapest University of Technology and 

Economics 

szantai@math.bme.hu 

Track(s): 46 

1 session

Stochastic Valuation for Financial 

Markets 

Invited 

Martin Rainer 

ENAMEC Inst. 

martin.rainer@enamec.de 




Track(s): 38 

4 sessions 

Supply Chain Planning 

Invited 

Moritz Fleischmann 

University of Mannheim 

Moritz.Fleischmann@bwl.unimannheim.de 

Herbert Meyr 

Technical University of Darmstadt 

Meyr@bwl.tu-darmstadt.de 

Track(s): 14 

3 sessions 

Supply Chain Planning [c] 

Contributed 

Moritz Fleischmann 

University of Mannheim 

Moritz.Fleischmann@bwl.unimannheim.de 

Herbert Meyr 

Technical University of Darmstadt 

Meyr@bwl.tu-darmstadt.de 

Track(s): 14 

6 sessions 

Sustainable Living: Cognitive, 

Social, Economical, Ecological 

and World View 

Invited 

Ali Gökmen 


agokmen@metu.edu.tr 

Inci Gokmen 


igokmen@metu.edu.tr 

Pedamallu Chandra Sekhar 

Dana-Farber Cancer Institute 

pcs.murali@gmail.com 




Track(s): 47 

1 session 


System Dynamics Modeling 

Invited 

Markus Schwaninger 

Universität St.Gallen 

markus.schwaninger@unisg.ch 

Stefan Groesser 

University of St. Gallen 

stefan.groesser@unisg.ch 

Track(s): 41 

4 sessions 

Teaching OR/MS 

Invited 

Antonia Carravilla 

FEUP / INESC Porto 

mac@fe.up.pt 

Track(s): 22 

4 sessions 

Timetabling and Rostering 

Invited 

Dario Landa-Silva 

University of Nottigham 

dario.landasilva@nottingham.ac.uk 

Track(s): 24 

6 sessions 

Transportation and Logistics 

Invited 

Maurizio Bielli 

Institute of Systems Analysis and 

Informatics 

bielli@iasi.cnr.it 

Track(s): 27 

7 sessions 

Transportation and Logistics [c] 

Contributed 

Maurizio Bielli 

Institute of Systems Analysis and 

Informatics 

bielli@iasi.cnr.it 

Track(s): 27 

2 sessions 

Transportation Planning 

Invited 

Herbert Kopfer 

University of Bremen 


Knut Haase 

Technische Universität Dresden 

knut.haase@tu-dresden.de 

Track(s): 17 

11 sessions 

Variational Inequalities, 

Complementarity Problems and 


Invited 

Stephan Dempe 

Technische Universitaet Freiberg 

dempe@math.tu-freiberg.de 

Joaquim Judice 

University of Coimbra 

Joaquim.Judice@co.it.pt 

Track(s): 42 

4 sessions 

Vector and Set-Valued 

Optimization 

Invited 

Vicente Novo 

Universidad Nacional de Educacion a 

Distancia 

vnovo@ind.uned.es 

Bienvenido Jiménez 

Uned 

bjimenez@ind.uned.es 

César Gutiérrez 

Universidad de Valladolid 

cesargv@mat.uva.es 

Track(s): 44 

3 sessions 

Vehicle Routing 

Invited 

Jean-François Cordeau 

HEC Montréal 

jean-francois.cordeau@hec.ca 

Stefan Ropke 

Technical University of Denmark 

sr@transport.dtu.dk 

Track(s): 15 

12 sessions 

Vehicle Routing [c] 

Contributed 

Jean-François Cordeau 

HEC Montréal 

jean-francois.cordeau@hec.ca 

Stefan Ropke 

Technical University of Denmark 

sr@transport.dtu.dk 

Track(s): 15 

3 sessions 

311


Young People for System Theory, 

Optimization and Education 

Invited 

Alexis Pasichny 

National Technical University of 

Ukraine "Kyiv Politechnic Institute" 

alexis.pasichny@gmail.com 

Kateryna Pereverza 


Ukraine 

pereverza.kate@gmail.com 

Alexander Makarenko 


Ukraine "KPI" 

makalex@i.com.ua 

Ielyzaveta Korotchenko 


Ukraine "Kyiv Politechnic Institute" 

Korotchenko.liza@gmail.com 

Track(s): 47 

2 sessions 

312

A 

A.C. Rocha, Ana Maria TA-48, TB-48 

Ackermann, Fran MC-35 

Afsar, H. Murat WB-25 

Airoldi, Mara MB-30 

Akar, Hanife WD-21, WE-21 

Akartunali, Kerem TB-27 

Akkan, Can WE-28 

Aksakal, Erdem MB-12 

Alcaraz, Javier TF-03 

Algaba, Encarnación TB-26 

Allevi, Elisabetta WA-33 

Almada-Lobo, Bernardo MB-34, TA-34 

Almeida, João TD-19 

Alparslan Gok, Sirma Zeynep WB-26, 

WD-26 

Alvarez, Jose Fernando WD-22 

Alvarez-Valdes, Ramon MC-20 

Alvelos, Filipe WC-05 

Amodeo, Lionel MA-04, WC-04, 

MC-05 

Andersson, Henrik WC-22, TC-38 

Andrade, Jose luis TD-22 

Antunes, Carlos Henggeler TC-33 

Aparisi, Francisco MD-05 

Arantes, Amílcar TF-22 

Araujo, Jonata TC-03 

Argyris, Nikolaos MB-44 

Armentano, Vinícius TD-21 

Artigues, Christian TF-25, WE-25 

Asproth, Viveca TC-44 

Assoumou, Edi TA-37 

Astorino, Annabella WB-19 

Atkin, Jason MC-08 

Öberg, Lena-Maria MB-11 

Önder, Halil TD-47, TF-47 

B 

Baes, Michel WB-34 

Baesens, Bart WC-30 

Baier, Robert MB-39, TB-39 

Baldacci, Roberto MD-15 

Barrieu, Pauline WB-33, WD-33 

Barros, Ana Isabel TC-45, TD-45 

Battiti, Roberto WA-05 

Batur, Gul Didem WA-28 

Bauman, Evgeny WC-23 

Bayraktaroglu, Ayse Elvan WE-08 

Bektas, Tolga MC-15 

Belen, Selma WC-44, WD-44, WE-44 

Bell, Peter MB-22 

Belov, Gleb TB-20 

Belton, Valerie TB-01 

Ben Amor, Sarah TF-37 

Benkhellat, Zahira MA-05 

Bentz, Cédric MD-10 

Bergantinos, Gustavo TA-26 

Bielza, Concha TA-42 

Bierwirth, Christian TB-17, TD-17 

Biga, Veronica MA-11 

Bjørndal, Mette TA-33 

Blazewicz, Jacek WC-02, MA-24, 

MB-24, MC-24 

Bloos, Melanie MA-17, TA-17, TF-17 

Bock, Stefan MC-17 

Boggia, Antonio WA-37 

Bonato, Anthony WE-42 

Borges, Paulo TC-32 

Boros, Endre MC-29 

Bortfeldt, Andreas MB-20 

Bosso de Freitas, Gisele TD-41 

Boucekkine, Raouf MA-39 

Bourreau, Eric WA-25, WD-25 

Bouyssou, Denis TC-01 

Branda, Martin MF-28 

Brandao, Jose TB-03 

Brans, Jean-Pierre TA-44 

Branzei, Mariana Rodica WB-26, 

WC-26, WD-26 

Brás, Raul TD-32 

Brieden, Andreas TB-45 

Briskorn, Dirk TD-07, TC-08 

Broekmeulen, Rob WD-45 

Brugha, Cathal MB-31 

Bruglieri, Maurizio WB-27 

Brunner, Jens TB-22 

Buchheim, Christoph WC-35 

Burinskien, Marija WA-43 

Burke, Edmund WC-08, WD-08 

C 

Cafieri, Sonia WA-35 

Caglar, Burcu TC-36 

Caimi, Gabrio Curzio MC-16 

Canca, David MF-21 

Cancer, Vesna MD-35 

Cardoso, Domingos MB-07, WC-09 

Cardoso, Gonçalo TB-33 

Carpente, M a Luisa TC-26 

Carrizosa, EmilioTA-13, TB-13, MF-42 

Casado, Leocadio G. TC-43 

Caserta, Marco MA-34 

Castro, Pedro MF-20 

Cayirli, Tugba TA-22 

Chabchoub, Habib MA-37 

Chandra Sekhar, Pedamallu MA-47 

Chen, Yi-Chun TF-12 

Chirenje, Kelvin T WE-38 

Christiansen, Marielle MC-01, WB-22 

Chubanov, Sergei WB-08, WD-36 

Clark, Alistair MD-34, TB-34 

Clarke, Nancy WD-42 

Claro, João WA-05 

Clasen, Christian WA-41 

Colaco, Susana MA-22 

Collier, David MC-30 

Condotta, Alessandro WC-28 

Consigli, Giorgio WE-30 

Constantino, Miguel MC-32 

Cook, Wade MC-06 

Corberan, Angel TB-15 

Corbett, Charles TD-14 

Correia, Isabel MB-13 

Costa, Lino TD-48 

Costa, Ruy MD-22 

Cristescu, Gabriela TF-34 

Cunha, Maria TF-47 

SESSION CHAIR INDEX 

D 

Dall’Aglio, Marco TD-26 

Daniele, Patrizia WA-48 

Dargam, Fatima TC-37, TD-37 

Dash, Gordon TF-21 

De Baets, Bernard TD-04 

de Carvalho, J. M. Valério TB-02, 

TD-20 

de Frutos, Javier TB-38 

de Rooij, Frans WE-46 

De Schepper, Steven TA-44 

de Sousa, Amaro MC-40 

de Werra, Dominique ME-01 

Deelstra, Griselda WD-14 

Della Croce, Federico MC-43 

Demange, Marc TA-10 

Demir, Aysegul TA-14 

Dempe, Stephan MA-42 

Despotis, Dimitris MA-06, TC-07 

DeTombe, Dorien MD-31 

Di Giacomo, Laura MC-27 

Dias, Luis C. TA-07 

Diehl, Moritz WB-40 

Dimitriyadis, Irini WE-43 

Dolgopolova, Irina MC-47, MD-47 

Dolgui, Alexandre WD-39, WE-39 

Dorndorf, Ulrich WC-36 

Doumpos, Michael MD-30 

Drewes, Sarah WD-35 

Duer, Mirjam TD-43 

Duggan, Jim TA-41 

Duhamel, Christophe TD-03, TC-04 

Dutta, Joydeep MD-42 

E 

Eden, Colin MB-35 

Egriboyun, Feyzullah TC-25 

Ehrgott, Matthias MD-23 

Emrouznejad, Ali WE-06 

Erdogan, Gunes TA-15 

Erensal, Yasemin C. TC-12 

Eriksson, Ola MD-32 

Erlwein, Christina MD-27 

Ernst, Andreas TF-08 

Escudero, Laureano Fernando TC-05 

Eshragh Jahromi, Ali WB-42 

Espírito Santo, Isabel TC-48 

Estellita Lins, Marcos MF-06, MC-31 

Euchi, Jalel TD-03 

Euler, Reinhardt MA-10 

F 

Faccio, Maurizio MD-04 

Fügenschuh, Armin MF-09, TF-09, 

TF-16 

Farago, Andras MF-40 

Fawzi, Bessaih TA-40 

Fehr, Max WB-33, WD-33 

Feichtinger, Gustav TD-39 

Felici, Giovanni TC-24 

Fernandes, Edite M.G.P. WA-10, TF-48 

Fernandez, Elena WE-13 

Fernández Barberis, Gabriela TC-30 

Ferreira, Carlos TC-23 

313

SESSION CHAIR INDEX EURO 24 - Lisbon 2010 

Ferreira, Fernanda A. TB-19 

Ferreira, Flávio TD-19 

Ferreira, Helena MB-19 

Ferreira, Luis MD-19 

Ferrucci, Francesco MC-17 

Fertis, Apostolos MF-25 

Festa, Paola TB-24 

Fethi, Meryem Duygun MF-07 

Figueira, José Rui MF-23 

Filar, Jerzy MA-07 

Fischetti, Matteo WC-05 

Floudas, Chris TF-35 

Fodstad, Marte MC-28 

Fonlupt, Jean MB-10 

Formanowicz, Piotr MA-24 

Forsell, Nicklas WC-32 

Forster, Geoffrey WD-24 

Fortz, Bernard MA-40, MD-40, MF-41 

Fourer, Robert MC-21 

Francesco, Cesarone MA-27 

Franco, L. Alberto TB-35 

Freire de Sousa, Jorge WC-27 

French, Alan TD-42 

Fuduli, Antonio WB-19 

Funk, Matthias WE-20 

Furmans, Kai WA-45 

G 

Gabriel, Steven MC-11 

Galli, Laura MB-16 

Gallo, Mariano WA-17 

Gan, Heng-Soon WE-22 

Günther, Markus TC-46 

Gürel, Sinan WC-07 

Garatti, Simone WA-21 

Garcia-Melon, Monica TB-12 

Gardi, Frédéric WE-05 

Gökmen, Ali MA-47 

Göttlich, Simone MB-09, TD-09 

Geiger, Martin Josef TD-05 

Gel, Esma MC-14 

Genoese, Massimo TB-46, TC-46, 

TD-46 

Georgiou, Andreas MD-11 

Geraldes, Carla A. S. MD-14 

Giallombardo, Giovanni WA-40 

Giannikos, Ioannis MA-13 

Ginda, Grzegorz MA-12 

Gkoumas, Konstantinos TD-16, WA-16 

Glass, Celia WD-28 

Gnudi, Adriana WB-48 

Goensch, Jochen TA-43 

Gokmen, Inci MA-47 

Gomes, A. Miguel TF-20 

Gomes, Marta Castilho MD-33 

Gomide, Fernando MC-36 

Gonçalves, José Fernando TA-20 

Gonçalves, Rui TC-19 

Gonzalez, Marcela TB-32 

Gourdin, Eric WE-02 

Gouveia, Luis WC-15, MB-40 

Gozzi, Fausto MD-39 

Grad, Sorin-Mihai TB-09, WD-09 

Gradisar, Miro TD-44 

Griebsch, Susanne WD-46 

Grigoroudis, Evangelos MF-30 

314 

Gritzmann, Peter TB-45 

Grobleben, Katharina TA-27 

Gronalt, Manfred MA-32 

Grunwald, Jessika TF-42 

Grusho, Alexander MB-18 

Gungor, Burcu MB-26 

Gussmagg-Pfliegl, Elisabeth TB-03 

H 

Habenicht, Walter MC-05 

Hadjiconstantinou, Eleni MD-02 

Hakanen, Jussi MC-07, TD-23 

Hügel, Katrin TB-41 

Hasle, Geir WA-31, WB-31, WC-31 

Haus, Utz-Uwe TB-29 

Hayashi, Shunsuke MF-46 

Höfer, Mathias TD-47 

Helin, Janne TF-32 

Hendrix, Eligius M.T. TF-43 

Heredia, F.-Javier WC-19 

Herrmann, Sascha TD-31 

Herroelen, Willy WA-07 

Heyde, Frank TF-23 

Hildmann, Marcus WB-23 

Hochreiter, Ronald WA-29 

Holmberg, Stig C MB-11, TC-44 

Htiouech, Skander WB-04 

Hudec, Miroslav TD-36 

Huisman, Dennis TC-16, WC-16 

Hunjak, Tihomir WA-30 

Husain, Iqbal WC-34 

Hussain, Zakria MD-26 

Hvattum, Lars Magnus WE-05, TC-15 

I 

Ilk, Ozlem WD-47 

Ioakimidis, Christos MF-33 

J 

Jablonsky, Josef MF-12 

Jaehn, Florian MB-08, TB-08 

Jütte, Silke WC-24 

Jensson, Pall TA-46 

Jiang, Houyuan MF-41 

Jimenez-Lopez, Mariano MB-36, 

MC-36 

Joerin, Florent WC-37 

Josa-Fombellida, Ricardo WC-14 

Judice, Joaquim MB-42 

Junglas, Daniel WB-36 

Junker, Ulrich MA-43 

K 

Kandakoglu, Ahmet WB-12 

Küfer, Karl-Heinz MB-17 

Kürüm, Efsun WB-17 

Karagiannis, Giannis TA-06 

Karimov, Azar TA-25 

Karmitsa, Napsu WD-19 

Karpak, Birsen TA-12 

Karpov, Igor WB-28 

Kartal Koc, Elcin WD-17 

Kayakutlu, Gulgun WC-12 

Keisler, Jeffrey MC-44 

Khachay, Michael WA-23, TB-42, 

TC-42 

Kim, Kap Hwan TF-40 

Kimms, Alf MF-43 

Kinoshita, Eizo WD-12 

Kis, Tamas WB-21 

Klatt, Tobias WC-47 

Kliewer, Natalia MD-16 

Kljajic, Miroljub TB-44 

Kocabiyikoglu, Ayse MD-41 

Kochetov, Yury TC-03 

Koksal, Gulser WA-11, WB-11 

Kopfer, Herbert MA-17, MB-17 

Koppang, Haavard TB-31 

Kostrzewska, Marta WB-46 

Kovalyov, Mikhail WD-39, WE-39 

Krass, Dmitry TC-13 

Kristensen, Anders Ringgaard MF-32 

Kristjansson, Bjarni MD-21, TC-21 

Kroon, Leo WB-16 

Kropat, Erik MA-18, MD-18, MF- 

18, TA-18, TB-18, TC-18, 

TD-18, TF-18, WA-18, WB- 

18, WC-18, MC-39, WC-44, 

WD-44, WE-44, MC-46 

Kruger, Alexander WE-19 

Kubiak, Wieslaw WB-39 

Kuhn, Daniel TF-36 

Kuhn, Heinrich WC-45 

Kunsch, Pierre MD-44 

Kuzmina, Lyudmila MC-39 

L 

Lacomme, Philippe MC-03, MD-03, 

MF-03, WD-04 

Laguna, Manuel TF-03 

Lamond, Bernard MB-21 

Landete, Mercedes WA-13 

Lau, Hoong Chuin TF-28 

Laumanns, Marco WE-14, MC-16 

Lavor, Carlile MC-02 

Le Thi, Hoai An MD-45 

Lee, Lai-Soon WB-20 

Lee, Loo Hay TF-40 

Lee, Wenyih TB-14 

Lehmann, Thomas MD-09 

Leopold-Wildburger, Ulrike WD- 

02, MA-38, MB-38, MC-38, 

MD-38 

Levin, Yuri MB-41 

Levina, Tatsiana MB-41 

Leyva-Lopez, Juan Carlos TB-30 

Li, Jing-An TD-40 

Liao, Shuangqing WA-24 

Liberti, Leo WB-35, WE-35 

Libura, Marek TD-02 

Liers, Frauke MB-29 

Liesiö, Juuso MA-44 

Ljubic, Ivana MA-02, WC-13 

Lo, Mei-Chen TD-12 

Lodi, Andrea WE-35 

Lopez, Pierre WA-08 

Loukil, Taicir TA-23 

Lozano, Sebastián TD-06 

Lozovanu, Dmitrii MF-47, TA-47 

Lukasiak, Piotr MB-24 

Luptacik, Mikulas WC-06 

M 

Maassen, Klaus-Christian TF-27

Maïzi, Nadia TB-37, TD-38 

Macedo, Rita WE-15 

Maculan Filho, Nelson MC-02 

Madlener, Reinhard MA-33 

Madrid-Sanchez, Jaisiel MC-26 

Mahjoub, A. Ridha TB-10 

Makarenko, Alexander WA-47 

Maksa, Gyula TC-34 

Mannino, Carlo TF-31 

Müller-Merbach, Heiner MF-48 

Marciniak, Dorota WE-26 

Maria Franca, Norese WB-37 

Markov, Michael WD-23, TB-25 

Marques, Rui WB-06 

Martein, Laura WA-34 

Martello, Silvano TE-01 

Martens, David WC-30 

Martin, Quintin TD-04 

Martins, Isabel MC-32 

Martins, José TA-19 

Masmoudi, Youssef MA-37 

May, Angelika WD-43 

Mazauric, Vincent WB-32 

Mäkelä, Marko M. WD-19 

Möst, Dominik WC-33 

Meisel, Frank TB-17, TD-17 

Meissner, Joern WD-38, MA-41 

Melo, Gustavo MA-05 

Mendes, Armando WE-20 

Mersha, Ayalew Getachew MC-42 

Mesquita, Marta MF-16 

Meyer, Christoph Manuel MD-17 

Meyr, Herbert MB-14 

Miettinen, Kaisa MC-07, TD-23 

Miglierina, Enrico WB-44 

Milanic, Martin MD-29 

Milioni, Armando MD-06 

Min, Hokey WD-13 

Mingers, John MA-35 

Mingozzi, Aristide TA-02 

Minner, Stefan WC-38 

Mishra, Nishant MC-41 

Moccia, Luigi WA-40 

Moench, Lars WB-07 

Mohamed, Nurul TC-17 

Mohd Razali, Noraini MA-03 

Molho, Elena TF-44 

Montibeller, Gilberto MF-35, TA-35 

Morabito, Reinaldo MF-34 

Moreno-Pérez, José A. MA-03 

Moretti, Stefano TF-26 

Morton, Alec MA-30, MB-30, MC-30, 

MB-44 

Mottl, Vadim WC-23, WD-23 

Moura, Ana MA-21 

Mousa, Abdelrahim MB-19 

Mues, Christophe WA-20, WC-30 

Muraviev, Roman MB-27 

Muyldermans, Luc TC-04 

N 

Namboothiri, Rajeev MF-17 

Neboian, Andrei MC-33 

Nemeth, Sandor Zoltan MC-09 

Newman, Alexandra TB-21 

Ngueveu, Sandra Ulrich MB-03 

EURO 24 - Lisbon 2010 SESSION CHAIR INDEX 

Nickel, Stefan MD-01 

NoËl, Jean-François WD-37 

Nossack, Jenny TA-08 

Noyan, Nilay TA-28 

O 

Oberti, Pascal WD-37 

Odile, Pourtallier TD-38 

Oguz, Ceyda MD-24 

Oikonomidis, Anastasios TF-10 

Oliveira Soares, João WB-30 

Oliveira, Bruno M.P. M. MD-19 

Oliveira, Jose Fernando TA-01, MA-20 

Onalan, Omer MB-25 

Oplatkova, Zuzana MA-09 

Osman, Ibrahim MD-04 

Ozcan, Ender WC-25 

Ozceylan, Eren WA-10 

Ozkan, Sevgi MF-11, TB-11, TC-11, 

TF-11, WC-11, WD-11 

Ozogur-Akyuz, SureyyaMD-26, MF-26 

Ozturk, Gurkan MA-26, MF-26 

P 

Pacheco, Joaquín WB-15 

Paias, Ana WD-15, MF-16 

Pais, Tiago WE-24 

Paixão, José WB-02 

Pales, Zsolt TD-34 

Paquete, Luis MA-23, MC-23 

Paradi, Joseph TB-06 

Parrado-Hernandez, Emilio MC-26 

Parreira, Telmo TB-19 

Pascoal, Marta TB-23 

Pasichny, Alexis WA-47 

Pavone, Mario MF-24 

Pehlivan, Canan MA-47 

Pekec, Sasa MD-43 

Perea, Federico TA-16, TD-16 

Pereira, Sandrina MB-33 

Pereverza, Kateryna WB-47 

Perez, Gloria MA-28 

Pesch, Erwin TB-08, WC-08, WD-08 

Petrovic, Dobrila TD-28 

Peyghami, Mohammad Reza WE-36 

Pflug, Georg WD-29 

Pham Dinh, Tao MF-45, TA-45 

Pickl, Stefan MC-38 

Pinheiro, Diogo MC-19 

Pinto, Alberto MF-19, TA-19 

Pinto, Alberto A. MA-19 

Pirilä, Pekka TF-33 

Pirlot, Marc TA-30 

Pisinger, David TD-01, MB-15 

Pitsoulis, Leonidas MD-05 

Pla, Lluis Miquel TA-32, WA-32 

Pla-Santamaria, David WD-30 

Plyasunov, Alexander MB-05 

Pohlmann, Tobias TD-27 

Porschen, Stefan TA-29 

Portela, Maria TC-06 

Potra, Florian MC-09 

Potts, Chris MC-08 

Pradenas, Lorena MF-02 

Pralat, Pawel WC-42 

Pratsini, Eleni WE-14 

Prekopa, Andras TF-46 

Prins, Christian MA-15 

Prodhon, Caroline MC-03, MD-03, 

MF-03 

Pruyt, Erik MF-44 

R 

Rainer, Martin TB-38, TD-38 

Ramik, Jaroslav MF-36, TA-36 

Ramos, Ana Luísa WA-27 

Ratprasert, Pasu TF-04 

Rauner, Marion MF-22 

Raupp, Fernanda WD-04 

Rave, Claudia MC-37 

Rönnqvist, Mikael MF-05 

Reston Filho, José Carlos TA-04 

Reuther, Markus MA-16 

Rezapour, Shabnam MF-14 

Ries, Bernard MC-10, TC-10 

Ries, Jana TD-05 

Rincon-Zapatero, Juan Pablo WC-14 

Rivera Agudelo, Juan Carlos TB-04 

Robinson, Stewart TD-35 

Rodríguez-Chía, Antonio Manuel MD- 

13, MF-13 

Romanin-Jacur, Giorgio TB-47 

Rommelfanger, HeinrichMA-36, TB-36 

Rossi, André MB-04, MC-04, MF-04, 

TA-04, WD-05, WE-37 

Rossi, Riccardo WD-27 

Rouwette, Etiënne TC-35 

Ruiz, Ruben WE-33 

S 

Sagir, Mujgan TA-09 

Salhi, Said WB-13 

Salles, André MC-27 

Salman, Sibel TD-13 

Salo, Ahti MA-44 

Sama, Miguel WA-44 

Sami, Nadia TF-30 

Santana, Roberto TA-42 

Santos, José MB-23 

Santos, Nicolau MF-04 

Santos-Peñate, Dolores R. TF-13 

Scaparra, Maria Paola MC-13 

Schellhorn, Henry WD-41, WE-41 

Schmid, Lukas TB-41 

Schulz, Volker WB-40 

Schwaninger, Markus TC-41 

Schwindt, Christoph MA-08 

Sebastian, Hans-Jürgen TC-31 

Seeland, Klaus WC-41 

Sena, Vania WA-06, TB-07 

Sener, Emrah TC-25 

Seow, Hsin-Vonn WD-18, WB-20 

Sevaux, Marc WD-05 

Shakhlevich, Natalia TD-08 

Shikhman, Vladimir MD-46 

Siepak, Marcin MB-04 

Simsek, Koray MA-25 

Sinuany-Stern, Zilla WA-36 

Siskos, Yannis MF-30 

Slikker, Marco WA-26 

Slowinski, Roman TD-29 

Smirlis, Yannis MB-06 

315

SESSION CHAIR INDEX EURO 24 - Lisbon 2010 

Smith, Honora MA-11, MD-37, MF-37 

Sniedovich, Moshe MA-45 

Sofianopoulou, Stella WD-06 

Soler Arnedo, Manuel TA-21 

Soto, Maria WB-05 

Sousa, Joao Miguel da Costa MD-36 

Speranza, M. Grazia WA-02, TA-05, 

TB-05, TF-15 

Spieksma, Frits MF-10 

Spinler, Stefan MC-25 

Spoorendonk, Simon MF-15 

Stefani, Silvana WB-29 

Steinhardt, Claudius TB-43 

Sterna, Malgorzata TF-07 

Sternbeck, Michael WC-45 

Stoyan, Yuri MD-20 

Strijov, Vadim WA-23, WB-23, TB-42, 

TC-42 

Sucha, Premysl WD-07 

Suhl, Leena WA-09 

Szachniuk, Marta MC-24 

Szantai, Tamas WA-46 

T 

Taillard, Eric WC-04 

Tarantilis, Christos TD-15 

Taylan, Pakize WC-17, WE-17 

Tchemisova, Tatiana WB-09, TD-11, 

WA-11, WC-21 

Teghem, Jacques TA-23 

Telhada, Joao TF-24 

Tempelmeier, Horst MA-14 

Tereso, Anabela Pereira MC-12 

Terlaky, Tamas TC-09 

Tomlin, John TB-28 

Topcu, Y. Ilker MD-12 

Toth, Paolo MB-02 

Touati Moungla, Nora MC-48, MD-48 

Trautmann, Norbert MA-08 

Triulzi, Umberto TF-29 

Tseng, Hwai-En MB-05 

Tsoukiàs, Alexis MB-43 

Turan, Gyorgy TC-29 

Turkay, Metin TD-24 

Tuyttens, Daniel WB-04 

316 

U 

Ugur, Ömür TA-38, TB-38, WD-46 

Uhan, Nelson TC-02 

Ukovich, Walter TC-27 

Ulukan, Ziya WA-12 

Uney-Yuksektepe, Fadime MB-26 

V 

Valls, Vicente WA-39 

van den Akker, Marjan TB-16 

Van den Broeck, Dennis TD-10 

van den Heuvel, Wilco WB-38 

Vanden Berghe, Greet WB-24 

Vanhoucke, Mario MF-08 

Vanmaele, Michèle WC-46 

Vícha, Josef MF-36, TA-36 

Veliov, Vladimir MF-39, TA-39 

Ventura, Juan TF-06 

Vespucci, Maria Teresa MB-28 

Vidal, Victor MD-35 

Vidalis, Michael TC-14 

Vieira, Susana MD-36 

Vigna, Elena WB-14 

Villa, Tiziano MF-29 

Vilutiene, Tatjana WC-43 

Vitoriano, Begoña MB-37 

Vla, Katarina MD-28 

Volkovich, Zeev (Vladimir) MC- 

18, MD-18, MF-18, TA-18, 

TB-18, TC-18, TD-18, TF- 

18, WA-18, WB-18 

W 

Walther, Ursula WB-41 

Wanrooy, Gerard TF-05 

Weber, Christoph TD-33 

Weber, Gerhard-Wilhelm MA-01, MA- 

07, MB-07, MC-07, WC-08, 

WD-08, MA-09, TB-09, TF- 

09, WA-09, WB-09, WC- 

09, WD-09, WA-10, WB-10, 

MA-11, TB-11, TC-11, TD- 

11, WA-11, WB-11, WC- 

11, WD-11, WE-17, MD- 

18, MF-18, TA-18, TB-18, 

TC-18, TD-18, TF-18, WA- 

18, WB-18, MA-19, MF- 

19, WC-20, MD-25, MF-25, 

TA-25, TD-25, WB-26, WD- 

26, MA-31, WD-36, WE-36, 

MF-37, MA-38, MB-38, MC- 

38, MD-38, TA-38, TC-38, 

39, WC-40, WD-40, WE-40, 

41, WE-43, WC-44, WD-44, 

WE-44, MA-46, MB-46, MC- 

46, TA-46, MA-47, MF-47, 

TA-47, MA-48, MB-48 

Weber, Richard WE-18 

Weber, Valentin WC-20 

Weglarz, Jan WC-39 

Weintraub, Andrés MB-32 

Wen, Min WA-15 

Wensing, Thomas WA-38 

Wenstøp, Fred TA-31 

Werner, Frank MD-08 

White, Leroy MA-35 

Woeginger, Gerhard J. MB-01 

Wong, Elaine TF-04 

Worthington, Dave TC-22 

Wozabal, David TB-28, WC-29 

Wu, Cheng-Ru WA-42 

Wu, Yue TD-40 

Y 

Yalaoui, Farouk MA-04, TB-04, WA-04 

Yannacopoulos, Denis TD-30 

Yelkenci, Simge WA-04 

Yerlikaya Ozkurt, Fatma WD-17 

Yilmaz, A. Egemen WB-10 

Younesi, Erfan TA-24 

Yu, Yugang WB-45 

Z 

Zabarankin, Michael WE-29 

Zadnik Stirn, Lidija MA-45, MB-45 

Zanoni, Simone MC-34 

Zaraté, Pascale TC-37, TD-37 

Zhao, Huiling MC-22 

Zopounidis, Constantin MD-30 

Šelih, Jana WB-43

Symbols 

, Sonia TD-09, TB-27 

Anghinolfi, Davide WC-25, TC-27 

El Hachemi, ’Nizar TD-15 

Lin, Yadi WA-42 

A 

A.C. Rocha, Ana Maria TB-48 

Aïssani, Djamil WB-18 

Abada, Ibrahim MC-11 

Abascal, Julio MC-26 

Abbink, Erwin TC-16 

Abdel Jawad, Malek MC-25 

Abdelkader, Merakeb TD-39 

Abdelwaheb, Rebai WC-04 

Abdi, M Reza TB-12, TC-41 

Abdul-Jalbar, Beatriz WB-38 

Abe, Jair Minoro MF-26, TD-36 

Abginehchi, Soheil TB-47 

Abramova, Nina MC-35 

Abreu, Rute MC-22, MD-32 

Achtziger, Wolfgang MB-48 

Acikgoz, Bernur WA-06 

Ackermann, Fran TB-01, MB-35, 

MC-35 

Ackermann, Heiner MB-17 

Adacher, Ludovica TB-08 

Adam, Maria TF-29 

Adan, Ivo MF-14 

Adany, Ron MC-18 

Adjiman, Claire TF-35 

Afacan, Gülgün TF-11 

Afanasyeva, Larisa WE-39 

Afonso, Paulo TA-18 

Afsar, H. Murat TD-09, WA-25, WE-25 

Agarwal, Harshita WD-24 

Agarwal, Manju MD-38 

Agarwal, Suraj MF-21 

Aghezzaf, El-Houssaine TC-14, WE-39 

Agnetis, Alessandro TF-27 

Agra, Agostinho WB-22, WE-22, 

TA-40 

Agrawal, Vinu MF-21 

Aguilar Madeira, Jose WC-21 

Aguilera, Jorge WE-37 

Ahlatcioglu, Aykut WE-35 

Ahlgren, Erik TB-37 

Ahlip, Rehez MB-25 

Ahlroth, Lauri WA-25 

Ahmed, Shabbir WB-35 

Ahmed-nacer, MohamedWA-05, TC-18 

Ahn, Heinz MF-06 

Airola, Antti TD-29 

Airoldi, Mara MC-30 

Aissani, Djamil WB-18 

Ak, Ronay TC-36 

Akar, Hanife WD-21, WE-21 

Akartunali, Kerem TB-27 

Akay, Diyar TA-07 

Akca, Yeliz TF-15 

Akca, Zeliha MF-17 

Akinc, Deniz WD-47 

Akkan, Can WE-28 

Aksakal, Erdem MB-12, TD-12, TC-36 

Aksen, Deniz MC-13, TF-15 

Aksoy, Nimet WC-24 

Aktar Demirtas, Ezgi TC-06, WA-06 

Aktas, Emel MA-21, TF-42 

Aktin, Tülin MC-15, TD-20 

Akturk, Gizem MA-31 

Akyüz, Mehmet Hakan TB-13 

Akyuz, Unal MA-31 

Al-Salamah, Muhammad WB-11 

Aladag, Alper MF-04 

Alagador, Diogo TD-32 

Alalouf, Amir WA-36 

Alaoui Selsouli, Marouane WD-39 

Alasehir, Oguzhan TF-11 

Alçada, Luís MB-23 

Albareda Sambola, Maria TC-05 

Albayrak, Y. Esra TC-12 

Albino, Luis TC-16 

Albornoz, Victor MF-32, TA-46 

Albrecht, Thomas MC-16 

Alcan, Pelin MC-04 

Alcantud, José Carlos R. TB-31 

Alcaraz, Javier TF-03 

Alderson, David TC-45 

Alejandro, Boada WC-44 

Alekseeva, Ekaterina MA-42 

Aleluia Reis, Lara MA-33 

Alemany, Mareva MB-36 

Alexandris, Georgios MA-13 

Alfandari, Laurent TA-10, WD-25 

Algaba, Encarnación TA-16, TB-26 

Aliev, Iskander TF-09 

Alirezaee, Mohammad Reza WE-06, 

TC-07 

Alirezaee, Nassrin WE-06 

Alki¸S, Nurcan TF-11 

Allahyar, Maryam MA-06 

Allevi, Elisabetta WA-33, MA-40, 

WA-48 

Allibe, Benoit TA-37, TB-37 

Almada-Lobo, Bernardo MB-34, 

MD-34, TA-34, TB-34 

Almaraz Luengo, Elena WB-17 

Almeder, Christian WB-07, TA-08 

Almeida, Adiel Teixeira de MA-44 

Almeida, Ana MC-23 

Almeida, João TD-19 

Almeida, João Nunes de MD-33 

Almeida, José Paulo MB-23 

Almeida, Leandro MB-19 

Almeida, Maria TD-02 

Almeida, Rui Jorge MA-36 

Almgren, Torgny WD-28 

Alon, Noga MB-01 

Alonso, Maria Teresa MC-20 

Alonso-Ayuso, Antonio TC-05, MA-28, 

TC-48 

Alonso-Meijide, Jose M a WA-26 

Alonso_meijide, José María TC-26 

Alp, Özge Nalan TD-11, MB-12, 

MF-33 

Alparslan Gok, Sirma Zeynep WC-26 

AUTHOR INDEX 

Alper, Doron WB-06 

Alptekinoglu, Aydin TA-43 

Alt, Walter MF-39 

Altan Sakarya, Ay¸se Burcu TD-47 

Altay, Nezih WD-38 

Altekin, F. Tevhide TA-27, 28 

Altin Kayhan, Aysegul MD-40 

Altinel, I. Kuban MA-09, TB-13, 

TC-38, MA-40 

Altinel, Kuban WD-05 

Altinoz, Okkes Tolga WB-10 

Altiok, Tayfur WD-47 

Altunta¸s, Mehmet Yalçın TA-12 

Altuzarra, Alfredo MF-12, TC-37 

Alumur, Sibel A. WE-13 

Alvarez Carrillo, Pavel TB-30 

Alvarez, Ada WB-15 

Alvarez-Valdes, Ramon MC-20 

Alvelos, Filipe WC-05, TA-20, MA-32, 

MB-34, MB-40 

Alvelos, Helena MF-19, TA-44 

Alves, Andre MD-34 

Alves, Cláudio MC-02, TA-02, TB-02, 

WE-15, MA-20 

Alves, Maria João MC-42 

Alzate, Carlos TA-42 

Amaral, Andre MC-43 

Amaral, Paula TC-43 

Amaro, Ana Cristina Santos MC-14 

Amatatsu, Hirofumi MA-06, WD-06 

Amaya, Ciro Alberto MA-15 

Amberg, Bastian MD-16 

Amberg, Boris TB-16 

Ambrosino, Daniela TC-27 

Amelia, Bilbao-Terol MC-36 

Amen, Prof. Dr. Matthias WA-29 

Amodeo, Lionel MF-03, MA-04, TB- 

04, WA-04, WC-04, MC-05 

Amores, Antonio F. MF-06 

Amorim, Filipa TC-33 

Amorim, Pedro MD-34 

Anatoliy, Dorovskykh MC-18 

Andersen, Sigrid de Mendonca WA-12 

Anderson, Edward WB-33 

Andersson Granberg, Tobias MB-46 

Andersson, Henrik WB-22, WC-22, 

WE-22, TC-38 

Andrade, Fernando do Valle SilvaTC-06 

Andrade, Jose luis TD-22 

Andrade, Marinho Gomes WE-41 

Andrade-Campos, António MC-20 

Andradottir, Sigrun TB-47 

André, Vitor WA-43 

Andreeva, Galina TA-38 

Andronikidis, Andreas WA-12 

Androulakis, George S. TD-48 

Androutsopoulos, Konstantinos WB-12, 

MC-48 

Angel-Bello, Francisco R. WB-15 

Angelelli, Enrico WC-29 

Angelis, Vassilis MB-13, TD-14, 

WD-18, MB-25 

Angelis-Dimakis, Athanasios MB-33 

Angulo-Meza, Lidia MD-06, TC-06 

317

AUTHOR INDEX EURO 24 - Lisbon 2010 

Angun, Ebru Mevlude MD-28 

Anjos, Miguel MB-29, TB-29 

Anouze, Abdel Latef TF-06 

Antczak, Maciej MB-24 

Anton, Bjoern WE-17 

Antunes de Souza, Rosely MF-18 

Antunes, Antonio MF-45, TF-47 

Antunes, Carlos Henggeler MA-06, 

TC-33 

Anvar, Seyed Hesameddin MC-12 

Apanaviciene, Rasa WB-43 

Aparicio, Juan TB-26 

Aparisi, Francisco MD-05 

Apaydin, Aysen MD-36 

Aponte, Susanne TF-17 

Appa, Gautam MF-10, TC-10, WE-35 

Araújo, Miguel TD-32 

Araújo, Olinto MC-20 

Aragones-Beltran, Pablo TB-12 

Arampatzis, George MB-33 

Arana-Jiménez, Manuel TF-44 

Aranda Almansa, Joaquin MB-20 

Araneda-Fuentes, Cristina TB-43 

Arantes, Amílcar TB-22, TF-22 

Arapoglu, R. Aykut TB-03 

Aras, Necati WD-05, MA-09, MC-13, 

TD-13, TC-38, MA-40 

Araujo, Silvio MA-34 

Arbelaitz, Olatz MC-26 

Archetti, Claudia TD-15 

Archetti, Francesco TC-24 

Archibald, Thomas MF-18 

Arenales, Marcos TB-32 

Arenas-Parra, Mar WD-30, MC-36 

Argaez, Miguel WB-21 

Argyris, Nikolaos MC-30, MB-44 

Arian, Amin WD-12 

Arias, Alicia WC-06 

Arias, Marta TC-29 

Aringhieri, Roberto TA-22 

Arkhypova, Elena WC-11 

Armentano, Vinícius MC-20, TD-21 

Armstrong, Stanislava MA-45 

Arrar, Nawel WB-18 

Arroyo, José Manuel MB-33 

Arslan, Alp Muzaffer TC-12 

Arslan, Özge WD-36 

Artalejo, Jesus TF-05, WA-46 

Artigues, Christian WA-07, 08, WA-25, 

WE-25 

Arulselvan, Ashwin TC-10 

Arutyunov, Aram TC-39 

Aryanezhad, M.b. MB-04, WC-04, 

MB-34, TA-34 

Ascó, Amadeo MC-08 

Aspergh, Giada TB-47 

Asproth, Viveca TC-44 

Assoumou, Edi TA-37, TB-37 

Astorino, Annabella WB-19 

Aswal, Abhilasha TB-18, MF-26 

Atamer, Busra TA-14 

Atasever, Ilknur WC-34 

Atbas, Serap WC-29 

Athanasiou, Efstratios MC-14 

Atkin, Jason MC-08, MA-45 

Atmaca, Ediz TA-09 

318 

Attia, Ahmed MF-33 

Attouch, Hedy MA-07 

Aubry, Alexis WE-37 

Avcikurt, Cevdet MF-37 

Avelino, Catarina MF-47 

Averaimo, Pietro TD-17 

Avila, Antonio Flavio TB-32 

Avkiran, Necmi WC-06 

Avros, Renata MB-18, MC-18 

Ayanoglu, Can MA-47 

Aydin, Mustafa TF-11 

Aydin, Nursen MB-45 

Ayhan, Hayriye TB-47 

Aykanat, Pınar MB-11 

Aytekin, Alper TC-42 

Azad, Md. Abul Kalam TB-48 

Azar, Adel WA-10 

Azcarate, Cristina WA-36 

Azevedo, Nuno MC-19 

Azi, Nabila MC-17 

Aziz, Azmin Azliza WB-15 

Aziz, Haris TC-02 

Azizi, Majid WA-12, WD-12 

Azizoðlu, Meral WB-04, TD-48 

Azzi, Anna MD-04 

Álvarez-López, Alberto A. MD-19 

Álvarez-Mozos, Mikel WA-26 

Çetin, Yasemin TF-11 

Çoban, Özlem WB-26 

Öberg, Lena-Maria MB-11, TC-44 

Önal, Hayri TA-19 

Öncan, Temel TB-03, TB-13 

Önder, Halil TD-47 

Özbey, Halil WD-40 

Özdemir, Rifat Gürcan WE-08, TD-20 

Özden, Eda WA-45 

Özgürler, Mesut MA-07 

Özkut, Cemal Murat MB-05 

Özlen, Melih TD-48 

Özmutlu, H. Cenk TC-36 

Üstün, Abdullah Korkut TC-06, WA-06 

Üstünkar, Gürkan MB-26, MD-26 

özkan, özlem MA-37 

B 

Baños, Alfonso TA-48 

Babaei, Salman WE-24 

Babic, Zoran TA-18 

Babington, Sally MB-35 

Bacellar, André MA-31 

Backåker, Lars TA-27 

Badillo Piña, Isaias WB-39 

Baek, Jun-Geol WA-11 

Baes, Michel WB-34 

Baesens, Bart WB-20, WC-30 

Baesler, Felipe TB-04, TA-14 

Bagheri, Behzad WB-46 

Bagirov, Adil WB-19, WD-19 

Bahalke, U. WC-04, WE-40 

Bahiense, LauraMB-02, MA-11, TB-48 

Bahn, Olivier WB-32 

Bahramgiri, Mohsen TD-25, WE-41 

Baidya, Tara Keshar Nanda WB-06 

Baier, Robert MB-39, MF-39 

Bajak, Szabolcs TF-34 

Bakal, Ismail Serdar TA-14 

Bakhrankova, Krystsina TA-08 

Bakken, Bent Erik TB-31, MD-35 

Bako, Andras WA-46 

Bakoulas, Konstantinos TA-06 

Balas, Egon MB-02, TA-02, WB-21 

Baldacci, Roberto TA-02, MD-15, 

MF-15 

Ballestin, Francisco WA-39 

Bana e Costa, Carlos TD-30, TD-37, 

TC-45 

Banal-Estanol, Albert TB-46 

Bang-Jensen, Joergen TB-20 

Baniasadi, Pouya TA-29 

Bürgisser, Michael WB-34 

Bürgy, Reinhard MD-04 

Büskens, Christof MD-09 

Baptista, Elesandro WD-08 

Baptista, Susana MA-46 

Baptiste, Hervé MD-48 

Baraçlı, Hayri TD-11, MB-12 

Barbagallo, Annamaria WA-48 

Barbarosoglu, Gülay TD-13 

Barbósa-Póvoa, Ana Paula MB- 

04, WA-13, MC-14, TA-15, 

TC-17, WC-18, TB-33, MD- 

47, TB-47, WA-47 

Barbosa, Edna WB-28 

Barcelo, Jaume WA-27 

Bareche, Aïcha WB-18 

Barfod, Michael Bruhn TC-37 

Barkana, Buket MF-26 

Barnett Lawton, Jorge MC-28 

Barreiros, Artur WA-09 

Barrena, Eva MD-13 

Barrieu, Pauline WD-33 

Barros, Ana Isabel TC-45, TD-45 

Barros, Elisa WE-21 

Barros, Oscar WE-18 

Barth, Dominique TB-18 

Bartl, David WC-26 

Barzily, Zeev MB-18 

Bashiri, Mahdi TD-05 

Basligil, Huseyin MC-04, WC-40 

Bastani, Kaveh TD-22 

Bastert, Oliver WB-36 

Bastic, Majda MA-31 

Bastos, João MA-19 

Basu, Sumanta WD-04 

Battarra, Maria TA-15 

Batur, Gul Didem WA-28 

Baudach, Jens WA-31 

Bauman, Evgeny WC-23, WD-23, 

TB-25 

Baumann, Philipp MA-08 

Baumgartner, Alfonzo WD-28 

Bayindir, Z. Pelin TA-14 

Baykal, Nazife MB-26 

Bayram, Armagan WD-44 

Bayramoglu, M.Fatih TB-25, TC-42 

Börjesson, Martin TB-37 

Ba¸so˘glu, ˙Ismail WD-41 

Beasley, J. E. MD-20 

Beaudoin, Daniel MB-32 

Bebcakova, Iveta MF-36 

Becceneri, José Carlos WC-05 

Becejski-Vujaklija, Dragana WE-18

Becker, Denis MC-28 

Bednarczuk, Ewa WB-44 

Begen, Mehmet MB-22, WA-24 

Begicevic, Nina WD-07 

Behret, Hülya MB-12, TD-36 

Behzadi, Mohammad Hassan MB-06, 

MD-06 

Bekesas, Luiz WE-40 

Bekesi, Jozsef TC-04 

Bektas, Tolga MC-15, WE-35 

Beldek, Ulas WE-17 

Belderrain, Mischel Carmen N. MC-35, 

MB-44 

Belen, Selma WD-44, MC-46 

Belenguer, Jose M. MA-15 

Beliën, Jeroen WA-24 

Bell, Peter MB-22, MF-43 

Bellenguez-Morineau, Odile WE-28 

Bellini, Fabio MF-19 

Bello, Lenys TA-13 

Bello-Dambatta, Aisha MC-47 

Belmecheri, Farah MF-03 

Belotti, Pietro WA-35 

Belov, Gleb MF-20, TB-20 

Belton, Valerie MA-30, MF-35 

Beltran-Royo, Cesar TD-02, MD-28 

Ben Amor, Sarah TF-37 

Benabid, Abir WB-08 

Benavent, Enrique MA-15 

Bencivenga, Cristina WB-29 

Benedito, Ernest MC-34 

Beneki, Christina TF-29 

Benfkir, Rahim TD-41 

Benhamiche, Amal TA-10 

Benhizia, Faten TC-16 

Benito, Antonio WD-30 

Benke, Philipp MA-08 

Bennell, Julia MC-08, WC-16 

Benoist, Thierry WE-05 

Beraldi, Patrizia WE-30 

Berbig, Dominik WB-45 

Berdugo-Kushnir, Vered WB-06 

Berenguel, Jose Luis TC-43 

Beresnev, Vladimir TF-03 

Bergantinos, Gustavo TA-26 

Berger, Rosemary MF-17 

Bergman, Bo TD-22 

Berk, Erhan TA-07 

Berlinska, Joanna TF-07 

Berman, Oded TC-13 

Bermúdez, José D. TF-21, TD-25 

Bernardini, Annalia MA-12 

Berning, Bettina TD-31 

Berrachedi, Abdelhafid WA-18 

Berro, Alain MD-05 

Berteloot, Koen WC-30 

Berthold, Timo TA-08 

Bertocchi, Marida MB-28, MA-40 

Bertolazzi, Paola TC-24 

Bertsimas, Dimitris WE-05 

Best, Michael MD-27 

Bettinelli, Andrea TC-17 

Beullens, Patrick TD-05, TD-46 

Beyersdorff, Olaf TA-29 

Bharadwaj, Vijay TB-28 

Bhatia, Deepak MA-14 

EURO 24 - Lisbon 2010 AUTHOR INDEX 

Bhattacharjya, Debarun WE-14 

Bianchessi, Nicola TD-15 

Bianzino, Aruna Prem TF-26 

Bieksa, Darius WC-43 

Bielza, Concha TA-42 

Bierwirth, Christian TB-17 

Bifulco, Gennaro Nicola WD-27 

Biggiero, Luigi WA-27 

Bijak, Katarzyna TA-38 

Bilbao, Celia WD-30 

Bilbao, Jesus-Mario TB-26 

Bilegan, Ioana MA-41 

Bilgin, Burak WB-24 

Billaut, Jean-Charles TF-05 

Bilmes, Jeff TD-02 

Binder, Anne MC-16 

Biniecki, Wojciech MB-24 

Birattari, Mauro TD-05 

Birbil, S. Ilker WC-05, WC-24, MB-45, 

MC-48 

Birgin, Ernesto G. TF-21 

Bisdorff, Raymond TA-30 

Bjørndal, Mette TA-33 

Blanco, Victor WE-36 

Blander Reinhardt, Line MD-15 

Blank, Frederik TD-47 

Blanquero, Rafael TA-13 

Blaszczynski, Jerzy TD-29 

Blazewicz, Jacek MA-24, MB-24, 

MD-24 

Blesl, Markus TB-37 

Blockmans, Tom WD-26 

Bloemen, Axel TD-45 

Bloos, Melanie MA-17 

Bobek, Samo TD-44 

Bock, Stefan TD-07, TC-08, TC-15 

Bodnar, Taras WD-29 

Boggia, Antonio WA-37 

Bogojevic Arsic, Vesna TD-25 

Boile, Maria WA-40 

Boland, John WC-40 

Boland, Natashia TB-27 

Bolat, Bersam TB-36 

Bollapragada, Ramesh MA-14 

Bomze, Immanuel MA-02, TC-43, 

TD-43 

Bonachela Capdevila, Francisco MA-24 

Bonami, Pierre WB-35 

Bonandi, Marcelo WD-40 

Bonassi, Stefano TF-26 

Bonates, Tiberius MC-02 

Bonato, Anthony WE-42 

Bonnard, Nicolas TA-37 

Bonneuil, Noel TC-39 

Bonomo, Flavia MD-29 

Bont, Leo MB-32 

Boogert, Alexander TD-33 

Borges, Diogo TC-06 

Borges, Jorge MC-33 

Borges, Jose MA-32, MC-32, MD-32, 

TC-32 

Borges, Paulo MC-32, TC-32 

Borglund, Erik TC-44 

Borgwardt, Steffen TB-45 

Borndörfer, RalfMB-16, TF-16, WB-16 

Borne, Sylvie WD-15 

Borodin, Dmitry WC-12 

Boros, Endre TC-29 

Bortfeldt, Andreas MB-20, MA-21 

Bosch, Maximo WC-45 

Boschian, Valentina TC-27 

Bose, Dipankar WB-46 

Bossens, Frédéric WC-46 

Bossley, Kevin MC-44 

Bosso de Freitas, Gisele TD-41 

Bossy, Mireille TD-38 

Bot, Radu Ioan WD-09 

Botequim, Brigite MC-32, TC-32 

Botsaris, Charalampos TF-29 

Botter, Rui Carlos TD-40 

Botton, Quentin MB-40 

Boualem, Mohamed WB-18 

Bouasker, Olfa MC-07 

Boucekkine, Raouf MA-39, MD-39 

Boucherie, Richard MB-30 

Boudhar, Mourad WD-28 

Boudjemaa, Redouane WB-09 

Boukredera, Djamila WA-21 

Bouroubi, Sadek MB-07 

Bourreau, Eric WA-25, WE-25 

Boutiche, Mohamed Amine WA-18 

Bouvry, Pascal MB-23, MD-45 

Bouyssou, Denis TA-30 

Boyaci, Tamer MD-14 

Boysen, Nils MB-08, WC-36 

Bozdag, Cafer Erhan TC-36 

Bozic, Caslav WA-23 

Bozkaya, Burcin MA-21 

Bradl, Peter TB-41 

Braekers, Kris TC-03 

Brahou, Vassiliki WA-36 

Branchini, Rodrigo TD-21 

Branda, Martin MF-28 

Brandao, Jose TB-03 

Brandner, Hubertus WA-20 

Brandstädt, Andreas MA-10 

Brandt, Felix TC-02 

Branzei, Mariana Rodica TB-26, 

WC-26 

Branzei, Oana WC-26 

Brauner, NadiaWD-05, MD-10, WC-20 

Bravo, Mila WD-30 

Brás, Pedro MA-20 

Brás, Raul TD-32, MC-46 

Bredström, David WB-31 

Breiter, Andreas WD-45 

Breschan, Jochen MA-21 

Bretschneider, Sarah TF-27 

Bretto, Alain TA-10 

Brezina, Ivan MC-03, WB-13 

Briand, Cyril WE-37 

Briat, Vincent MC-11 

Brieden, Andreas TB-45 

Brimberg, Jack WB-13 

Briskorn, Dirk TD-07, TC-08 

Brodnik, Andrej TC-04 

Broekmeulen, Rob WC-45 

Brotcorne, Luce MA-41 

Brown, David MF-25 

Brown, Iain WC-30 

Brugha, Cathal WB-30, MD-31, TA-31 

Bruglieri, Maurizio WB-27 

319


Bruneel, Herwig MD-18 

Brunelli, Ricardo TB-32 

Brunner, Jens TB-22 

Bruno, Giuseppe TD-17 

Bruns, Florian WC-36 

Brutti, Pierpaolo TD-26 

Bucheli Guerrero, Víctor MD-37 

Bucher, David WD-38 

Buchheim, Christoph WC-35 

Buer, Tobias TF-17 

Buhrkal, Katja TB-17 

Bulajic, Milica TA-25 

Bulavsky, Vladimir MC-09 

Bulbul, KeremMD-04, WC-05, WC-24, 

WB-28 

Bulinskaya, Ekaterina WE-39 

Buljubasic, Mirsad WD-25 

Burachik, Regina TF-23 

Burai, Pál TC-34 

Buratto, Alessandra MD-39 

Burinskien, Marija WA-43 

Burke, Edmund MF-02, WE-05, MC- 

08, MD-24, WE-24, MA-45 

Burkhardt, Thomas MA-38, WA-41 

Burlando, Roberto TA-31 

Burnaz, Sebnem MD-12 

Burt, Christina MA-40 

Buscaylet, Fabrice TC-21 

Buscher, Udo WB-07 

Bushenkov, Vladimir TF-32 

Bussieck, Michael MC-21, WE-35 

Bustos, Mercedes MC-37 

Buyukozkan, Gulcin MD-36 

Bykadorov, Igor TA-47 

C 

C. B. Camargo, Victor TB-34 

Cañal, Verónica WD-30, MC-36 

Caballero, Rafael TD-04 

Cabello, Jose Manuel TD-23 

Cabeza, Mar TD-32, MC-46 

Caccetta, Louis TA-36 

Cacchiani, Valentina MB-16 

Cadarso, Luis MA-16 

Cadima, Jorge MC-32 

Cadoux, Florent TA-37 

Cafieri, Sonia WA-35, WB-35 

Caglar, Burcu TC-36 

Caimi, Gabrio Curzio WC-16 

Cakir, Elif MA-37 

Cakmak, Emre WA-38, 39 

Calabuig-Tormo, Carola TB-12 

Caldeira, Pedro WB-24 

Caldwell, Vanessa MB-35 

Calheiros de Melo Junior, Aurelio MA- 

31 

Call, Mikael MC-40 

Calvas, Baltazar WC-41 

Calvete, Herminia I. MA-42 

Calvillo, Gilberto MC-02 

Camacho, Carmen MA-39 

Camanho, Ana TC-06, TF-06, WA-06, 

WD-06 

Cambazard, Hadrien WD-25 

Cambini, Riccardo TC-34 

Camilleri, Guy MD-43 


Campi, Marco WA-21, MD-46 

Campos, Clara M. TF-13 

Camus, Roberto WD-27 

Can, Anıl TA-08 

Canales, Cristian TA-46 

Canca, David MC-06, TA-16, MF-21 

Cancer, Vesna MD-35 

Canepa, Alessandra MC-25 

Caner, Serra WA-11 

Cangalovic, Mirjana TC-10 

Canos, Lourdes MB-36, MC-36 

Canos, Maria J. MB-36, MC-36 

Caprara, Alberto MB-16, WC-35 

Captivo, Maria Eugénia TC-23, TD-28 

Carè, Algo WA-21 

Cardeneo, Andreas WC-08, MD-17 

Cardoen, Brecht MF-05, WA-24, 

MC-30 

Cardonha, Carlos WB-16 

Cardoso, Domingos MB-07 

Cardoso, Gonçalo TB-33 

Cardoso, Margarida TF-18 

Cardoso, Margarida F. WD-21 

Cardoso, Sónia TA-15 

Caris, An TC-03 

Carlier, Mathilde TC-08 

Carmo, José Luís TC-09 

Carneiro Brandão, Luana TC-06 

Carnero, Maria Carmen TD-30 

Carolina, Saldana WC-44 

Carosi, Laura TC-34 

Carotenuto, Pasquale WC-27 

Carpente, M a Luisa TC-26 

Carpentier, Pierre MC-39 

Carrasco-Gallego, Ruth TA-27 

Carravilla, Antonia MB-22 

Carreras, Ashley MA-35 

Carreras, Francesc TC-26 

Carrión, Miguel WC-33 

Carrizosa, Emilio TA-13 

Carroll, Paula MD-40 

Carvalho, Ana WA-47 

Carvalho, Filipa TD-02 

Carvalho, Joel MA-13, TF-15 

Carvalho, Marco TD-20 

Carvalho, Pedro TD-06, WB-06 

Carvalho, Sameiro MD-14, MD-30, 

MA-32 

Carvalho, Silvia WC-40 

Casado, Leocadio G. TC-43 

Casas-Méndez, Balbina-Virginia TC-26 

Caserta, Marco WA-20 

Casier, Aurélie MF-41 

Cassioli, Andrea TA-48 

Castelli, Mauro TC-24 

Castermans, Gerd WC-30 

Castro Lobo, Maria Stella MF-06, 

WD-06 

Castro, Jordi WC-19 

Castro, Pedro MF-20, TF-35 

Castro, Santiago TA-35, MC-44 

Catala, Andreu MC-26 

Catalão-Lopes, Margarida WB-30, 

MC-38 

Catanzaro, Daniele MC-02, MB-40 

Catibusic, Faik WD-25 

Cattier, Francois TA-37 

Caulkins, Jonathan TC-39 

Caurla, Sylvain WC-32 

Cavique, Luís WE-20, TD-21 

Cavuslar, Gizem TB-03 

Cayirli, Tugba TA-22 

Cayla, Jean-Michel TA-37 

Cáceres, M a Teresa MD-13 

Cárcaba, Ana TF-06 

Celeste, Francis TA-45 

Cello, Marco MF-40 

Cempirek, Vaclav WB-45 

Cepek, Ondrej TC-29 

Cerdeira, J. Orestes MB-07, TD-32, 

MC-46 

Cerveira, Adelaide TA-40 

Cervera, Joaquín TA-48 

Cesar, Mihael TD-44 

Cesaret, Bahriye MA-04 

Ceselli, Alberto TC-17 

Cetiner, Demet MF-43 

Cevizci, Dicle WD-44 

Chabchoub, Habib TD-03, MA-37 

Chabrier, Alain WE-14 

Chacón, Casi WA-26 

Chagas Júnior, Milton MB-44 

Chagas, Nilo MF-06 

Chaichiratikul, Pairoj MF-03 

Chalimourda, Christina MB-13 

Champion, Donna TB-35 

Chan, C.y. WD-15 

Chan, Yao-ban MA-40 

Chancelier, Jean-Philippe MC-39 

Chang, Che-Wei WC-11, WA-42 

Chang, Chingpou WB-11 

Chang, Pao-Long TD-11 

Chang, Yang-Chi TA-46 

Chao, Po-Liang WB-10, TD-11 

Chao, Xiuli WB-17 

Chaojun, Xu TB-21 

Chaouachi Siala, Jouhaina WB-05 

Chardy, Matthieu WA-31 

Chartres, Sophie TD-46 

Chatterjee, Ashis WB-46 

Chattopadhyay, Asis Kumar MD-11 

Chaudhry, Sohail TC-18 

Chauvet, Fabrice TB-19 

Chaves, Antonio WD-04 

Chébre, Mériam WE-37 

Chehade, Hicham WC-04, MC-05 

Chellathurai, Thamayanthi MC-27 

Chemla, Daniel WD-25 

Chen, Chie-Bein TD-18 

Chen, Chie-bein WB-10, MA-11, 

TD-11 

Chen, Chien-Ming TD-14 

Chen, Chiu-Chin TC-11, WC-11 

Chen, Frank Y WC-18 

Chen, Huey-Kuo MA-46 

Chen, Jih-An WB-11 

Chen, Li-Fei WD-47 

Chen, Li-Je MA-11 

Chen, Ming-Hui WB-10 

Chen, Sheu-hua TB-04, MA-06 

Chen, Si MB-40 

Chen, Yan-Cheng WD-47

Chen, Yi-Chun TC-12, TF-12 

Chen, Yi-Kang TB-06 

Chen, Yi-Shan TA-25 

Chen, Yuang-sung TB-06 

Chen, Yuh-Liang TF-18 

Chernov, Nikolai MA-20 

Cheung, C.f. WD-15 

Cheung, Kwok TD-38 

Chevalier, Alain WB-30 

Chew, Ek Peng WE-24, TF-40 

Chiang, Hsin-Yu WC-12 

Chiarella, Carl WD-46 

Chibeles-Martins, Nelson MB-04 

Chien, Chin Fang TB-07 

Chien, Yuyao TA-20 

Chieng, Ming-Hua TD-11 

Chimani, Markus MA-02 

Chinnam, Ratna Babu TA-14, MB-15 

Chiu, Ya-Ling WD-11 

Chiusolo, Katia WA-13 

Chmielewski, Mariusz WA-30 

Cho, Wei Ting WB-10, TD-18 

Choi, In-Chan TC-18 

Choi, Sungho MF-42 

Chokri, Slim TB-36 

Cholette, Susan MB-41 

Chou, Ying-Chyi WE-38 

Christiansen, Marielle WB-22, WC-22, 

WE-22 

Christophel, Philipp WB-36 

Chryssoverghi, Ion MF-39 

Chu, Feng TF-13 

Chu, Hsiao-Wen WE-38 

Chubanov, Sergei WD-36 

Chun, Young WA-11 

Chung, Cheng-Chi MB-12 

Church, Richard MA-21, MB-32 

Churilov, Leonid TA-41 

Ciavotta, Michele MC-04 

Cicková, Zuzana MC-03, WB-13 

Ciftci, Mehmet Tahir MA-26 

Cihan, Ahmet MA-03, TF-25 

Cinicioglu, Esma Nur WE-38 

Cinko, Murat WD-47 

Cipriani, Ernesto WA-16, WB-27 

Cismondi, Federico MD-36 

Clark, Alistair TA-34 

Clarke, Nancy WD-42 

Claro, João WA-05 

Clasen, Christian WA-41 

Clausen, Uwe TB-21 

Clautiaux, François TB-02, MA-20, 

TB-20 

Clímaco, João TC-23, TB-31, TB-35 

Clemente, Mónica MB-38 

Clewlow, Les WD-46 

Cobuloglu, Halil TB-33 

Cochran, James TA-01, TD-10 

Codina, Esteve MA-16 

Coelho, Antönio TF-17, MF-18 

Coelho, Mayk TF-48 

Coelho, Ormeu TB-48 

Coelho, Paulo MA-46 

Coffrin, Carleton MC-48 

Cohen, Guy MC-39, TA-47 

Cojocaru, Monica-Gabriela MB-42 


Colaco, Susana MA-22 

Colebrook, Marcos WB-38 

Coletsos, John MF-39, WC-44 

Colicchia, Riccardo WA-13 

Collazo, Rodrigo MA-11 

Collier, David MC-30, TF-30 

Colmorn, Richard TF-17, WB-27 

Colombo, Marco MC-21 

Colomer, M. Angels MF-24 

Colorni, Alberto WB-27, MA-43, 

MB-43 

Comas Marti, Joana M. MB-35 

Condotta, Alessandro TD-08, WC-28 

Conejo, Antonio J. WC-33 

Conforti, Domenico TD-28 

Consigli, Giorgio WE-30 

Constans, Sophie WE-36 

Constantino, Miguel TC-22, MC-32, 

MD-32 

Contardo, Claudio TD-16 

Contreras, Ivan WE-13 

Cook, Wade MC-06 

Coppini, Nivaldo WD-08, WD-40, 

WE-40 

Corberan, Ana TF-21 

Corberan, Angel TB-15 

Corbett, Charles TD-14 

Corchero, Cristina MB-28 

Cordeau, Jean-FrançoisWE-13, WA-15, 

WB-22, TB-23 

Cordeau, Jean-Francois TA-15 

Cordes, Philip TF-17, WB-27 

Cordier, Jean-Philippe MF-22 

Corman, Francesco WC-16 

Cornaz, Denis MB-10, TB-10 

Corominas, Albert MC-34 

Correia, Aldina WE-19 

Correia, Isabel MB-13, WE-13 

Correia, Telmo MF-19 

Cortés, Cristián WE-15 

Cortina, Carla WA-37 

Cosgun, Ozlem WD-45 

Costa, Alvaro MC-27 

Costa, Alysson TB-32 

Costa, Ana MA-46 

Costa, Anabela MA-44 

Costa, Anildo MB-33 

Costa, Lino TA-04, TB-48, TC-48 

Costa, Marie-Christine MD-10 

Costa, Ruy MD-22 

Cotret, Julien MC-31 

Couceiro, Miguel TC-29 

Coudert, David WD-42 

Coutinho-Rodrigues, João MB-23 

Couto, Francisco MC-24 

Covas, Miguel TD-37 

Cozzini, Alberto WD-23 

Crainic, Teodor Gabriel MF-03, TA-15, 

MF-17, TA-17, TB-19 

Creemers, Stefan WA-07 

Crema, Alejandro WE-36 

Crespi, Giovanni Paolo TF-34 

Crespo Abril, Fortunato TD-32 

Crespo, Olivier TF-47 

Cristescu, Gabriela TD-34 

Cruz, José MB-19 

Cruz, Jose MD-44, WA-48 

Csetnek, Ernö Robert WD-09 

Cui, Jian TB-28 

Culik, Miroslav TD-33 

Cunha, Alexandre MB-02 

Cunha, João MB-23 

Cunha, Maria MA-46, TF-47 

Custodio, Ana Luisa WC-21 

Cutello, Vincenzo MF-24 

Cywicka, Dominika TA-32 

Czarnecki, Marco MA-07 

D 

D’Acierno, Luca WB-16, WA-17 

D’Alpaos, Chiara WB-37, MD-39, 

MB-43 

D’Ambrosio, Claudia MF-05 

D’Amours, Sophie MA-32, MB-32 

D’Ariano, Andrea WC-16 

D’Ecclesia, Rita TF-29, WB-29 

da Silva Neto, Mikey MF-42 

Dagdeviren, Metin MB-12, TD-12, 

TC-36 

Dagliyan, Onur TD-24 

Dai, Bo MB-17 

Dalby, Paul TF-35 

Dall’Aglio, Marco TD-26 

Damart, Sébastien WB-37 

Dambreville, Frédéric TA-45 

Daniele, Patrizia WB-48 

Dück, Viktor MD-16 

Daou Pulido, Amir José MF-33 

Daoud, Slim WC-04 

Dargam, Fatima TD-37 

Darlay, Julien MD-10, WD-25 

Darouich, Hanaa TF-32 

Darvish, Parviz TC-33 

Daryina, Anna WA-34 

Das, Shubhabrata WD-43 

Dash, Gordon TF-21 

Date, Paresh MC-25, WD-46 

Datta, Dilip MF-23 

Daugeliene, Ala WB-43 

Dauzere-peres, Stéphane TC-16 

David, Balazs TC-04 

David, Fátima MD-32 

David, Israel WA-36 

Davidau, Olivier TD-38 

Davison, Matt MD-27 

Davoodi, Alireza MB-06 

Davtalab Olyaie, Mostafa TF-09 

Dérian, Nicolas MB-24 

Díaz Díaz, José Carlos WB-25 

Döyen, Alper TD-13 

De Almeida, David TC-16 

de Armas, Jesica MF-20 

De Asmundis, Roberta MC-24 

De Baets, Bernard TD-04, TD-29 

de Brock, Bert TB-18, MB-30 

De Brucker, Klaas MC-31 

de Carvalho, J. M. Valério MC-02, 

TA-02, TB-02, WC-05, WE- 

15, MA-20, TA-20, TD-20, 

21, MB-34 

De Causmaecker, PatrickMA-24, TF-28 

De Feyter, Tim MD-11 

321


De Francesco, Juan Pablo MF-06 

De Giorgi, Enrico MF-25 

De Giovanni, Pietro MC-34 

de Jong, Cyriel TD-33 

de Kok, Ton MB-15, MC-15 

de Koster, René TF-40, WA-45 

de la Croix, David MA-39 

de La Fortelle, Arnaud TA-47 

de Lange, Gertjan MD-22, WE-46 

De Leone, Renato MF-42 

De Maria, Elisabetta TA-24 

de Martinis, Valerio WA-27 

De Moor, Bart MA-24 

De Pascale, Gianluca TF-27 

de Regt, Dorien TC-22 

De Reyck, Bert MA-34 

de Rooij, Frans MD-21, WE-46 

De Santis, Marianna TA-48 

De Scheemaekere, Xavier WD-14 

De Schepper, Steven TA-44 

De Smet, Yves MF-23, MD-30, TA-30, 

TC-30 

de Smidt-Destombes, Karin TC-45 

de Sousa, Amaro WC-05, MB-40, 

MC-40 

de Souza, Cid MB-02 

de Vicente, Javier TD-32 

de Werra, Dominique MD-10 

Dedoussis, Vassilis WD-06 

Deelstra, Griselda WC-46 

Degerli, Mustafa TF-11 

Degraeve, Zeger MA-34 

Dehghan Krooki, Rafat WA-12 

Dejaeger, Karel WB-20 

Dekker, Rommert MC-01, WD-38, 

TF-40 

Del Rio Vilas, Victor MF-35 

Delacote, Philippe WC-32 

Deleris, Lea WE-14 

Delgado Pineda, Miguel MB-20 

Delgado, Alexandrino Duarte WB-22 

Delibasic, Boris TF-18 

Dell’Amico, Mauro WB-25 

Dell’Olmo, Paolo WA-13 

Dell’Orco, Mauro WC-27 

Dell, Robert TC-45 

Della Croce, Federico MC-43 

Dellaert, Nico MB-15 

Demange, Marc TA-10 

Demeulemeester, Erik WA-07, WA-24 

Demir, Aysegul TA-14 

Demir, Leyla WA-04 

Demirba¸s, Korkut TD-47 

Demirta¸s, Kerem WB-04 

Demirta¸s, Nurgül TD-11, MB-12 

Dempe, Stephan MC-42, MD-42 

Dempsey, James TC-22 

Denault, Michel WC-33 

Denizel, Meltem MC-41 

Densham, Greta S. WE-22 

Dereniowski, Dariusz WB-39 

Dereu, Guillaume WA-25, WE-25 

Derpich, Ivan WB-21 

Desaulniers, Guy WB-22, WC-22, 

WC-24, 25, WB-31 

Deshmukh, Abhijit MC-41 

322 

Despic, Ozren TA-06 

Despotis, Dimitris MB-06 

Detienne, Boris MF-02 

DeTombe, Dorien MB-31 

Devolder, Olivier WB-40 

Dhillon, Inderjit WD-19, WA-34 

Dhouib, Emna TA-23 

Di Francesco, Massimo TA-15 

Di Giacinto, Marina WB-14, MD-39 

Di Giacomo, Laura MB-27, MC-27 

Di Luca, Camilla TD-26 

Di Pace, Roberta WA-27, WD-27 

Di Pillo, Gianni TD-43 

Di Puglia Pugliese, Luigi TA-02 

Di Vincenzo, Rosalba WA-48 

Diallo, Madiagne WB-06, MD-37 

Diarrassouba, Ibrahima TB-10 

Dias, Gustavo TB-48 

Dias, José G. MA-07, WC-29 

Dias, Luis MA-13, TF-15 

Dias, Luis C. MA-06, TB-30, TB-35, 

TD-37 

Diaz, Javier MA-03, TA-21 

Diethe, Tom MD-26 

Dietz, Simon MC-30 

Dikmen, Onur WC-08 

Dimaki, Katerina WD-18, MB-25 

Dimitriou, Vasileios MD-11 

Diniz, Milena MC-35 

Dirsch, Ruth WA-41 

Divjak, Blazenka WD-07 

Djeffal, Elamir WB-24 

Djeffal, Lakhdar WB-24 

Dobrodolac, Tunde MC-38 

Dodin, Bajis WA-07 

Doerner, Karl MC-03, TB-03, TC-15 

Dogan, Ibrahim TA-14 

Dogru, Mustafa MB-34 

Dohn, Anders WC-24 

Dokka, Trivikram MF-10 

Dolgui, Alexandre WD-39 

Dollevoet, Twan TC-16, WC-16 

Dominguez, Enrique MF-13 

Dominguez, Luis TF-35 

Domschke, Pia MB-09 

Donchev, Tzanko MB-39 

Donuk, Bilge TD-10 

Dopazo, Esther MA-12 

Doppstadt, Christian MB-15 

Dorbec, Paul MD-10 

dos Reis, Amália Faria MA-11 

Dotoli, Mariagrazia TC-27 

Doumpos, Michael MD-30 

Dovica, Ivan MB-16 

Dowart, Moses MB-37 

Do ˘Gdubay, Murat MF-37 

Drenovak, Mikica TF-29 

Drewes, Sarah WA-35, WD-35 

Drigo Filho, Elso TD-41 

Dris, Djamal TC-18 

Drosos, Dimitrios TD-30, TF-30 

Drouet, Laurent MA-33 

Drouineau, Mathilde WB-32 

Drovandi, Guido TC-24 

Drozdowski, Maciej TF-07 

Du, Bisheng MD-14 

Duarte Afonso, Humberto MD-02 

Duarte, Alfonso TB-30 

Duarte, António TD-21, MB-34 

Duarte, Isabel MC-19 

Dudzinska-Baryla, Renata WE-43 

Duer, Mirjam TD-43 

Dugardin, Frédéric TB-04 

Duggan, Jim TA-41 

Duhamel, Christophe MB-03, TD-03, 

TC-04 

Dulá, José WE-06 

Duran, Guillermo TB-05, MD-29 

Duran, Serhan TD-13 

Duscha, Vicki MA-47 

Dussault, Jean-Pierre TA-21 

Dutta, Joydeep TF-23, MD-42 

Duvvuri, Sri Devi TA-43 

Duzdar, Irem TA-12 

Duzgun, Ruken WB-10 

Dyson, Robert TF-05, MA-35 

Dytczak, Miroslaw MA-12 

E 

Ealet, Fabienne TA-45 

Eben-Chaime, Moshe MD-18 

Ebrahimian, A.h. MB-04 

Ebrahimkhany Ghazy, Soheila WA-10 

Eckhause, Jeremy MC-11 

Eden, Colin MB-35 

Edinger Munk Plum, Christian WD-22 

Edirisinghe, Chanaka WC-09 

Efendigil, Tugba WA-30 

Egan, Michael MC-30 

Eghbaal, Vahid MB-21 

Eglese, Richard MC-15 

Egriboyun, Feyzullah TC-25 

Ehmke, Jan Fabian TD-27, WB-27 

Ehrgott, Matthias MA-23, WA-44 

Eirinakis, Pavlos MB-10 

Eisenblätter, Andreas TF-31 

Eisenschmidt, Elke TB-29 

Ejov, Vladimir MF-18, TA-29 

Ekin-Karasan, Oya WE-13 

El Fallahi, Abdellah MC-03, WD-05 

El Khoury, Hiba MA-18 

El-ga’aly, Ashraf WB-11 

Elbassioni, Khaled MC-29 

Eldabi, Tillal TD-35 

Elhallaoui, Issmail WC-24, WB-31 

Elhedhli, Samir MD-02 

Elie, Ghanassia WB-32 

Elimam, Abdelghani WA-07 

Ellero, Andrea TA-47 

Elloumi, Sonda TA-23 

Emambocus, Mohammad TD-42 

Emel, Erdal MA-03 

Emiris, Dimitrios TF-27, TC-38 

Emmanouilides, Christos TC-46 

Emrouznejad, Ali WE-06 

Engau, Alexander MB-29 

Engelbeen, Celine TF-22 

Engell, Sebastian TB-28 

Engels, Volker TD-31 

Engineer, Faramroze TA-05 

Ensor, Andrew MD-23 

Ensslin, Leonardo MC-22, TA-35

Ensslin, Sandra TA-35 

Eppe, Stefan MD-30 

Epprecht, Eugenio MD-05 

Epstein, Rafael TB-05 

Eraslan, Ergun MB-12 

Erden, Feyza WD-21 

Erdogan, Gunes TA-15 

Erensal, Yasemin C. TC-12 

Ergun, Ozlem MA-17 

Eriksson, Ola MD-32 

Erismis, Burak WD-40 

Erjavec, Jure TD-44 

Erlwein, Christina MD-27 

Ernst, Andreas TF-08 

Erol, Serpil WA-28 

Ersoy, Cem WD-05, MA-40 

Erte, Idil MA-37 

Ertek, Gurdal WB-26 

Ertogral, Kadir WC-08 

Esbensen, Eystein Fredrik TC-15 

Escardino-Malva, Alberto TB-12 

Escrivá, Gema MF-04 

Escudero, Laureano Fernando TC- 

05, WB-16, MA-28, MD-28, 

MA-41, TC-48 

Esen, Mustafa WE-35 

Eshghi, Kourosh WB-05 

Eshragh Jahromi, Ali TB-09, WB-42 

Esperet, Louis WD-25 

Espinoza, Daniel WE-18 

Estellita Lins, Marcos MF-06, WD-06, 

MC-31, MC-35 

Estellon, Bertrand WE-05 

Euchi, Jalel TD-03 

Euler, Reinhardt MF-10 

Eusébio, Augusto MD-23 

Evans, Ian TB-27 

Even, Raphaël TC-08 

Evers, Lanah TD-45 

Ewe, Hendrik MB-17 

Ewing, Lee TC-45 

Exler, Oliver MD-09 

F 

F. Almeida, Joao Flavio WE-06, TB-34 

F. Campos Velho, Haroldo WC-05 

F. Vaz, A. Ismael WC-21 

Fabbri, Giorgio MA-39, MD-39 

Fabian, Csaba I. MA-27 

Facchin, Paola TB-47 

Faccio, Maurizio MD-04 

Faco’, Joao Lauro D. MD-47 

Fadjrir, Trishna MD-24 

Faezipour, Mehdi WA-12, WD-12 

Fagerholt, Kjetil TC-15 

Fagundez, Fabio MD-47 

Faias, Marta MB-19 

Falbo, Paolo WB-29, WA-33 

Fampa, Marcia WD-04 

Fanti, Maria Pia TC-27 

Fügenschuh, Armin TF-16 

Farago, Andras MF-40 

Farahi, Mohammad Hadi TF-48 

Farhang Moghaddam, Babak MB-03 

Farhash, Elham MB-03 

Faria, Daniel MA-24 


Farias, Ricardo WC-40 

Farkhi, Elza MB-39 

Fasano, Giorgio MA-20 

Fatemi Ghomi, S.m.t. WE-40 

Fathian, Mohammad MC-12 

Faulin, Javier MD-03, WC-19, MA-22 

Fawzi, Bessaih TA-40 

Fátima, Marilda WB-28 

Federico, Salvatore WB-14, TA-39 

Fedossova, Alina MF-26 

Fedotova, Olga MD-12 

Fehr, Max WB-33 

Feichtinger, Gustav TC-39 

Feil, Balazs TC-43 

Feillet, Dominique TA-40, MA-41 

Felden, Carsten TF-42 

Feldmann, Gunnar MC-41 

Felici, Giovanni TC-24 

Felletti, Daniele WB-29 

Ferguson, Mark MC-41 

Fernandes, Ana WA-13 

Fernandes, Edite M.G.P. TB-48 

Fernandes, Kiran TC-46 

Fernandes, Leão MD-47 

Fernandes, Luís MC-40, MF-45 

Fernandes, Sérgio TC-23 

Fernandez, Elena MD-01, TC-05, 

MD-13, TB-15 

Fernandez, Francisco Ramon MD-43 

Fernandez, Jose TB-13 

Fernandez, Raul Eduardo MD-33 

Fernández Barberis, Gabriela TC-30 

Ferrandi, Fabrizio MF-29 

Ferrante, Anna TB-47 

Ferrao, Paulo MF-33, TB-33, TA-44 

Ferreira Neto, José Ambrósio TD-04 

Ferreira, Ana TA-21 

Ferreira, Brígida da Costa WE-35 

Ferreira, Carlos TC-23 

Ferreira, Deisemara TA-34 

Ferreira, Fernanda A. TB-19, TD-19 

Ferreira, Fernando MD-12 

Ferreira, Flávio TD-19 

Ferreira, Helena MB-19, TC-19 

Ferreira, José Vasconcelos WA-27 

Ferreira, Liliana MD-32 

Ferreira, Luis MD-12, MD-19, TF-22, 

WA-43 

Ferrer, Albert WE-19 

Ferretti, Ivan MC-34 

Ferro, Eduardo WE-18 

Ferrucci, Francesco TC-15 

Fertis, Apostolos MB-09, MF-25 

Festa, Paola MC-24, TC-24 

Fetsun, Andryi WB-41 

Feyzioglu, Orhan WA-05, TD-14, 

MC-48 

Feyzollahi, Mohammad Javad WA-10, 

WC-20, WB-24 

Fiala, Petr TB-12 

Fialho, Andre MD-36 

Fichtner, Wolf TD-33, WC-33, TB-46 

Fidelis, Krzysztof MA-24, MB-24 

Fiedler, Sebastian WC-08 

Fiestras-Janeiro, M a Gloria WA-26 

Fievez, Veerle TD-04 

Figlali, Alpaslan MA-03 

Figueira, José Rui MC-23, MD- 

23, MF-23, MA-30, MC-43, 

MB-44 

Figueiredo, Francisco MA-07 

Figueiredo, Manuel MD-12, MA-13, 

TF-15, WA-31 

Figueiredo, Marisa MC-23 

Figueiredo, Rafael M. A. de MD-37 

Figueiredo, Rosa Maria MC-02, WE-22 

Filar, Jerzy TB-08, 09, MF-25, TA-29, 

MD-38, WB-42 

Finbow, Stephen WE-42 

Findik, Duygu TF-11 

Fink, Miloslawa MF-35 

Finkelstein, Stan MD-36 

Finkenstadt, Barbel MD-19 

Fiosins, Maksims TD-27 

Firat, Murat TD-07, TF-25 

Fischer, Kathrin MF-14 

Fischetti, Matteo MB-16 

Fiszman, Roberto MF-06, WD-06, 

MC-35 

Fitzpatrick, Shannon WC-42 

Flachskampf, Paul MD-35 

Flamini, Marta TB-08 

Flapper, Simme Douwe MC-07, TA-14, 

TA-27 

Fliedner, Malte MB-08 

Flindt Muller, Laurent WB-25 

Flisberg, Patrik TA-05, MB-32 

Flitman, Andrew TA-41 

Florentino, Helenice TA-19 

Flores, Mauricio WB-39 

Floudas, Christodoulos TD-15, WA-35 

Fodstad, Marte WC-22, MC-28 

Fogelholm, Carl-Johan TD-23 

Fondacci, Rémy WE-36 

Fonseca, Maria da Conceicao MC-32 

Fonseca, Raquel TF-36 

Fontes, Dalila TA-23 

Fontes, Dalila Martins TA-04 

Fontes, Fernando A. C. C. TA-23 

Formanowicz, Dorota MD-24 

Formanowicz, Piotr MD-24 

Forsell, Nicklas WC-32 

Forster, Geoffrey WD-24 

Fortemps, Philippe TA-23, TD-29 

Fortz, Bernard MB-40, MD-40, MF-41 

Fotakis, Dimitris WD-26 

Fourer, Robert MC-21 

Fra Paleo, Urbano TD-04 

Fragkos, Ioannis MA-34 

Fragnelli, Vito TF-26 

Fragoso, Rui TC-32, TF-32 

Framinan, Jose M MA-04, WC-07, 

WC-39 

Francesco, Cesarone MA-27 

Francfort, Stanislas WA-31 

Francisco, Mario TC-48 

Franco, Giuditta MF-24 

Franco, L. Alberto MF-35, TB-35, 

TC-35 

Frangioni, Antonio WB-19 

Franke, Rüdiger TD-47 

Franko, Ceki MB-05 

323


Franquesa, Carles TB-15 

Fröhling, Magnus MF-13 

Freire de Sousa, Jorge WC-27 

Freitas, Paulo TD-18 

Freixas, Josep WE-26 

French, Alan TD-42 

Frenk, J.b.g. MB-45 

Frey, Andreas MB-11 

Fridman, Lea TB-06 

Friedlander, Ana MA-42 

Friedrich, Bernhard TD-27 

Fries, Carlos Ernani TD-06, WE-18 

Froyland, Gary MF-16 

Fu, Michael MF-28 

Fuchsberger, Martin MC-16 

Fukuda, Emiko MA-19, TD-45 

Fukushima, Masao MF-46 

Fulga, Cristinca WE-29 

Fulop, Janos WC-20 

Funaki, Yukihiko TB-26 

Funk, Matthias WE-20 

Furian, Nikolaus TB-20 

Furini, Fabio MD-02 

Furmans, Kai WA-45, WB-45 

Fusco, Gaetano WA-16 

Fuss, Sabine WC-32 

G 

Gabay, Daniel WB-14 

Gabrel, Virginie TB-10 

Gabriel, Steven MC-11 

Gafarov, Evgeny MD-08 

Gaggero, Mauro MF-40 

Gaivoronski, Alexei MC-28 

Galé, Carmen MA-42 

Gallay, Olivier TF-17, WB-27 

Gallego, Jordi TC-32 

Galli, Laura MB-16 

Gallo, Giorgio TA-31 

Gallo, Mariano WB-16, WA-17 

Galvão Dias, Teresa MD-03, MF-16, 

MB-23 

Gambardella, Luca Maria WC-25 

Gammon, Mark TD-45 

Gan, Heng-Soon WE-22 

Günther, Markus TC-46 

Gürbüz, Tuncay TC-12 

Gürel, Sinan WC-07 

Güven, Caglar MA-33 

Garatti, Simone WA-21, MD-46 

Garín, María Araceli MA-28 

García Quiles, Sergio MC-13, MF-13, 

TC-13 

García, Irma WB-15 

García-Bouso, Ana MC-39 

García-Jurado, Ignacio TC-26 

García-Sanz, María D. TA-30 

Garcia, I. TB-13, TC-43, TF-43 

Garcia-bernabeu, Ana WD-30 

Garcia-Melon, Monica TB-12 

Garcia_Gonzalo, Jordi MD-32, TC-32 

Gardi, Frédéric WE-05, WC-25 

Gargallo, Pilar MF-12, TC-37 

Gargano, Francesco TD-17 

Garibaldi, Jonathan MC-24 

Garriga, Xavier WC-07 

324 

Gaspar, Miguel WA-06 

Gassel, Christian MC-16 

Gastaldi, Massimiliano WD-27 

Gaston, Kevin MC-46 

Gather, Thorsten MA-08 

Gathy, Maude WD-14 

Gauder, Markus TD-47 

Gaudioso, Manlio WB-19 

Gavalec, Martin MF-36 

Gavranis, Andreas WA-36 

Gavranovic, Haris WD-25 

Gay, David M. MC-21, WD-35 

Gazzola, Rosaura TB-32 

GÜnay, Noyan Sebla WA-39 

Gómez Ibañez, Alvaro TD-46 

Gómez-Limón, José A. TC-32 

Gönülol, Semiye MA-03 

Görmer, Jana TD-27 

Göttlich, Simone MB-09 

Gebhard, Marina MB-13, TB-22 

Gehring, Hermann MA-17 

Geiger, Martin Josef TD-05 

Geipele, Ineta WB-43 

Geissler, Bjoern WD-35 

Gel, Esma MC-14, MA-46 

Gelareh, Shahin WD-22 

Gendreau, Michel MF-03, TC-15, 

TD-15, TF-15, MC-17, MF- 

17, TA-17, WC-25 

Gendron, Bernard WC-13 

Genoese, Massimo WC-33, TB-46 

Genovese, Andrea TD-17 

Gentile, Guido WA-16 

Georgiadis, Patroklos MC-14 

Georgiou, Andreas MD-11, WA-12 

Georgiou, Ion MA-35 

Ger, Metin TD-47 

Geraghty, John MA-03 

Geraldes, Carla A. S. MD-14 

Geraldes, Carlos MA-11 

Gerchak, Yigal WE-39 

Gerdts, Matthias MD-09, MB-39, 

MF-39, TB-39 

Gergin, Zeynep WE-33 

Gershwin, Stan WC-28 

Gettinger, Johannes MB-44 

Gevezes, Theodoros WD-04 

Geyer, Alois WA-29 

Gfrerer, Helmut WB-44 

Ghaddar, Bissan MB-29 

Ghaemi Nasab, Fateme TD-36 

Gharehgozli, Amir Hossein TF-40 

Ghasem Esfahani, Mostafa WC-09 

Ghiami, Yousef WD-45 

Ghoreyshi, Seyed Mohammad MB-04, 

MB-34 

Ghosh, Diptesh TF-03, WD-04, MF-40 

Ghotbaddini, Maryam WC-20, WB-24 

Giacometti, Rosella MB-28 

Giallombardo, Giovanni WA-40 

Giannikos, Ioannis MA-13 

Gicquel, Céline WE-33, MA-34 

Giffhorn, Edilson MC-22, TA-35 

Gila-Arrondo, Aranzazu TB-13 

Gilanyi, Attila TF-34 

Ginchev, Ivan TF-34 

Ginda, Grzegorz MA-12 

Giner-Bosch, Vicent TA-48 

Ginestar, Concepción TC-32 

Giordani, Ilaria TC-24 

Giraldo, Fabian MC-37 

Girard, Thierry WC-04 

Girginer, Nuray MD-25 

Gittins, John TD-08 

Gkoumas, Konstantinos WA-16 

Gladysz, Barbara TD-32 

Glass, Celia WD-28 

Glau, Kathrin WC-46 

Glawischnig, Mex MF-22 

Glazebrook, Kevin WC-45 

Glensk, Barbara MA-33 

Glineur, François WB-40 

Gnecco, Giorgio MF-40 

Gnudi, Adriana WA-48 

Godinho, Ana TF-24 

Godinho, Maria Teresa TB-15 

Godskesen, Steffen WA-25 

Goel, Ankur WC-14 

Goensch, Jochen MF-43 

Goentzel, Jarrod MC-28 

Goerdt, Andreas TA-29 

Goes, Gabriela MB-19 

Goetz, Renan MA-39 

Gogkou, I. TA-24 

Gogus, Itir MD-41 

Gokay, Selim TC-25 

Golany, Boaz MD-19 

Goldacker, Götz MB-20 

Goldengorin, Boris WD-40 

Goldschmidt, Olivier WB-08 

Goletsis, Yorgos TA-24, WA-36 

Golias, Mihalis WA-40 

Gollowitzer, Stefan WC-13 

Gomes, A. Miguel WA-05, MB-20 

Gomes, Catarina WB-42 

Gomes, Eliane TB-32 

Gomes, Maria Isabel WA-13, TC-17 

Gomes, Marta Castilho MD-33 

Gomes, Orlando MB-19 

Gomes, Rui MD-03 

Gomes, Teresa MC-40 

Gomes, Tiago MA-32 

Gomez Gonzalez, Daniel WC-21 

Gomez, Trinidad TD-04 

Gomez-Corral, Antonio WB-18 

Gomez-Navarro, Tomas TB-12 

Gomide, Fernando MC-36 

Gonçalves, Graça MB-07 

Gonçalves, José Fernando TA-20 

Gonçalves, José Manuel TF-32 

Gonçalves, Michele WB-28 

Gonçalves, Rui TC-19 

Goncalves, Gilles WB-24 

Gondek, Verena WA-28 

Gondzio, Jacek MC-21 

Gonen, Amnon MA-38 

Gonzaléz-Olabarria, José MD-32 

Gonzalez, Eduardo TF-06 

Gonzalez, Ignacio WD-30 

Gonzalez, Marcela TB-32 

Gonzalez-R, Pedro L. MF-21 

González Alastrué, José AntonioWC-19

Goossens, Dries MF-10, TD-10 

Gopaladesikan, Mohan TC-02 

Gorelik, Viktor WC-12 

Gorgone, Enrico WB-19 

Gori, Stefano WA-16, WB-27 

Gorski, Jochen MA-23 

Gotzamani, Katerina WA-12 

Gounaris, Chrysanthos TD-15 

Gourdin, Eric MD-40 

Gouveia, Luis MA-02, TB-15, WC-15, 

WD-15, MB-40, MC-40 

Gouveia, Maria MA-06 

Govaert, Tim TC-14 

Gozzi, Fausto WB-14, MD-39, TA-39 

Grad, Sorin-Mihai WD-09 

Gradisar, Miro TD-44 

Grahl, Jörn TA-42 

Granichin, Oleg MC-18 

Grapsa, Theodoula N. TD-48 

Grasas, Alex TA-43 

Grasman, Scott MD-03 

Grass, Dieter TC-39 

Grasselli, Martino WB-14 

Gravot, David MC-21 

Gröflin, Heinz MD-04 

Greco, Salvatore TD-29, WA-37, 

WE-37 

Gregus, Marie Theres MB-25 

Gribkovskaia, Irina WD-39 

Griebsch, Susanne WD-46 

Griffa, Simone WB-37 

Griffiths, Jeff TD-28 

Grigoroudis, Evangelos MF-30 

Grimaud, Frédéric WD-39 

Gritzmann, Peter TB-45 

Grobleben, Katharina TA-27 

Groesser, Stefan TC-41 

Gronalt, Manfred WA-15, MA-32 

Grosso, Andrea MC-43 

Grothey, Andreas MC-21, MB-45 

Grunwald, Jessika TF-42 

Grusho, Alexander MB-18 

Grzechca, Waldemar TF-07 

Guardao, Luis MF-04 

Guarisco, Michael WE-14 

Guarracino, Mario MC-24 

Guden, Huseyin WB-45 

Gue, Kevin WA-45 

Guedes, Maria do Carmo TB-39 

Guenther, Hans-Otto TD-17, WC-31, 

TA-34 

Guerassimoff, Gilles TA-37 

Gueret, Christelle MA-15 

Guerrero, Carlos TB-13 

Guerrero, William MA-15 

Guerriero, Francesca TA-02, TD-28 

Guerrin, Francois TD-41, MD-44 

Guerry, Marie-Anne MD-11 

Gueye, Serigne WC-35 

Gugenheim, Dan TB-19 

Guha, Puja MA-19 

Guide, Daniel WA-11 

Guido, Rosita TD-28 

Guignard-Spielberg, Monique WE-35 

Guimarães, Luis MB-34 

Guimarães, Rui WD-06 


Guinand, Frederic MB-23 

Gujjula, Rico WC-31 

Guler, Kemal WD-26 

Gulkac, Hande MA-03, WA-38 

Gulli’, Daniel WA-40 

Gultas, Ilkay WE-36 

Gundogdu, Ceren TC-11 

Gundogdu, Erdem WA-05 

Guner Goren, Hacer MA-34 

Guner, Ali MB-15 

Guner, Banu WB-09 

Guney, Evren WD-05 

Gungor, Burcu MB-26 

Gupta, Arindam MD-11 

Gur Ali, Ozden MD-26 

Gurevich, Gregory WD-07, WE-17, 

TC-18 

Gurný, Petr MD-25 

Gurrieri, Massimo TD-29 

Gurzhiy, Vladimir TD-46 

Gussmagg-Pfliegl, Elisabeth TB-03 

Gutiérrez, César TF-44, WA-44 

Gutiérrez, Ester TD-06 

Gutiérrez, Iliana MC-37 

Gutierrez, Genaro WC-14 

Gutierrez, Jose M WB-38 

Gutierrez, Marco TD-13 

Gutierrez, Rafael S. WD-38 

Gutjahr, Walter WB-15 

Gwiggner, Claus TB-27 

Găianu, Mihail TD-34 

H 

Haase, Sabrina MF-46 

Habenicht, Walter MC-05, MC-43 

Hadad, Yossi TB-06, WE-17 

Hadjiconstantinou, Eleni MD-02, 

MF-03 

Hafizoglu, Ahmet MC-14 

Haghi, Narges WE-35 

Hagiwara, Motohiro TA-38 

Hagmayer, York TC-35 

Haguihara, Êmili WD-08 

Hahn, Gerd J. MB-14 

Haijema, René WC-38 

Hain, Sebastian MA-36 

Hajykazemian, Hassan MB-21 

Hakanen, Jussi TD-23 

Halachmi, Ilan TA-32 

Hamacher, Horst W. MB-46 

Hamaci, Samir TD-41 

Hamarat, Caner MF-44 

Hamdi, Khaoula MB-03 

Hamel, Andreas TF-23 

Hamzaçebi, Coskun TB-25, TC-42 

Hamzaoui, Leila MA-14 

Hanafi, Saïd MC-02, WE-15 

Hanen, Claire WB-08, WC-28 

Hanke, Michael WA-29 

Hansen, Neele TB-16 

Hansen, Pierre TD-01, WB-35 

Hanson, Jared MD-31 

Hanzalek, Zdenek WD-07, WC-11 

Hübner, Alexander MB-14, WC-45 

Hügel, Katrin TB-41 

Hülsmann, Michael TF-17, WB-27 

Haoxun, Chen TF-13, MB-17 

Harper, Paul MD-37 

Hartl, Richard MC-03, TB-03, TA-08, 

TC-15, TC-39 

Hartmann, Alexandra MD-17 

Hartnell, Bert WD-42 

Hasgul, Servet WB-09 

Hasle, Geir WA-31 

Hassanzadeh, Elmira MF-44 

Hassanzadeh, Yousef MF-44 

Hassin, Refael TD-13 

Hattingh, Giel MA-09 

Hatzl, Johannes TB-02 

Haurant, Pierrick WD-37 

Haus, Utz-Uwe TB-29 

Hautphenne, Sophie TB-18 

Havlik, Petr WC-32 

Hayashi, Shunsuke MF-46 

Hayel, Yezekael MA-41 

Hayez, Quantin TC-30 

Haythorpe, Michael TA-29 

Hazir, Oncu WB-07 

Házy, Attila TC-34 

Höfer, Mathias TD-47 

Hörmann, Wolfgang WD-41 

Hebrard, Emmanuel WD-25 

Heemeryck, Annelies MA-12, WA-43 

Heerda, Jan MA-48 

Hege, Laura WE-33 

Heid, Werner MD-17 

Heimerl, Christian MA-08 

Hein, Robert TA-44 

Heinimann, Hans Rudorf MA-21, 

WD-26, MB-32 

Heinz, Stefan TA-08, WA-25 

Heipcke, Susanne TC-21 

Hejducki, Jacek TD-03 

Hejducki, Zdzisław TD-03 

Hekimo˘glu, Mert Hakan MD-41 

Helber, Stefan MD-34 

Held, Anne TF-33 

Helin, Janne TF-32 

Helo, Petri TC-11 

Hempsch, Christoph TC-31 

Henao, Felipe TB-35 

Hendrix, Eligius M.T. TA-13, TC-43, 

TF-43 

Henriques, Carla TC-33 

Henriques, Pedro MB-06 

Hens, Luc MA-31 

Herdem, Canan WD-44 

Heredia, F.-Javier WC-19, MB-28 

Hernandez, MIguel WB-21 

Hernandez, Monica TD-04 

HernÁndez, Beatriz WB-44 

Herold, Johannes MC-11 

Herrmann, Sascha TD-31 

Herzog, Florian WB-23 

Heyde, Frank TF-23 

Hiete, Michael WC-47 

Higuchi, Hideki TD-45 

Hijazi, Hassan WB-35 

Hildmann, Marcus WB-23, WC-23 

Hillebrandt, Julia TC-31 

Hillege, Hans MB-30 

Hindle, Giles MD-35 

325


Hinojosa, Miguel A. MA-38 

Hinz, Juri WB-33, WD-33 

Hiramoto, Megumi WD-11 

Hirsch, Patrick WA-15, TA-22, MA-32 

Ho, Sin C. MF-03 

Ho, Ying-Chin WB-09 

Hoai Minh, Le MD-45 

Hoang Hai, Pham MA-31 

Hochreiter, Ronald WA-29 

Hoda, Samid TB-29 

Hogg, Jonathan MC-21 

Hohzaki, Ryusuke MA-19, WB-26, 

TD-45 

Hojda, Maciej TF-08 

Holmberg, Kaj MC-40, TF-42 

Holmberg, Stig C MB-11, TC-44 

Homfeld, Henning TF-16 

Hon, Cheng-Chuang WB-12 

Hongler, Max-Olivier TF-17, WB-27 

Honma, Yudai WA-13 

Hontelez, Jan TC-45 

Hoogeveen, Han TB-16 

Hooker, John N. TB-29, TB-31 

Horbach, Andrei TD-07, TC-08 

Horozoglu, Nayat TF-26 

Hoseini, Seyed Mohieddin TD-06 

Hosseini, Seyed Ahmad WB-46 

Hosseinzadeh Lotfi, Farhad MB-06, 

MD-06 

Hotta, Keisuke WC-09 

Hou, Jiang-Liang TF-18 

Hougaard, Jens Leth WE-43 

Howard, John WB-26 

Howell, Michael MD-36 

Hritonenko, Natali MA-39 

Hsiao, Yuan-Du WA-10 

Hsu, Chun-Chin MA-11 

Hsu, Pi-Fang WC-12 

Htiouech, Skander WB-04 

Hu, Tung-Cheng TA-25 

Huang, Chiao-Ling MA-11 

Huang, Dongbin WD-26 

Huang, Kai-I WB-11 

Huang, Mei-Hsiu TD-18 

Huang, Rongbing MC-13 

Huang, Shiuan-Yiang WB-10 

Huang, Tze Chin TB-38 

Huber, Christian WB-45 

Huber, Sebastian MC-25 

Huchzermeier, Arnd WD-45 

Hudelmaier, Andreas TF-28 

Huelle, Judith MF-12, TC-35, WC-47 

Huepping, Bjoern TB-39 

Hugues, Tchouankem TD-27 

Huguet, Marie-José WA-08 

Hui-Chih, Hung WE-24 

Huisman, Dennis TC-16, WC-16 

Hujala, Teppo TF-30 

Hulianytskyi, Leonid WA-18 

Humpola, Jesco MF-09 

Hunjak, Tihomir WA-30 

Hunter, Paul WC-42 

Hurkens, Cor TD-07, WB-25 

Husain, Iqbal WC-34 

Hussain, Deedar MD-12 

Huysmans, Johan WB-20 

326 

Hvattum, Lars Magnus TC-15, WE-22 

H˛edrzak, Magdalena TA-32 

I 

Isljamović, Sonja TB-11 

Iacobellis, Giorgio TC-27 

Ibrahimov, Maksud MC-05 

Ikeda, Takeshi TD-08 

Ilk, Ozlem WD-47 

Imahori, Shinji TA-20 

Imai, Akio TB-17 

Improta, Gennaro TD-17 

Inci, Ahmet Can MA-25 

Inderfurth, Karl MC-34 

Inkaya, Tulin WD-17, TB-42 

Innorta, Mario MB-28 

Inoue, Eduado MB-44 

Ioakimidis, Christos MC-33, MF-33, 

TA-44 

Ioannou, George TB-14, WB-31 

Ionescu, Lucian MD-16 

Ip, W.h. WD-15 

Irnich, Stefan TB-03, MF-15 

Ishizaka, Alessio MB-44 

Isik, Mine MD-12 

Istvan, Szuts WA-46 

Iyigun, Cem TB-42 

Izady, Navid TC-22 

Izmailov, Alexey WA-34 

Izquierdo, Josep M TD-26 

J 

J. W. James, Ross WC-09 

Jabali, Ola MC-15 

Jablonsky, Josef MF-12, MB-43 

Jablonsky, Petr TA-38 

Jacomino, Mireille WE-37 

Jaehn, Florian MB-08, WC-36 

Jafarizadeh, Babak WE-41 

Jagannathan, Rupa TD-22 

Jager, Christina MA-48 

Jagla, Jan-Hendrik MD-21 

Jahangiry, Pedram WE-41 

Jahanshahloo, Gholam Reza MD-06 

Jain, Manish TA-05 

Jalali-Naini, S.g.r. MD-14 

Janaqi, Stefan WE-37 

Jang, Jinbong TC-18 

Jang, Steve TA-41 

Jans, Raf MA-34, WB-38 

Jansen, Michiel MF-14 

Janssens, Gerrit TC-03 

Jütte, Silke WC-24 

Jaramillo, Patricia MC-37 

Jarboui, Bassem WC-04 

Jaschob, Mathias MA-23 

Jaunzems, Andrejs WA-42 

Javadi, Akbar MC-47 

Jørgensen, Erik MF-32 

Jönsson, Petrus WB-31 

Jeannet, Bernard WA-35 

Jegelka, Stefanie TD-02 

Jenkins, Barbara MF-41 

Jensen, Thomas Sejr WA-25 

Jepsen, Mads Kehlet MD-15, TD-15 

Jha, Sanchita MF-21 

Jhang, Jhu-Ning MF-07 

Jiang, Houyuan MD-41 

Jiang, Xinjia TF-40 

Jiménez, Bienvenido TF-44, WA-44 

Jiménez-Losada, Andrés WA-26 

Jimenez-Lopez, Mariano MB-36 

Jlassi, Jihen MA-37 

Joao, Isabel M. TD-30 

Joerin, Florent WC-37 

Johar, Farhana WA-15 

Johnson, Johnnie TF-10, WB-17 

Joncour, Cédric WB-09 

Jones, Dylan MB-36, MB-44 

Jongen, Hubertus Th. MB-42, MD-46 

Josa-Fombellida, Ricardo WB-14 

José Martins Ferreira Filho, VirgílioTB- 

48 

Jost, Vincent WB-25, MC-48 

Joubert, Johan TA-17 

Jouini, Oualid WA-24 

Jovanovic, Milos TF-18 

Jozefczyk, Jerzy MB-04 

Jozefowiez, Nicolas MD-23 

Jozefowska, Joanna WB-39 

Jozefowski, Lukasz WB-39 

Juan, Angel A. MD-03, TC-03, WC-19, 

MA-22 

Juan, Pin-Ju WB-12 

Judice, Joaquim MB-42, MC-42, 

TC-43, MF-45 

Juenger, Michael MA-02 

Jung, Gimun WA-20 

Junglas, Daniel WB-36 

Jurun, Elza MB-25 

Justino, João WD-07 

K 

Ka˘nková, Vlasta MA-28 

Kabak, Özgür TF-42 

Kabyl, Kamal WA-18 

Kaci, Souhila MD-48 

Kaczmarczyk, Waldemar TB-34 

Kafash, Mohamad Hosein TD-34 

Kaimakamis, George TB-25 

Kainich, Alexandros WC-44 

Kaiser, Gernot TB-22 

Kaiser, Markus TA-13 

Kajiji, Nina TF-21 

Kalashnikov, Vyacheslav MC-09, 

MC-42 

Kalashnykova, Nataliya MC-09, MC-42 

Kallio, Markku WC-41 

Kamisli Ozturk, Zehra TA-09 

Kamisli, Melik MD-25, WC-29 

Kan, Cihangir MB-05 

Kana, Vivien MB-43 

Kanat, Irfan Emrah TF-11 

Kandakoglu, Ahmet MC-12 

Kang, Boda WD-46 

Kangas, Annika TF-30 

Kao, Shih-Chou WD-17, TA-31 

Küfer, Karl-Heinz MB-17, TD-23, 

MF-46 

Kürüm, Efsun MD-25 

Kapamara, Truword TD-28 

Kapelko, Magdalena WA-06

Kapoor, Reena WD-09 

Kar, Anirban TA-26 

Kara, Bahar Yetis WE-13 

Karabacak, Bilge MF-11 

Karabulut, Ilayda MC-15 

Karabulut, Ozlem WB-04 

Karaca, Nihan WE-36 

Karadayi, Melis Almula MD-12 

Karagiannis, Giannis TA-06 

Karakostas, George WD-26 

Karaman, Hazal TD-14 

Karamanis, Dimitrios TB-25 

Karamzin, Dmitry TC-39 

Karasakal, Esra WB-04, MB-05 

Karimi, Hossein TD-05 

Karimi-Nasab, M. MB-04, TA- 

04, WC-04, MD-14, MB-34, 

TA-34, WE-40 

Karkazis, John MA-13, MB-13 

Karmen, Pazek TB-44 

Karmitsa, Napsu WD-19 

Karpak, Birsen TA-12 

Karpov, Igor WB-28 

Karsten, Frank WA-26 

Kasap, Nihat TB-28 

Kasimbeyli, Refail MA-26 

Kaspar, Ralf MF-12 

Kasper, Ulf WC-33 

Kasprzyk, Rafal WA-30 

Kassa, Rabah WA-21 

Kat, Bora MA-33 

Katsev, Ilya TB-26 

Katsis, Christos TA-24 

Katsuda, Hideki TA-38 

Kaur, Parmjit MA-35 

Kaur, Simranjit TF-09 

Kaut, Michal TB-19 

Kavakli, Halil TD-24 

Kawamura, Hidenori TD-08, WC-12 

Kawas, Ban MA-27 

Kaya, Murat TB-14, WB-26 

Kaya, Onur TC-14, TF-15 

Kayakutlu, Gulgun TA-12, TC-12, 

WC-12, TC-36 

Kayaligil, Sinan WD-17 

Kaymak, Uzay MA-36 

Keisler, Jeffrey MC-44 

Keles, Dogan TD-33 

Kellerer, Hans WA-28 

Kendall, Graham MD-24 

Keppo, Ilkka TF-33 

Kerbache, Laoucine MA-14, MA-18 

Keren, Baruch TB-06, TC-18 

Kermek, Dragutin WA-30 

Kern, Walter TC-02 

Kersch, Mike WA-29 

Kesek, Zbigniew MB-31 

Keshvari, Abolfazl TC-07 

Keskin, M. Emre MA-40 

Keskinocak, Pinar TA-05, TD-13, 

MC-14 

Kezi, Csaba TF-34 

Khachay, Michael MA-18, WA-23 

Khamisov, Oleg MC-42 

Khan, Mohammed TA-46 

Khanafer, Ali TB-20 


Khandelwal, Ankit TD-09, TB-27 

Khazaei, Javad MB-28 

Khemakhem, Mahdi MA-37 

Khosravi, Banafsheh WC-16 

Khuong, Paul-Virak WC-13 

Kieffer, Yann WD-05, WC-20, WD-25 

Kiekintveld, Christopher TA-05 

Kiesel, Antje TF-22 

Kiesel, Rudiger WB-33 

Kiesling, Elmar MB-44, TC-46 

Kilic, Kemal MC-36 

Kilic, Okan MA-31 

Kim, Dongmin WD-19, WA-34 

Kim, Eun-Seok WD-28 

Kim, Kap Hwan TF-40 

Kim, Sangho MF-37 

Kim, Seokjin MC-13 

Kim, Seong-Jun MF-42 

Kim, Seongmoon MB-46 

Kim, Sung-Shick WA-11 

Kim, Woo Chang MA-25 

Kimms, Alf TF-27, MF-43 

Kinoshita, Eizo WD-12, MB-31 

Kirchner, Kathrin TF-18 

Kiriakopoulos, Konstantinos TB-25 

Kirkizlar, Eser TB-47 

Kis, Tamas WB-21 

Kjeldsen, Karina WB-22 

Kjeldsen, Niels WA-25 

Klamroth, Kathrin TD-07, TA-13, 

MA-23, TD-24 

Klatt, Tobias MF-12, WC-47 

Klatzmann, David MB-24 

Kleber, Rainer MC-34 

Kleindorfer, Paul MC-33 

Kleine, Oliver TC-41 

Klemmt, Andreas WB-07 

Klerides, Evelina MD-02 

Kliewer, Natalia MD-16, TB-16, TD-40 

Kline, Jeffrey WE-22 

Klinkert, Andreas WA-24 

Kljajic, Miroljub TB-44, TC-44 

Kljajic-Borstnar, Mirjana TC-44 

Kloster, Oddvar WA-31 

Knüppel, Peter TF-42 

Knauer, Matthias MD-09 

Knoke, Thomas WC-41 

Knust, Sigrid WC-36 

Ko, Hong Seung WD-11 

Kobayashi, Takashi TA-25 

Koberstein, Achim MD-16 

Kocabiyikoglu, Ayse MD-41 

Koch, Ronald MB-09 

Kochetov, Yury MA-42 

Kochetova, Nina MA-42 

Kodera, Jan WE-44 

Koester, Christian TB-43 

Kofjac, Davorin TB-44, TC-44 

Kogan, Alex TC-29 

Kokkinis, Basil MF-39 

Koksalan, Murat MC-43 

Kolak, Orhan Ilker MC-48 

Kolarov, Ivan WC-11 

Kolb, Oliver MB-09 

Kolisch, Rainer MA-08 

Kolliopoulos, Stavros WD-26 

Kolodzey, Alexander. MB-18 

Komiya, Toru MA-19, TD-45 

Kondrakov, Ivan WB-23 

Kongo, Takumi TB-26 

Konnov, Igor WA-48 

Konstantas, George WD-18 

Koole, Ger WA-24 

Kopanska-Brodka, Donata WE-43 

Kopfer, Heiko MB-17 

Kopfer, Herbert MA-17, MB-17, 

MC-17, MD-17 

Koppang, Haavard TA-31 

Korgin, Nikolai TD-11 

Korotchenko, Ielyzaveta WB-47 

Korpeoglu, Evren WD-26 

Kort, Peter M. MA-19, TC-39 

Kortbeek, Nikky MB-30 

Korzilius, Hubert TB-35 

Koschtial, Claudia TF-42 

Koseoglu, Pinar MA-37 

Kosowski, Adrian MC-10 

Kostarelou, Eftychia WA-36 

Kostoglou, Vassilis WD-08 

Kostrzewska, Marta WB-46 

Kotiadis, Kathy TD-35 

Koulis, Alexandros TF-29 

Kourentzes, Nikolaos MC-28 

Kouvela, Anastasia MF-10, TC-10 

Kovacevic Vujcic, Vera TC-10 

Kovacevic, Ivana TA-18 

Kovacevic, Raimund WD-29 

Kovacs, Edith WA-46 

Kovalev, Sergey WD-39 

Kovriga, Svetlana MC-35 

Kozanidis, George WA-36 

Kpoumie, Amidou WB-37 

Kramer, Alpar Vajk MA-10 

Kramkowski, Stefan MD-16 

Krarup, Jakob MF-48 

Krasnogor, Natalio MF-24 

Krasotkina, Olga WC-23, WD-23 

Krass, Dmitry TC-13 

Kratica, Jozef TC-10 

Kratz, Marie WD-43 

Kraus, Sarit MC-18 

Kravchenko, I. A. MC-18 

Krawczyk, Jacek MD-38 

Krejic, Natasa MD-27, 28, MF-28 

Kremers, Enrique TD-46 

Krempasky, Thorsten TA-13 

Kresz, Miklos TC-04 

Kristensen, Anders Ringgaard MF-32 

Kristjansson, Bjarni MD-21, TC-21 

Kritzinger, Stefanie MC-03 

Krklec, Natasa MD-28, MF-28 

Krkosková, Šárka MF-37 

Krokhmal, Pavlo TC-18, WD-29 

Kronek, Louis-Philippe TC-42 

Kroon, Leo MA-16, MB-16, WC-16 

Kropat, Erik WD-44, MC-46 

Kruger, Hennie MA-09 

Krumke, Sven TB-16 

Kruse, Rudolf TD-29 

Kruse, Susanne WC-46 

Krushinsky, Dmitry WD-40 

Krymova, Ekaterina WA-23 

327


Kubali, Deniz TC-14 

Kubiak, Wieslaw WB-39 

Kubo, Mikio MA-45 

Kucera, Petr TC-29 

Kucherenko, Sergei TC-43 

Kuchta, Dorota TD-32, TC-36 

Kucuk, Mahide WC-34 

Kucuk, Yalcin WC-34 

Kucukaydin, Hande MA-09 

Kuhn, Daniel TF-36 

Kuhn, Harold W. TE-01 

Kuhn, HeinrichMB-14, WA-38, WC-45 

Kuilman, Jan MA-31 

Kula, Ufuk WE-08, WD-45 

Kulak, Osman TD-17 

Kumar, Krishna MA-19 

Kumar, Shruthi S MA-22 

Kumar, Soumojit MC-25 

Kumaresan, Miles MD-27 

Kumbaroglu, Gürkan TB-33 

Kunc, Martin MA-35 

Kunkel, Martin MD-09 

Kunsch, Pierre MD-44 

Kurumatani, Koichi TD-08 

Kurz, Sascha MF-02 

Kutlu, Ahmet Can MB-12 

Kuula, Markku WC-41 

Kuzmanovic, Marija TB-42 

Kuzmina, Lyudmila MC-39 

Kvasov, Dmitri TF-43 

Kyngas, Jari TF-24 

Kyriazis, Panagiotis TB-14 

L 

Labbé, Martine TC-01, MC-02, MB-40, 

MD-40, MF-41 

Labbi, Wafaa WD-28 

Labib, Ashraf TB-12 

Lacomme, Philippe MB-03, TD-03, 

WA-08 

Lacroix, Mathieu TB-10 

Lagarde, Lionel TF-16 

Lahrichi, Nadia MF-03, TA-17 

Lai, Guoming MC-27 

Lai, Tzyy Jane WD-11 

Laib, Fodil WD-44 

Lamanna, Rosalba TC-48 

Lamine, Amine MA-37 

Lamond, Bernard MB-21 

Lan, Yu Wen TB-38 

Lancia, Giuseppe MB-02 

Landa, Paolo TA-22 

Landete, Mercedes TF-03, TC-05, 

MF-13 

Landry, Jean-François TA-21 

Lang, Jens MB-09 

Lange, Jean-Charles MB-13 

Langenberg, Nils MA-48 

Lübbecke, Marco TA-08 

Laporte, Gilbert TA-13, 

WE-13, MC-15, TA-15, WA- 

15, MD-23, TB-23 

Larabi, Mohand WA-08 

Lari, Isabella MC-29 

Lario Esteban, Francisco MB-36 

Larrañaga, Pedro TA-42 

328 

Larsen, Christian MD-14, TB-47 

Larsen, Jesper WA-15, TB-17, WC-24 

Larsen, Rune TB-20, WA-25 

Laslo, Zohar WD-07, TC-18 

Latouche, Guy MD-18 

Lau, Hoong Chuin TF-28 

Laugier, Alexandre TD-07 

Laukkanen, Timo TD-23 

Laumanns, Marco MB-09, WE-14, 

MC-16, WC-16 

Laundy, Richard WB-36 

Lavangnananda, Kittichai MB-23 

Lavor, Carlile MC-02 

Layter Xavier, Vinicius TF-43 

Lazarev, Alexander MD-08, WB-28 

Lazarin, Daniel MC-17 

Létocart, Lucas WD-25, WB-35 

López, Jorge TB-26 

López-Redondo, Juana TB-13, TC-43 

Löhne, Andreas TF-23 

Le Thi, Hoai AnMD-45, MF-45, TA-45 

Leão, Celina Pinto TA-21 

LeBel, Luc MB-32, TD-32 

Leblebicio˘glu, Kemal WE-17 

Lebrón, Esperanza WA-26 

Lebreiro, Nuno WA-31 

Lee, Hong Tau TB-04, MA-06 

Lee, Hsu-Hua TD-18 

Lee, Jon WA-35 

Lee, Lai-Soon MB-03 

Lee, Loo Hay WE-24, TF-40 

Lee, Wen Yi TA-25 

Lee, Wen-Chiung WC-07 

Lee, Wenyih TB-14 

Lefèvre, Claude WD-14 

Legato, Pasquale TD-17, WA-40 

Lehaux, Thomas WE-37 

Lehmann, Thomas MD-09 

Lehtonen, Heikki TA-32 

Lejeune, Miguel MA-25 

Leleur, Steen TC-37 

Lemos, Felipe TD-20 

Leon, Coromoto MF-20 

Leon, Maria Amparo TD-04 

Leon-Rubio, Santiago TB-12 

Leoni, Peter WC-46 

Leopold, Ulrich MA-33 

Leopold-Wildburger, Ulrike MA-38 

Leow, Mindy WC-30 

Lesaja, Goran WB-34 

Leskinen, Pekka TF-30 

Leskovar-Špacapan, Gabrijela WB-47 

Lessard, François WC-25 

Lessmann, Stefan WA-20 

Leung, Kwai-sun MA-18 

Leus, Roel WA-07 

Levin, Yuri MB-41 

Levina, Tatsiana MB-41 

Levy, João Quinhones MD-33 

Leyva-Lopez, Juan Carlos TB-30 

Li, Dan TC-09 

Li, Duan TD-43 

Li, Jing-An TD-40 

Li, Jingpeng WE-05 

Li, Qing MC-41 

Li, Tz-Chiang TB-07 

Li, Yan Chen TB-07 

Lian, Zhaotong WE-44 

Liang, Tzu-Yin TA-31 

Liao, Chia-Chun WA-42 

Liao, Shuangqing WA-24 

Liao, Ta-Wei WB-09 

Liberati, Diego MF-29 

Liberatore, Federico TC-13, MB-37 

Liberti, Leo MC-02, WA-35, WB-35 

Libura, Marek TD-02 

Lidén, Bertil MB-32 

Lidouh, Karim MF-23 

Liefooghe, Arnaud MC-23 

Lien, Chung-Chang WB-10, TD-18 

Liern, Vicente MB-36, MC-36 

Liers, Frauke MB-29, TB-29 

Liesiö, Juuso MA-44 

Lifvergren, Svante TD-22 

Lilia, Zaourar WD-05 

Lillo, Felipe MD-23 

Lin, Chin-Piao TC-07 

Lin, Chin-Tsai TA-25 

Lin, Shin-Yi WB-12 

Linnala, Mikko MA-45 

Lino, M.Pilar WA-39 

Linton, Jonathan MA-14 

Lioris, Jennie TA-47 

Liour, Yuanchau MA-11, TD-11 

Liptrot, Tom MF-41 

Lis, Marcin W. TA-32 

Lisboa, Adriano TD-48 

Lisse, Nora WD-43 

Listes, Ovidiu TC-21, WE-46 

Litsardaki, Maria TD-37 

Litvak, Nelly MF-18, MB-30 

Liu, Fuh-Hwa TA-06 

Liu, Jiyin WC-31 

Liu, Kuangyi MF-18 

Liu, Qian MB-41 

Liu, Shasha WD-11 

Liu, Xiaoming WE-44 

Liu, Yang MC-38 

Lium, Arnt-Gunnar WC-22 

Ljubic, Ivana MA-02, WC-13, MB-40 

Llop, Jose Ignacio WA-39 

Llorca, Natividad TA-26, TB-26 

Lo, Mei-Chen TD-12, MC-38 

Loan, Dinh Thi Phuong MA-31 

Lobo, Victor MC-20 

Locatelli, Marco TA-48 

Lodi, Andrea WC-35 

Loeffen, Ronnie WD-14 

Loher, Marcel TB-41 

Loiseau, Irene MA-02 

Lokman, Banu MC-43 

Longo, Giovanni TD-12, MC-16, 

WD-27 

Lopes, Isabel Cristina TD-20 

Lopes, Maria do Carmo WE-35 

Lopes, Maria João MD-13 

Lopes, Rui Borges TC-23 

Lopez Montoya, Ruben Luis WA-44 

Lopez, Claudia MD-20 

Lopez, Francisco MA-16 

Lopez, Pierre WA-08 

Lopez-Herrero, Mj WA-46

Lorena, Luiz A. N. TB-02, MD-47 

Lorenzo, Leticia TA-26 

Lorenzo-Freire, Silvia TA-26 

Losada, Chaya MC-13 

Losch, Nadine MF-18 

Louati, Hekma MA-37 

Loukil, Taicir TA-23 

Loulou, Richard WB-32 

Lourenço, Helena Ramalhinho TC-03, 

TF-13, WC-19 

Lourenço, João Carlos TD-37 

Lourenço, Lídia MB-07 

Louzis, Dimitrios MF-25 

Lova, Antonio MA-04 

Lozano, Antonio J. MD-16, TA-16 

Lozano, Sebastián MC-06, TD-06 

Luè, Alessandro WB-27, MB-43 

Lucas, Helena TF-47 

Lucena, Abílio MB-02 

Lucertini, Giulia WB-37 

Lucheroni, Carlo TF-29 

Lucidi, Stefano TD-43, TA-48 

Luckert, Marty WA-41 

Luhandjula, Monga K MD-46 

Lukasiak, Piotr MA-24, MB-24 

Luna, Mônica M. M. TD-06, WE-18 

Lundberg, Kristian TF-42 

Lundell, Andreas WB-35 

Luo, Jiabin TD-40 

Luo, Kai MA-14 

Luptacik, Mikulas WC-06 

Luque, Mariano TD-23 

Lusa, Amaia WA-09 

Lusby, Richard TB-17, WC-24, WB-25 

Lust, Thibaut MA-23 

Lustosa, Leonardo TB-43 

Lutsenko, Mikhail MF-19 

Luz, Carlos J. MB-07 

Luz, Eduardo WC-05 

Luzanin, Zorana WB-19 

Lyra, Christiano WC-40, MB-45 

Lyytikäinen, Tapani TA-32 

M 

Maassen, Klaus-Christian TF-27 

Maïzi, Nadia WB-32, TA-37, TB-37, 

TD-38 

Mabin, Vicky MB-35 

Macedo, Mark WB-05 

Macedo, Rita WE-15 

MacGillivray, Gary WD-42 

Machado Cardoso Junior, Moacyr MF- 

42 

Macharis, Cathy MA-12, MC-31, 

WA-43 

Mack, Alexander WA-06 

Maculan Filho, Nelson MC-02, WC-02 

Madas, Michael TB-08, WA-42 

Madlener, Reinhard MA-33, MC-33 

Madureira, Ana WC-31 

Maenhout, Broos MF-08 

Magbagbeola, Joshua TC-11 

Maggioni, Francesca MC-28, MA-40 

Magos, Dimitris MB-10, MF-10, TC-10 

Mahecha, Nancy WC-44 

Mahey, Philippe TA-21 


Mahjoub, A. Ridha TA-10, TB-10 

Mahmoudzadeh, Houra WB-05 

Makajic-Nikolic, Dragana MD-18 

Makarenko, Dmitry TD-11 

Maksa, Gyula TF-34 

Makulska, Joanna M. MF-32 

Makuschewitz, Thomas MC-14 

Malczewisk, Jacek MF-23 

Malliappi, Fotini WD-22 

Mallor, Fermin WA-36 

Mamasis, Konstaninos TA-17 

Mammadov, Musa WE-19 

Mamon, Rogemar MD-27 

Manaa, Adel WC-28 

Manca, Vincenzo MF-24 

Maneshi, E. WC-04 

Manfredotti, Cristina WE-30 

Manger, Robert WD-28 

Mannini, Livia WB-27 

Manolitzas, Panagiotis TD-30, TF-30 

Mansi, Raid MC-02 

Mansini, Renata MC-03 

Müller, Johannes TD-38 

Müller, Thomas TC-31 

Müller-Merbach, Heiner MF-48 

Maquera, Gladys MC-37 

Maquera, Jorge MC-37 

Mar Molinero, Cecilio MC-22, MB-31 

Maratini, Riccardo WD-27 

Marí, Laura WC-19 

María, Gómez-Rúa TA-26 

Marín, Alfredo MD-13, MF-13 

Marín, Ángel MA-16 

Marín-Solano, Jesús WC-14 

Marchant, Thierry TA-30 

Marchese, Mario MF-40 

Marcikic, Aleksandra MA-38 

Marciniak, Dorota WE-26 

Marecek, Jakub MF-02 

Marenco, Javier MD-29 

Marentakis, Charis TF-27, TC-38 

Mareschal, Bertrand TC-30, TF-37 

Margalida, Antoni MF-24 

Margot, Francois MC-27 

Mariano Junior, Marcio WE-06, TB-34 

Marinkovic, Boban MF-47 

Marins, Fernando TC-32 

Markopoulou, Chrysi MB-38 

Markov, Michael WC-23, WD-23, 

TB-25 

Marmion, Marie-Eléonore TC-42 

Maroti, Gabor MA-16, MB-16 

Maroto, Concepcion TC-32 

Marques, Alexandra MA-32 

Marques, Ana MB-19 

Marques, Carlos TF-32 

Marques, Inês TD-28 

Marques, Rui TD-06, WB-06 

Marques, Vitor MC-36 

Marschner, Andreas TC-31 

Martín, Jacinto MA-43 

Martein, Laura TC-34 

Martel, Jean-Marc TF-37 

Martens, David WC-30 

Martic, Milan TB-42 

Martin, Alexander TB-08, TF-16, 

MB-24, WD-35, TD-38 

Martin, Arnaud MD-43 

Martin, Quintin TD-04 

Martin, Sébastien TB-10 

Martin, Simon TD-46 

Martin-Campo, Francisco Javier TC-48 

Martinez de Albeniz, Javier TC-26 

Martinez Sykora, Antonio MA-15 

Martinez, Carme WA-09 

Martinez, Cristian TB-05 

Martinez, Maria MA-22 

Martinez, Yailen TF-28 

Martins, Ana Alexandra TF-18 

Martins, Isabel MC-32 

Martins, José TA-19 

Martins, Luís B. TA-21 

Martins, M. Joao MB-07 

Martins, Pedro TB-24 

Mas, Marta WA-09 

Massei, Gianluca WA-37 

Massetti, Emanuele TB-33 

Massim, Yamani WA-04 

Massol, Olivier MC-11 

Masuda, Ryuichi WB-26 

Matallin-Saez, Juan Carlos WC-06 

Mataracioglu, Tolga MF-11 

Mateo, Manuel WC-07 

Mathieu, Fabien MD-40 

Matias, João WE-19 

Matos Dias, Joana WE-35 

Matos, Henrique TF-35, MD-47 

Matsuo, Jun TD-08 

Mattfeld, Dirk Christian WB-27 

Mattia, Sara TF-31 

Mattoo, Rumana WC-34 

Matuszyk, Anna TB-06 

Maugeri, Antonino WA-48 

Mauri, Geraldo TB-02, MD-47 

Mautor, Thierry TB-18 

Mavralexakis, Theodoros TB-25 

Mavri, Maria MB-13, WD-18, MB-25 

Mawengkang, Herman MD-24, TB-24, 

TA-46 

May, Angelika WD-43 

Maya, Pablo WA-17 

Mazalov, Vladimir TF-26 

Mazauric, Vincent WB-32 

Mazur, Christoph MC-33 

Mazza, Rina Mary TC-27 

Mazzoldi, Laura MC-34 

Mármol, Amparo MA-38 

Mäkelä, Marko M. WD-19, WA-44 

Mäkinen, Antti WC-41 

Möhring, Rolf TA-08 

Möst, Dominik TD-33, WC-33, TB-46 

Mbuntcha-Wuntcha, Calvin MC-29 

McCaffrey, David MF-41 

McCarney, Geoff WA-41 

McCollum, Barry TD-39 

McComb, Sara TC-35 

McDonald, David WB-17 

McGarraghy, Seán MD-40 

McGill, Jeff MB-41 

Mckenna, Angela TD-22 

McKenny, Dan WA-41 

329


McMahon, Anne M. TA-12 

Medeiros, Helena C. TD-06 

Medeossi, Giorgio MC-16 

Medina, Afonso Celso TD-40 

Medina, Rosa MF-02 

Meeraus, Alex WE-35 

Mehrjoo, Razieh TD-36 

Meisel, Frank TB-17 

Meissner, Joern WD-38, MA-41, 

WC-45 

Mekiker, Cagatay TD-14 

Melachrinoudis, Emanuel WD-13 

Melechovsky, Jan MC-03, WC-15 

Mellano, Mario WC-27 

Melo, Teófilo TB-09 

Melo, Teresa MB-13 

Melo, Wendel WD-04 

Meloni, Carlo MC-04 

Memmedli, Memmedaga WC-17, 

MB-18, TB-36 

Mendes, Armando WE-20, TD-21 

Mendes, Maria Fernanda MA-35 

Mendes-Moreira, João WC-27 

Mendez, Carlos WC-19 

Mendonca, Vitor TA-19 

Meneguzzer, Claudio WD-27 

Menezes, Mozart MC-13, TC-13, 

MC-28 

Menoncin, Francesco WB-14 

Meral, Sedef TA-14, MA-47 

Mercan, Sinan TD-14 

Merino, María MA-28 

Merkuryeva, Galina WC-43 

Mersha, Ayalew Getachew MD-42 

Mert, Erkan MA-31 

Merz, Mirjam WC-47 

Mesa, Juan A. MD-16, TA-16 

Mesgarpour, Mohammad MC-08 

Mesquita, Marta MF-16 

Messina, Enza TC-24, WE-30 

Messine, Frederic WA-35, WB-35, 

TD-39, TC-43 

Messinger, Margaret-Ellen WC-42 

Messner, Stephan TF-22 

Mestre, Ana TB-47 

Mestre, Pedro WE-19 

Mesyagutov, Marat MF-20 

Meunier, Frédéric MB-10 

Meyer, Christoph Manuel MD-17 

Meyr, Herbert MB-14 

Mezei, József WA-44 

Mezzadri, Malik MC-29 

Miandoabchi, Elnaz WA-05 

Michaels, Dennis WC-35 

Michalewicz, Zbigniew MC-05 

Michel, Sophie WB-09 

Michelon, Philippe MF-02, WC-35, 

TA-40 

Middelkoop, Dick WC-16 

Midthun, Kjetil MC-28 

Miettinen, Kaisa TD-23 

Miglierina, Enrico TF-44, WB-44 

Miglionico, Giovanna WA-40 

Mijangos, Eugenio WC-19 

Mika, Marek WC-39 

Milanic, Martin MD-29 

330 

Milanovic, Zorica MC-16 

Milioni, Armando MD-06 

Miliotis, Panayiotis MB-10 

Milostan, Maciej MB-24 

Milstein, Irena TA-33 

Min, Hokey WD-13 

Mingers, John MF-48 

Mingozzi, Aristide MB-02, TA-02, 

MD-15, MF-15 

Minis, Ioannis TA-17 

Minner, Stefan WC-38 

Minoux, Michel WE-33, MA-34 

Miranda, Gara MF-20 

Miranda, Jaime WE-18, WD-24 

Miranda, Joao MD-22 

Mirbolooki, Mahnaz MB-06, MD-06 

Mishra, Nishant MC-41 

Missbauer, Hubert MD-34 

Mitra, Gautam MA-27 

Mitsakis, Evangelos TA-47 

Mittelmann, Hans TC-21, WC-35 

Miyagawa, Masashi WA-27 

Miyazaki, Kenji WC-44 

Mkaouar, Arij MA-37 

Mladenovic, Nenad WB-13 

Moccia, Luigi TB-17, WA-40 

Modarres, Mohammad WC-20, WB-24 

Moeini, Mahdi MD-45 

Moeller, Klaus MF-12, MF-18, TC-35, 

WC-47 

Moench, Lars WB-07 

Moewes, Christian TD-29 

Moghaddam, Atefeh MA-04 

Mohafid, Abdelmoula WD-39 

Mohais, Arvind MC-05 

Mohamad Raee Naeeni, SepidehWE-41 

Mohamed, Nurul TC-17 

Mohamed, Reghioui MC-03 

Mohammadian, Ghazaleh MF-25 

Mohd Razali, Noraini MA-03 

Mohr, Esther WA-29 

Mojtahedzadeh Sarjami, Shabnam TA- 

36 

Molenaar, Douwe MA-09 

Molho, Elena TF-44, WB-44 

Molina, Julian TD-04 

Molina-Pariente, Jose M. MA-04, 

WC-07, WC-39 

Monaco, M. Flavia TC-27 

Moncel, Julien MD-10 

Monge, Juan Francisco TF-03, MA-41 

Monnot, Jerome TA-10 

Monroy, Luisa MD-43 

Monsuur, Herman TD-45 

Montalvo Arango, Idel WA-20 

Montana, Giovanni WD-23 

Monteiro, Helena MA-22 

Monteiro, José TF-47 

Monteiro, M. Teresa Torres TC-48 

Monteiro, Paulo MC-40 

Montella, Bruno WB-16, WA-17 

Montemanni, Roberto WC-25 

Montero, Javier WC-21, MB-37 

Montevechi, José Arnaldo TC-32 

Montibeller, Gilberto MF-35, TA-35 

Montoya, Carlos WE-28 

Morabito, Reinaldo MF-34, TA-34 

Moradinezhad, D. MB-34 

Morais, A. Jorge WE-20 

Morais, Paulo TF-06 

Morales Matamoros, Oswaldo WB-39 

Moreau, Yves MA-24 

Moreira Junior, Fernando MD-38 

Moreno, Luis MA-03 

Moreno, Placido TD-06 

Moreno-jimenez, José María MF-12, 

TC-37 

Moreno-Pérez, José A. TF-13 

Moretti, Elena TA-47 

Moretti, Stefano TF-26 

Morita, Hiroshi TA-07 

Morita, Shuhei MA-19 

Morozova, Elena TB-09 

Morsi, Antonio WD-35 

Morton, Alec MA-30, MC-30, MB-44 

Moscardini, Alfredo WB-13 

Moser, Albert WC-33 

Moser, Robin TA-29 

Mosquera Rodríguez, Manuel Alfredo 

TA-26 

Mottl, Vadim WC-23, WD-23 

Moura, Ana MC-20, MA-21 

Moura, Pedro MA-02 

Mourão, Cândida WC-15 

Mourão-Miranda, Janaina MD-26 

Mourinho, João MB-23 

Mourtos, Yiannis MB-10, MF-10, 

TC-10 

Mousa, Abdelrahim MB-19 

Mousa, Mohammad MB-19 

Mousseau, Vincent MC-23, WE-37 

Moutari, Salissou TD-39 

Moutinho, Alexandra TC-33 

Mouysset, Sandrine TF-18 

Movahedi, Mohammad Mehdi TD-06 

Moz, Margarida MF-16, TB-23, WB-24 

Mruczkiewicz, Wojciech MD-24 

Muñoz, Jose MF-13, TB-13 

Muñoz, M. Pilar MB-28 

Mucherino, Antonio MC-02 

Muchnik, Ilya WC-23, WD-23 

Mueller, Marlene WC-46 

Mues, Christophe WA-20, WB-20, 

WC-30 

Muga, André TF-32 

Muguerza, Javier MC-26 

Mukhopadhyay, Somnath WD-38 

Mula, Josefa MB-36 

Mulder, Harm TC-45 

Mulvey, John MA-25 

Munapo, Elias WC-20 

Mund, Inbal MD-19 

Muniz, José Norberto TD-04 

Mur Torrentó, Rubén Javier MA-39 

Murat, Alper TA-13, MB-15 

Muraviev, Roman MB-27 

Murray, Alan MA-13 

Murthy, Ishwar TF-09 

Murukan Mypalli, Rishil MF-21 

Muselli, Marc WD-37 

Mustajoki, Jyri MA-43 

Mutlu, Esra TC-38

Mutzel, Petra MA-02 

Muyldermans, Luc TC-04, WA-11 

Muzzioli, Silvia WB-30 

Myllyviita, Tanja TF-30 

Myndyuk, Olga TF-46 

N 

Nabona, Narcis WC-19 

Nacima, Labadi MB-03, MC-03, 

WC-15 

Nagaoka, Sakae TB-27 

Nagaraj, Guruprasad MA-22 

Nagarajan, Mahesh TB-42 

Nagel, Andrea MA-17 

Nagy, Gábor TC-17 

Nagy, Mariana MD-22 

Najid, Najib. M. WD-39 

Nakagawa, Masashi WD-11 

Nakagawa, Yuji WC-09 

Nakai, Toru MB-45 

Nalca, Arcan MD-14 

Namboothiri, Rajeev MF-17 

Nandy, Debaprosanna MF-07 

Naoum-Sawaya, Joe MD-02 

Napalkova, Liana WC-43 

Nar, Fatih WE-05 

Narenji, Masoud MC-12 

Nascimento, João TF-47 

Nascimento, Mariá C. V. WB-05 

Nash, Jr., John F. ME-01 

Nasrabadi, Ebrahim MB-09 

Nasser-Carvalho, Luiz Felipe TC-35 

Nassiri, N. WC-44 

Natalia, Djellab WB-18 

Nattero, Cristiano WC-25 

Naumann, Marc MD-16 

Naumova, Mariya TF-46 

Naundorf, Jessica MA-36 

Navabi, M. TD-39, WC-44 

Navas, Jorge WC-14 

Naves, Guyslain WD-25 

Nazari, Asef TB-09, MF-25 

Nazarov, Denis MC-20 

Nazif, Habibeh MB-03 

Néron, Emmanuel MD-48 

Neboian, Andrei MC-33 

Nediak, Mikhail MB-41 

Nembhard, David WD-24 

Nemery, Philippe MB-44 

Nemeth, Sandor Zoltan MC-09 

Nemirovski, Arkadi MA-01 

Nenes, George TB-05 

Neralic, Luka TC-07 

Nesterov, Yurii WB-40 

Neto, Teresa MC-32 

Neumann, Ludmila MF-06 

Newman, Alexandra TB-21 

Newton, Louise TD-22 

Ng, Chi To MC-38 

Ng, Chi-Kong TD-43 

Ng, Kien-Ming WE-24 

Ng, Suk Fung MF-16 

Ngoc Khanh, Nguyen MA-31 

Ngueveu, Sandra Ulrich MD-15 

Nguyen Duc, Manh MF-45 

Nguyen, Giang MD-18, MF-18 


Nguyen, Viet Phuong MA-15 

Nguyen, Vu Duc TF-40 

Nickel, Stefan WE-13 

Nicolau, João TA-44 

Nicosia, Giuseppe MF-24 

Niedziółka, Jerzy W. TA-32 

Nielsen, Kurt WE-43 

Nielsen, Lars MA-16 

Nielsen, Lars Relund MF-32 

Niemi, Jarkko TA-32 

Niessner, Helmut MF-22 

Niknejad, Ali WE-06 

Nikolai, Tchernev WA-08 

Nikolakakou, Christina D. TD-48 

Nikulin, Yury WA-44 

Nilsson, Stefan TA-05 

Nishihara, Michi MB-27 

Nishimura, Etsuko TB-17 

Nisse, Nicolas WD-42 

Nitsch, Daniela MA-24 

Niu, Yi-Shuai MF-45 

Nizamitdinov, Akhlitdin WC-17 

Noailles, Joseph TF-18 

Noël, Jean-François WD-37 

Nobbe, Christoph WB-45 

Nobili, Paolo MA-10 

Noorollahi, E. MB-04, MB-34 

Norese, Maria Franca WB-37 

Nossack, Jenny MB-08 

Nouioua, Karim WE-05, WC-25 

Nouno, Gilbert MC-29 

Nova, Tertia Delia TB-24 

Novaes, Antonio G.N. TD-06, TF-17 

Novais, Augusto MB-04, WC-18 

Novak, Andreas MA-39 

Novello, Chiara WB-37 

Novo, Vicente TF-44, WA-44 

Novokmet, Novak MD-18 

Nowak, Agnieszka K. TB-06 

Nowak, Ivo MB-16 

Nowak, Tomasz MD-24 

Nowe, Ann TF-28 

Noyan, Nilay TF-04, MA-25, TA-28, 

MB-45, MC-48 

Nozari, Milad WE-41 

Numanagic, Ibrahim WD-25 

Numminen, Emil MB-25 

Nunes Vicente, Luís WC-21 

Nunes, Alcina WE-21 

Nunes, Ana Catarina WC-15 

Nunes, Eusebio TD-36 

Nunes, Luís TF-47 

Nunes, Manuel TA-18, TA-21 

Nunez, Marina TC-26 

Nurmi, Kimmo TF-24 

Nutt, Paul MB-35 

Nygreen, Bjørn TC-15 

Nyström, Christina MB-11, TC-44 

O 

O’Brien, Brendan WB-30 

O’Brien, Frances MA-35 

O’Hanley, Jesse MC-13, TC-13 

O’Sullivan, Barry WD-25 

O‘Kelly, Morton E. MD-37 

Obersteiner, Michael WC-32 

Oberti, Pascal WD-37 

Odegaard, Fredrik MB-22, TC-22 

Odile, Pourtallier TD-38 

Oggioni, Giorgia WA-33 

Oguz, Ceyda MA-04 

Oguz, Saziye Deniz TD-24 

Ohya, Takao WD-12 

Oikonomidis, Anastasios TF-10 

Ojalehto, Vesa TD-23 

Oktem, Hakan TD-24 

Okuno, Takayuki MF-46 

Olaso, Pablo WB-16, MA-28 

Olivares, Alberto TA-21, MC-39 

Oliveira Moreira, Mayron César TD-04 

Oliveira Soares, João WB-30, TD-37 

Oliveira, Amilcar MA-22 

Oliveira, Arlindo MF-29 

Oliveira, Aurelio WC-40, MB-48, 

TF-48 

Oliveira, Bruno M.P. M. MD-19 

Oliveira, Fabrício MB-33 

Oliveira, Fernando TC-33, TB-43, 

TB-46 

Oliveira, José WD-08, MA-13, TF-15, 

WA-31 

Oliveira, Jose Fernando MF-20 

Oliveira, Lucas MD-06 

Oliveira, Luis MF-33 

Oliveira, Manuela Maria WA-06 

Oliveira, Marisa MB-20 

Oliveira, Mónica TD-30, TB-47 

Oliveira, Pedro TC-48 

Oliveira, Rui TF-37 

Oliveira, Sandra Cristina WE-41 

Oliveira, Teresa MA-22 

Olivella, Jordi WD-24 

Olsson, Fredrik WC-38 

Olsson, Hanna MB-11 

Omari, Jawad MD-03, WB-31 

Omelchenko, Vadym TD-46 

Omerbegovic-Bijelovic, Jasmina MF- 

14, MD-33 

Omont, Nicolas TA-37 

Onal, Mehmet WB-38 

Oner, Kurtulus Baris WB-11 

Onggo, Stephan TB-05 

Onoda, Takashi WD-47 

Onsel, Sule MA-21, TF-42 

Or, Ilhan TB-33, WD-47 

Orbak, Ali Yurdun TC-36 

Orbay, Berk MB-05 

Ordin, Burak WE-17 

Ordonez, Fernando TA-05, TA-16, 

MC-18 

Oreskovich, Anne-marie TD-08 

Orlin, Jim MD-29 

Orlov, Alexej WB-47 

Orlovich, Yury TA-40 

Ormeci, E. Lerzan TC-14 

Orsenigo, Carlotta TA-42 

Ortega, Eva Maria TD-24 

Ortega, Francisco A. MD-13, TA-16 

Ortega, Isabel TD-24 

Ortega, Manuel MC-26 

Ortega-Ardila, Joan-Manuel TC-04 

Ortigosa, Pilar M. TB-13, TC-43 

331


Ortobelli, Sergio WC-29 

Ortuno, M. Teresa MB-37 

Oruc, Sercan WD-44 

Osuna, Edgar Elias TD-42 

Osuna-Gómez, Rafaela TF-44, WB-44 

Otola, Iwona MF-37 

Ouelhadj, Djamila TD-46 

Ouerdane, Wassila MA-43 

Ouorou, Adam WB-35 

Ovchinnikov, Anton MB-41 

Ozaydin, Ozay MD-12, MA-21 

Ozbas, Birnur WD-47 

Ozcan, Ender WA-25, WE-25, TD-46 

Ozdemir, Ozer WC-17, TB-36 

Ozdemir, Oznur MC-41 

Ozdemirel, Nur Evin WB-04, WD-17 

Ozdinc, Selin TD-13 

Ozel, Gizem TD-09 

Ozen, Ulas TB-14, MB-34 

Ozener, Okan MA-17 

Ozer, Ihsan WD-40 

Ozgün, Onur TD-19 

Ozkan, Sevgi MF-11, TF-11 

Ozkir, Vildan WC-40 

Ozluk, Ozgur MB-41 

Ozmutlu, Seda TC-36 

Ozogur-Akyuz, Sureyya MF-05, 

MB-26, MD-26, MA-31 

Ozsen, Leyla WB-46 

Ozsoy, Aykut MB-40 

Ozsoy, Aysegul WE-21 

Ozsoydan, Burcin TA-09 

Oztekin, Atakan MC-12 

Ozturk, Buse MA-31 

Ozturk, Gurkan MA-26 

Ozturk, Mercan MF-33 

Ozturk, Pinar MA-09 

P 

Pas, David TC-04 

Pacciarelli, Dario WC-16 

Pacheco, Joaquín WB-15 

Pacheco, Maria F MB-07 

Padoano, Elio TD-12 

Pagnoncelli, Bernardo TB-28 

Pahikkala, Tapio TD-29 

Pai, Fan-Yun WA-10 

Paias, Ana WD-15, MF-16 

Paillard, Gabriel MC-02 

Paillat, Jean-Marie MD-44 

Pais, Tiago WE-24 

Paiva, João TB-42 

Paiva, Renato MF-34 

Paixão, José MB-23, TF-32, MA-44 

Palacios, Daniel WE-26 

Pales, Zsolt TF-34 

Paletta, Giuseppe MC-04 

Palma dos Reis, António TD-21 

Paltayian, George WA-12 

Pan, Quan-Ke WE-33 

Panagiotidou, Sofia TB-05 

Pandelis, Dimitrios TB-39 

Pang, Gu TC-04 

Pang, James Z. MD-41 

Panic, Biljana TC-05, TA-18 

Pankratov, A. MA-20 

332 

Pankratova, Yaroslavna WE-26 

Pankratz, Giselher MA-17, TF-17 

Pantelic, Ognjen TC-05 

Paolotti, Luisa WA-37 

Paolucci, Massimo WC-25, TC-27 

Papadaki, Katerina TB-25, TF-26 

Papadopoulos, Christos TA-24 

Papageorgiou, Lazaros TF-35 

Papalamprou, Konstantinos WE-35 

Papanagnou, Christos TA-18 

Pape, Susanne MB-24 

Papoila, Ana MA-11 

Paquete, Luis MA-23 

Paquette, Julie TB-23 

Parada, Víctor MF-02 

Paradi, Joseph TB-06 

Paraskevopoulos, Dimitris WB-31 

Pardalos, Panos TC-10 

Pardo, Diego MC-26 

Paredes, Paula TF-32 

Parente, Manuel WE-14 

Park, Jonghyuck WA-11 

Parker, Kim MC-22 

Parkes, Andrew J. MF-02, WE-05 

Parlar, Mahmut MB-41 

Parragh, Sophie TB-23 

Parreira, Telmo TB-19 

Parreno, Francisco MC-20 

Pasalic, Enes MD-29, MF-29 

Pascoal, Marta TB-23 

Pasichny, Alexis WA-47 

Patay, Emmanuelle TC-42 

Patie, Pierre WD-14 

Pato, Margarida MB-07, MF-16, MA- 

22, TB-23, WB-24, TD-28 

Patrício, Pedro MC-40 

Patriksson, Michael WD-28 

Patrizi, Giacomo MB-27 

Patrone, Fioravante TF-26 

Paucar-Caceres, Alberto MD-44 

Paulo, Octavio MA-05 

Pauls-Worm, Karin WC-38 

Pavón Mendoza, Ruth TB-32 

Pavlovic, Ljiljana WA-09 

Pavone, Mario MF-24 

Pawlak, Grzegorz TF-07 

Pawlak, Krzysztof TA-32 

Pérez Sánchez, Carlos Javier MA-43 

Pérez, M. Angeles WA-39 

Pérez-Hurtado, Ignacio MF-24 

Pérez-Jiménez, Mario J. MF-24 

Peacock, Stuart MB-30 

Peccati, Lorenzo TF-05 

Pecin, Diego MD-15 

Pedro, João MC-40 

Pedro, Sílvia TB-24 

Pedroso, João Pedro MF-04, WB-05, 

MC-32 

Pedroso, Joao Pedro WB-29 

Peharda, Irena TC-45 

Peharda, Petra WD-07 

Pehlivan, Canan MA-47 

Peidro, David MB-36 

Peinhardt, Matthias MB-29 

Peixoto Santos, Cristina TB-48 

Peixoto, José Antonio TB-31 

Pekár, Juraj MC-03, WB-13 

Pekec, Sasa MA-43, MD-43 

Peker, Can MF-11 

Pekkurnaz, Didem WD-47 

Pelikan, Jan WC-28 

Pelizzari, Cristian WA-33 

Pellegrini, Paola TD-05 

Pena, Gloria MA-03 

Peraza, Edgar WE-36 

Perea, Federico MD-16, TA-16 

Pereira de Souza, Marcus Vinicius WB- 

06 

Pereira, Basílio MA-11 

Pereira, Fabio WB-28 

Pereira, Fernando MC-23, TC-39 

Pereira, Gonçalo TC-33 

Pereira, Guilherme MD-14 

Pereira, Ivo WC-31 

Pereira, Jorge MA-13, TF-15 

Pereira, Sandrina MB-33 

Pereira, Teresa MD-30 

Pereverza, Kateryna WA-47, WB-47 

Perez Gonzalez, Paz MA-04, WC-07, 

WC-39 

Perez, Garcia MB-20 

Perez, Gerardo MC-42 

Perez, Gloria MA-28 

Perez, Marcela MC-37 

Peris-Blanes, Jordi TB-12 

Perić, Tunjo TA-18 

Perko, Igor TD-44 

Perona, Iñigo MC-26 

Perrot, Nancy TA-10 

Persona, Alessandro MD-04 

Pesch, Erwin MB-08, WC-36 

Pesneau, Pierre TB-15 

Pesquita, Catia MC-24 

Pessoa, Artur MD-15 

Pessoa, Leonardo MF-06, MC-35 

Peter, Andrea TB-08 

Peter, Foldesi WA-46 

Petersen, Bjørn MF-15, TD-15, WB-25 

Petrelli, Marco WA-16 

Petridis, Ilias TC-38 

Petrovic, Dobrila TD-28 

Petrovic, Sanja TD-22 

Petrovic, Slavica P. MC-31 

Peyghami, Mohammad Reza WD-36 

Peypouquet, Juan MA-07 

Pflug, Georg WD-29 

Pham Dinh, Tao MD-45, MF-45, TA-45 

Pham, Hang Phuong MB-24 

Phan, Raksmey MB-03 

Pi, Wei-Ning TF-18 

Pichler, Alois WD-29 

Pickenhain, Sabine TA-39 

Pickering, Larry TA-05 

Picouleau, Christophe MD-10 

Pidd, Michael TB-05 

Pietrasz, Slawomir WC-37 

Pimentel, Carina WC-05, MB-34 

Pina Martins, Francisco MA-05 

Pina, Joaquim WB-30 

Pinar, Mustafa MD-40 

Pindoria, Sandip MD-21, TC-21 

Pinheiro, Diogo MC-19

Pinho de Sousa, Jorge MD-03, MF-04, 

MF-16 

Pinto Ferreira, M a Eduarda MB-20 

Pinto, Alberto TC-19 

Pinto, Alberto A. MB-19, MC-19, MD- 

19, TA-19, TB-19, TD-19 

Pinto, Telmo MA-20 

Pinto_Varela, Tânia MB-04, WC-18 

Pires, Ana TB-24 

Pires, Cesaltina MC-38 

Pires, João MC-40 

Pirilä, Pekka TF-33 

Pirkwieser, Sandro TF-25 

Pisinger, DavidMB-15, MD-15, WD-22 

Pistikopoulos, Efstratios TF-35 

PIta, James TA-05 

Pitsoulis, Leonidas MD-05, WE-35 

Piyade, Nuray MC-13 

Piyatumrong, Apivadee MB-23 

Pizarro Romero, Celeste TC-05, 

WB-16, MA-28 

Pizzolato, Nelio D MD-37 

Pla, Lluis Miquel WA-32 

Pla, LluisM MF-32 

Pla-Santamaria, David WD-30 

Plana, Isaac TB-15 

Pliska, Stanley MC-19 

Pliskin, Joseph WA-36 

Plyasunov, Alexander MB-05, MA-42 

Poggi de Aragão, Marcus MD-15 

Pohlmann, Tobias TD-27 

Pokutta, Sebastian TB-08, MB-24, 

WA-35, WD-35, TD-38 

Polat, Olcay TD-17 

Poli, Ange-Michel WC-17, WD-37 

Polyakovskiy, Sergey MF-10 

Polykarpou, Eleftheria TF-35 

Ponce-Cueto, Eva TA-27 

Ponomareva, Ksenia WD-46 

Porcar-Ramos, Alfonso TB-12 

Portela, Maria TC-06 

Poss, Michael MB-40, MD-40 

Postmus, Douwe TB-18, MB-30 

Posypkin, Mikhail MD-08 

Potra, Florian MC-09, MA-40 

Potthoff, Daniel TC-16 

Potts, Chris MC-08, WC-16, WD-22, 

WE-35 

Potvin, Jean-Yves MC-17 

Powell, Warren TF-08 

Pozo, Miguel Angel MD-13, TA-16 

Pradenas, Lorena MF-02 

Pralat, Pawel WE-42 

Prandtstetter, Matthias TF-25 

Pranzo, Marco MC-04, WC-16 

Prasanna, G. N. Srinivasa TB-18, 

MF-26 

Prata, Bruno MF-16 

Príncipe Anticona, Santos Valerio MC- 

37 

Prekopa, Andras TF-46 

Preusser, Tobias MF-46 

Prigent, Jean-luc MC-07 

Prins, Christian MC-03, MF-03, TD-03, 

MA-15, MD-15 

Prnaver, Katja MB-15 


Prodhon, Caroline TD-03, MA-15 

Prot, Damien WB-08 

Pruyt, Erik MF-44 

Psaromiligkos, Yannis TD-37 

Puente, Alejandra WC-45 

Puerto, Justo MD-13, WE-13, WE-36 

Pukkala, Timo MD-32 

Pulido Cayuela, Manuel A. TB-26 

Pulkkinen, Seppo WD-19 

Punkka, Antti MF-06 

Pureza, Vitória MC-17 

Puro, Lauri MD-43 

Q 

Qian, Jiani MC-15 

Qiu, Xian TC-02 

Qu, Shuo TD-08 

Quadri, Dominique MF-02 

Qualizza, Andrea TA-02 

Quek, Ser Aik TA-22 

Queyranne, Maurice WA-24 

Quintanilla, Sacramento WA-39 

Quttineh, Nils-Hassan TF-42 

R 

R. M. da Costa, Geraldo WC-20 

Raack, Christian TF-31 

Rabieh, Masood WA-10 

Rabinowitz, Gad TC-08 

Racz, Anett WB-36 

Rada, Miroslav MF-20 

Radjef, Mohammed Said WA-30, 

TD-48 

Radovanov, Boris MA-38 

Radovilsky, Zinovy TB-27 

Rafels, Carles TC-26, TD-26 

Rafferty, Martin TA-41 

Rafiee, Majid TA-28 

Raghavan, S. MB-40 

Ragwitz, Mario TF-33 

Raharjo, Hendry TD-22 

Rahmani, Masoud TA-33 

Raidl, Günther TF-25 

Rais, Abdur TC-22, WB-29 

Rakicevic, Zoran MF-14, MD-33 

Rakke, Jørgen Glomvik WC-22 

Ralph, Daniel WA-33 

Ralphs, Ted MF-17 

Ram, Camelia TA-35 

Ramalho, Esmeralda MB-06 

Ramalho, Joaquim MB-06 

Ramalho, Ruben TC-33 

Raman, Rajiv MC-10 

Ramón, Nuria MC-06 

Ramik, Jaroslav TA-36 

Ramirez, Carlos WB-21 

Ramlan, Rohaizan WA-27 

Ramond, François TF-16 

Ramos, Ana Luísa WA-27, TA-44 

Ramos, Francisco TF-06 

Ramos, Sofia B. WC-29 

Ramos, Tânia TC-17 

Ramos-Gallego, Ana Bel WE-13 

Rana, Rupal TB-43 

Range, Troels Martin WC-24 

Rückmann, Jan-J. MB-42 

Rapine, Christophe WB-08, WE-36 

Rastegar, Hasan TA-33 

Rathi, Shyamsunder TA-05 

Ratkovic, Branislava WA-13 

Ratprasert, Pasu TF-04 

Rauner, Marion MF-22 

Raupp, Fernanda WD-04 

Rave, Claudia MC-37 

Ravi, R. TD-13 

Ravizza, Stefan MC-08 

Ray, Bonnie WE-14 

Ray, Saibal MD-14 

Rayée, Gregory WC-46 

Rayee, Gregory WC-46 

Raz, David MD-20 

Rømo, Frode MC-28 

Ráčková, Adéla MF-37 

Rönnqvist, Mikael TA-05, WB-31, 

MA-32, MB-32 

Rebai, Ahmed MA-37 

Rebaine, Djamal WB-39 

Rebelo, Cecilia MA-22 

Rebennack, Steffen TC-10 

Recchioni, Maria Cristina TF-44 

Recht, Peter TA-40 

Refenes, Apostolos MF-25, MB-38 

Rego, César WC-09 

Rego, Leandro MB-19 

Rei, Constantino MC-22 

Rei, Walter MF-03, TF-15, TA-17 

Reichart, Christiane TB-22 

Reichmann, Gerhard TA-07 

Reiter, Ulrich TF-33 

Reizins, Toms WA-42 

Rejec, Valter TB-44 

Relvas, Susana MD-47 

Renaud, Arnaud TA-37 

Repoussis, Panagiotis TD-15, WB-31 

Requejo, Cristina WE-22, TA-40 

Respicio, Ana TB-23 

Reston Filho, José Carlos TA-04 

Restrepo, Juan Esteban MC-37 

Retel Helmrich, Mathijn WB-38 

Reti, Shane MD-36 

Reuther, Markus MA-16 

Reveco, Carlos WE-18 

Revollar, Silvana TC-48 

Rey, David WE-36 

Rey, Pablo A. WE-15, WD-24 

Rezaei, Shaghayegh TA-20 

Rezapour, Shabnam MF-14 

Riane, Fouad MF-22 

Ribas, Imma WC-07 

Ribeiro, Ana TB-39 

Ribeiro, Glaydston MD-47 

Ribeiro, Isabel MB-42 

Ribeiro, Luís TF-47 

Ribeiro, Rita TC-03, TD-37, WB-42 

Ribeiro, Simão MA-13, TF-15 

Riccardi, Rossana WA-33 

Ricci, Elena Claire TB-33 

Ricciardi, Nicoletta WA-13 

Richtarik, Peter WB-34 

Rieck, Julia MA-17 

Riera, Daniel MD-03 

Ries, Bernard MC-10 

333


Ries, Jana TD-05 

Riesgo, Laura TC-32 

Rietz, Jürgen TA-02, TB-02, MA-20 

Righini, Giovanni TC-17 

Rinaldi, Franca MB-02 

Rinaldi, Francesco TD-43, TA-48 

Rincon-Zapatero, Juan Pablo WB-14 

Ring, Andreas TC-44 

Rios-solis, Yasmin WA-34 

Rios-Zalapa, Ricardo TD-38 

Rivera, Juan Carlos TB-04 

Rivero, Carlos WB-17 

Rivreau, David WE-28 

Roberti, Roberto MB-02, MD-15, 

MF-15 

Robinson, Stewart TD-35 

Rocca, Matteo TF-34 

Rocchi, Lucia WA-37 

Roch, Oriol WC-14 

Rocha, Humberto WC-21, WE-35 

Rocha, Paula TF-36 

RodrÍguez-puerta, Inmaculada MD-19 

Rodríguez-Chía, Antonio Manuel TC- 

05, WE-13, TB-15 

Rodríguez-Sánchez, Sara Verónica MF- 

32 

Rodrigues, Ana Maria WC-15 

Rodrigues, Antonio TC-09, TD-18, 

TA-19 

Rodrigues, Carlos Diego MF-02 

Rodrigues, Helena Sofia TC-48 

Rodriguez Uria, Maria Victoria WD-30 

Rodriguez, J. Tinguaro MB-37 

Rodriguez-Chia, Antonio Manuel MD- 

13 

Rodriguez-Ravines, Romy-Elena MD- 

28 

Rodriguez-Uria, Maria Victoria MC-36 

Roidl, Moritz WA-45 

Rojas-Medar, Marko A. WB-44 

Roknedin Eftekhari, Reza WA-12 

Roland, Benoît MF-22 

Roland, Julien MD-30 

Rolando, Diana WB-37 

Roman, Diana MA-27 

Romanin-Jacur, Giorgio TB-47 

Romanova, Tatiana MA-20 

Romeijn, Edwin TC-22, WB-38 

Romero Morales, Dolores MA-41 

Romero, David MC-02 

Romeu de Carvalho, Fábio MF-26, 

TD-36 

Rommelfanger, Heinrich MA-36 

Ronconi, Débora TF-08 

Roncoroni, Andrea WE-29 

Rondier, Pierre WC-37 

Rong, Aiying WC-08 

Roos, Kees WB-34 

Roos, Pontus TC-22 

Ropke, Stefan TB-17 

Roque, Luís TA-23 

Rosa, Ana MC-22 

Rosa, Maria J. WD-21 

Rosa, Silvério MB-42 

Rosário, João TD-21 

Rosenberg, Ivo MC-29 

334 

Rosendahl, Knut MC-11 

Rosenthal, Edward TC-18 

Roshdi, Israfil TF-09 

Rossi, André WD-05, WE-37 

Rossi, Dario TF-26 

Rossi, Riccardo WD-27 

Rossi, Roberto MB-34, WC-38 

Rostamy-Malkhalifeh, Mohsen TD-36 

Rostamy-malkhalifeh, Mohsen MA-06, 

WA-10 

Rotaru, Kristian TA-41 

Rothblum, Uriel G. MD-19, TB-45 

Rougier, Jean-Louis TF-26 

Roussado, Jorge TC-16 

Rousseau, Louis-Martin TD-15, TA-17, 

WC-25 

Rousval, Benjamin WB-37 

Rouwette, Etienne TC-35 

Roy, Bernard MC-23 

Rozenkop, Antoine WD-25 

Rozman, Črtomir TB-44 

Rozycki, Rafal WC-39 

Rubinov, Alexander TF-34 

Rudec, Tomislav WD-28 

Rudolf, Gabor MD-29 

Rudzkien, Vitalija WA-43 

Rufián-Lizana, Antonio TF-44, WB-44 

Rufo Bazaga, María Jesús MA-43 

Ruhland, Johannes TF-18 

Ruiz, Ana Belen TD-23 

Ruiz, Bernardo WC-19 

Ruiz, Daniel TF-18 

Ruiz, Francisco TD-23 

Ruiz, Jose L. MC-06 

Ruiz, Ruben WE-33 

Ruiz-Garzón, Gabriel TF-44 

Ruiz-Gazen, Anne MD-05 

Ruiz-Tagle, Mauricio MA-12 

Rupérez Micola, Augusto TB-46 

Rustem, Berc TF-36 

Rusu, Alin TD-25 

Ruzika, Stefan MA-23, MB-46 

Rychlewski, Jeremi TC-16 

S 

S. Pereira, Luis TF-32 

S.Pinto, Leonor WC-15, MC-46 

Saameño, Juan José TB-13 

Saban, Daniela MD-29 

Saber, Ibrahim WC-04 

Sackett, Peter MC-24 

Sacone, Simona TC-27 

Saddoune, Mohammed WC-24 

Sadeh, Arik MD-20 

Sadjadi, Seyed Jafar WC-09 

Sadjadi, Seyed jafar MB-03 

Sadoghi, Amirhossein WD-45 

Sadoune, Mohamed WB-31 

Sadres, Natanel TC-42 

Sadykov, Ruslan MD-08, WB-09 

Safe, Martin MD-29 

Safia, Kedad-sdihoum WB-07 

Sagastizabal, Claudia MD-40 

Sagir, Mujgan TA-09 

Saharidis, Georgios WA-40 

Sahin, Guvenc TB-03, TA-27 

Sahin, Mustafa TB-03 

Sahling, Florian MD-34 

Sahlström, Leena TA-32 

Sahnoun, Zaidi WA-21 

Sahut, Jean-Michel TB-38 

Saidani, Nasreddine TF-13 

Saikouk, Tarik WA-10 

Saito, Susumu TA-38 

Sakalauskas, Leonidas WC-43 

Salakoski, Tapio TD-29 

Salazar González, Juan José MA-02 

Saldanha, Ricardo TC-16 

Saldanha, Rodney TD-48 

Saldanha-da-Gama, Francisco MB-13, 

WE-13 

Salhi, Said TC-17 

Saliaris, Dimitrios TA-17 

Saliba, Sleman TB-21 

Salima, Amrouche WA-18, MD-28 

Salimifard, Khodakaram WE-24 

Salles, André TC-19, MC-27 

Salman, Sibel MA-04, TD-13, TF-15, 

MA-46 

Salmeron, Javier WE-22, TC-45 

Salmeron, Jose L. TD-06 

Salmon, Ioannis TD-37 

Salo, Ahti MF-06, MA-30, MA-44 

Salomon, Valério TC-32 

Salouras, Yiannis TB-08, WA-42 

Salvador, Manuel MF-12 

Sama, Miguel WA-44 

Samarah, Rasha MB-19 

Samà, Albert MC-26 

Sami, Nadia TF-30 

Samii, Behzad TA-43 

Sammarra, Marcello TC-27 

San Matías, Susana MB-38, TA-48 

Sanatifar, M. TD-39 

Sanchez, Elena MF-07 

Sanchez, Francisca MA-38 

Sanchez, Reinaldo WB-21 

Sanchis, Jose Maria TB-15 

Sancho, Julia TB-26 

Sand, Guido TB-21 

Sandrini, Francesco WE-30 

Sanguineti, Marcello MF-40 

Santana, Roberto TA-42 

Santelices, Hugo TB-32 

Santoro, Miguel Cezar TD-20 

Santos Silva, Carlos TD-37 

Santos, Ademilton MB-44 

Santos, Andréa C. TC-04 

Santos, Anesio TF-48 

Santos, Dorabella WC-05, MB-40 

Santos, Eulália TA-40 

Santos, João MC-40 

Santos, Jorge TD-21 

Santos, José MB-23 

Santos, Lana Mara TB-32 

Santos, Maristela MD-34 

Santos, Nicolau MF-04 

Santos, Paula MF-19 

Santos, Ricardo Henrique TB-32 

Santos-Peñate, Dolores R. TF-13 

Saraç, Tugba WC-09 

Sarabando, Paula TB-30

Sarfi, Fatemeh WD-12 

Saricicek, Inci MF-26 

Sarioglan, Mehmet MF-37 

Sarkar, Uttam MC-25 

Sarmento, Manuela MF-19 

Sarmiento, Miguel TC-07 

Sarrico, Cláudia S. WD-21 

Sashima, Akio TD-08 

Sassano, Antonio MA-10 

Satiroglu, Sait TC-25 

Sato, Kimitoshi MD-41 

Sauvanet, Gaël MD-48 

Savariradjou, Alexandre MF-17 

Savelsbergh, Martin MA-17 

Savic, Gordana WE-15, MD-18 

Savin, Sergei MD-41 

Savourey, David WB-25 

Sawaki, Katsushige TB-38, MD-41 

Sá Esteves, Jorge WA-34 

Sánchez-Soriano, Joaquin TA-26, 

TB-26 

Sbihi, Abdelkader MA-07 

Scaparra, Maria Paola MC-13, TC-13, 

MB-37, MF-48 

Scarpel, Rodrigo MF-42 

Scarsini, Marco TD-26 

Schachtebeck, Michael WA-16 

Schaible, Siegfried WA-33 

Schaller, Jeffrey WB-28 

Schülldorf, Hanno TF-16 

Schönlein, Michael MC-14 

Schönung, Frank WB-45 

Scheerlinck, Karolien TD-04 

Scheffermann, Robert MD-17 

Scheidegger, Alexander MA-33 

Scheithauer, Guntram MF-20 

Schellenberg, Sven MC-05 

Scheller-Wolf, Alan WB-11 

Schellhorn, Henry WD-41 

Schemeleva, Kseniya WD-39 

Schenk-Mathes, Heike TB-43 

Schewe, Lars WD-35 

Schickinger, Thomas MB-16 

Schilde, Michael TC-15 

Schillings, Claudia WB-40 

Schimmelpfeng, Katja MD-34 

Schimpel, Ulrich WE-14 

Schirrmann, Arnd TF-04 

Schittkowski, Klaus MD-09 

Schlechte, Thomas TF-16, WA-25 

Schleich, Julien MD-45 

Schmid, Lukas TB-41 

Schmid, Verena MC-17 

Schmid, Wolfgang WD-29 

Schmidt, Daniel TD-27 

Schmidt, Günter WA-29 

Schmidt, Martin MF-09 

Schmidt, Stephan WB-40 

Schneider, Frank WC-36 

Schneider, Michael MB-15, WA-15, 

MF-17 

Schoebel, Anita TA-13, WA-16 

Schoen, Cornelia MF-33 

Schoenberger, Joern MC-17 

Scholz, Peter WB-41 

Scholz, Roland W. MA-33 


Scholz-Reiter, Bernd MC-14 

Schröter, Marcus TC-41 

Schreck, Helmut TF-20 

Schuett, Holger TD-17 

Schultmann, Frank MF-13, WC-47 

Schultz, Rüdiger MF-09 

Schulz, Jens TA-08 

Schulz, Volker WB-40 

Schumacher, Andre WA-25 

Schwaderer, Frank MF-13 

Schwaiger, Katharina MA-27 

Schwaninger, Markus TC-41 

Schwind, Michael MB-15, MF-17, 

TF-17 

Schwindt, Christoph MA-08 

Sciomachen, Anna TC-27 

Scolnik, Hugo TC-09 

Scozzari, Andrea MA-27 

Scrimali, Laura WA-48 

Seabra, Sofia MA-05 

Secomandi, Nicola MC-27 

Seeanner, Florian MB-14 

Seedig, Hans Georg TC-02 

Seeland, Klaus WC-41 

Segura, Baldomero TC-32 

Segura, José Vicente TF-21, TD-25 

Seidl, Andrea TC-39 

Seifert, Ralf W. MB-35 

Seipp, Florian MA-23 

Seixedo, Cristina MC-12 

Seker, Sukran MA-07 

Sellami, Khaled WA-05, TC-18 

Sellami, Lynda WA-05, TC-18 

Selosse, Sandrine TA-37, TB-37 

Sels, Veronique MF-08 

Semal, Pierre MB-13 

Semet, Frédéric WC-13, MD-23 

Sen, Alper WD-26 

Sender, Julia TB-21 

Sener, Emrah TC-25 

Sengün, Erkan WB-41 

Senne, Edson WD-04 

Seo, Inyong MF-42 

Seow, Hsin-Vonn MB-03, WD-18 

Sepúlveda, Juan Pedro TB-04, TA-14 

Sequeda, Maria Teresa WC-44 

Serafini, Paolo MB-02 

Serôdio, Carlos WE-19 

Serebriany, Artem WC-27 

Sergeyev, Yaroslav TF-43 

Serra, Daniel TC-05 

Serrão, Amílcar MF-07 

Sert, Caglayan TD-47 

Sevaux, Marc WD-05 

Sevil, Guven WC-29 

Sevkli, Mehmet WB-10 

Seyhan, Tolga TF-13 

Sezerman, Ugur MB-26 

Sezgin, Emre MF-11, TF-11 

Sgalambro, Antonino WA-13 

Sgarbossa, Fabio MD-04 

Sgoutas, Georgios TD-14 

Shadalooee, Mahdi TA-20 

Shah, Nilay TC-43 

Shah-Hoseini, B. TA-04 

Shahi, Reza WE-43 

Shaked, Uri MC-39 

Shakhlevich, Natalia TD-08, WC-28 

Shalymov, Dmitry MC-18 

Shamsuddin, Faiz WA-27 

Shamsuzzoha, Ahm TC-11 

Shananin, Alexander WB-23 

Shangari, Nitish WD-24 

Shapiro, Alexander WD-02 

SharifYazdi, Mehdi WB-46 

Sharikova, Agata MC-25 

Sharma, Mamta TB-18 

Sharma, Megha TF-03, MF-40 

Shawe-Taylor, John MD-26 

Shekhar, Saurabh WD-24 

Shen, Liji WB-07 

Shen, Siqian WB-35 

Sherali, Hanif MF-45 

Shi, Mengze WB-38 

Shi, Xiaohui TC-46 

Shibata, Takashi TD-18 

Shih, Kuang-Husn MA-11 

Shih, Su-Chuan TC-12 

Shikata, Yoshiaki MB-11 

Shikhman, Vladimir MB-42, MD-46 

Shishebori, D. MD-14 

Shunei, Norikumo WD-12 

Sicilia, Joaquín WB-38 

Siddappa, Sheela MF-26 

Siebert, Johannes MF-30 

Siepak, Marcin MB-04 

Signorelli, Marcelo WC-40 

Silva, Amanda MB-44 

Silva, Aneirson TC-32 

Silva, Angela MF-06, WD-06, MC-35 

Silva, Bruno MF-34 

Silva, Carlos TC-33 

Silva, Dulce WA-09 

Silva, Elsa WC-05, TA-20 

Silva, Francesc TF-21 

Silva, Francisco TF-13 

Silva, Helder WD-08 

Silva, Pedro C. MB-07 

Silva, Rodrigo Cesar MD-06 

Silva, Vanina Macowski Durski TF-17 

Silverio Campos, MagnoWE-06, TB-34 

Sim, Melvyn MF-25 

Simões, Filipe TF-22 

Simões, Maria Lurdes MB-42 

Simões, Pedro TD-06, WB-06 

Simeone, Bruno MC-29 

Simoes-Marques, Mario TC-45 

Simon de Blas, Clara WC-06 

Simon Martin, Jose WC-06 

Simonato, Jean-Guy WC-33 

Simsek, Koray MA-25 

Sinclair-Desgagné, Bernard WD-33 

Singh, Gaurav TF-08 

Sinuany-Stern, Zilla WB-06 

Sipahi, Seyhan TD-10 

Sipahioglu, Aydin MF-04, TA-09 

Sirenko, Sergii WA-18 

Siri, Silvia TC-27 

Sirvent, Inmaculada MC-06 

Siskos, Yannis MF-30 

Six, Adrien MB-24 

Skantzos, Nikos WC-46 

335


Skocdopolova, Veronika MB-43 

Skraba, Andrej TB-44, TC-44 

Skums, Pavel TA-40 

Skutella, Martin MB-09 

Sladky, Karel WC-14 

Slijepcevic-Manger, Tatjana WD-43 

Slikker, Marco TB-14, WA-26 

Slimani, Hachem TD-48 

Slowinski, Roman TF-05, TD-29, 

WE-37 

Smedley, Jo WE-21 

Smedsrud, Morten WA-31 

Smeers, Yves WA-33 

Smirlis, Yannis MB-06 

Smirnov, Georgi TB-39 

Smirnova, Nadezhda WA-26 

Smith, Cole TC-22, WB-35 

Smith, Gareth TD-42 

Smith, Honora MD-37 

Smith, Olivia TB-27 

Snow, Miguel TF-37 

Snyder, Lawrence TF-13 

Soares de Mello, João Carlos MD-06, 

TC-06 

Socha, Leslaw WB-46 

Soeiro Ferreira, José WB-05, WC-15 

Sofianopoulou, Stella WD-06 

Sol, Ceyda MB-26 

Solar, Ricardo MD-03 

Soldic-Aleksic, Jasna MB-38 

Soler Arnedo, Manuel TA-21 

Soler, Edilaine WC-20 

Soler-Dominguez, Amparo WC-06 

Solis, Adriano WD-38 

Sologub, Roman TC-42 

Solyali, Oguz WB-22 

Soma, Nei Yoshihiro TD-20 

Sommersguter-Reichmann, Margit TA- 

07, MF-22 

Soner Kara, Selin TA-28 

Soner, Mete TC-25 

Soopramanien, Didier TB-05 

Soper, Alan WA-28 

Soria, Daniele MC-24 

Soriano, Patrick TA-22 

Sorrentino, Gregorio TC-27 

Sosic, Greys TB-14, TB-42 

Sosurcka, E.s. MD-20 

Sotaquira Gutierrez, Ricardo TB-41 

Soto, Maria WD-05 

Souad, Larabi Marie-Sainte MD-05 

Soudi, A. MB-04 

Soumis, Francois WC-24, 25, WB-31 

Sourd, Francis TC-08, TF-16 

Sousa Junior, Afonso Farias MF-18, 

MD-38 

Sousa Santos, Beatriz TC-23 

Sousa, João Carlos TF-32 

Sousa, Joao Miguel da Costa MD-36 

Sousa, Lisete TB-24 

Sousa, Sergio TD-36 

Souza, Filipe MB-19 

Souza, Geraldo TB-32 

Souza, Reinaldo WB-06 

Spagnolo, Nicola MA-27 

Spalanzani, Alain WA-10 

336 

Spengler, Thomas MA-36 

Speranza, M. Grazia TD-15 

Spieksma, Frits MF-10 

Spinler, Stefan MC-25, MC-33, TA-44 

Spoorendonk, Simon MF-15 

Spyridakos, Athanasios TD-37 

Sra, Suvrit WD-19, WA-34 

Srdic, Aleksander WB-43 

Staffetti, Ernesto TA-21, MC-39 

Stahlbock, Robert WD-22 

Stancu, Andreea Madalina TD-34 

Stancu-Minasian, Ioan TD-34 

Stanger, Sebastian TB-22 

Stangl, Claudia MF-09 

Stanojevic, Milan WE-15 

Stützle, Thomas TD-05 

Starcevic, Dusan TD-45 

Staritsyn, Maxim TC-39 

Stauffacher, Michael MA-33 

Stålhane, Magnus WB-22 

Stecco, Gabriella TC-27 

Stefani, Silvana WB-29 

Stefanov, Ivan TD-04 

Steinbach, Marc MF-09 

Steiner, Roman TF-25 

Steinhardt, Claudius MF-43 

Stenger, Andreas MF-17, TF-17 

Stentoft, Lars WC-33 

Stepan, Adolf MF-22 

Stephen, Tamon TB-29 

Sterna, Malgorzata TF-07 

Sternbeck, Michael MB-14 

Stewart, Theodor MA-30, MF-35 

Steyaert, Bart MD-18 

Stier-Moses, Nicolas TA-16 

Stiglmayr, Michael TD-24 

Stigter, Tibor TF-47 

Stojanovic, Marina WA-09 

Stokic, Dejan WB-23 

Stoklasa, Jan MF-36 

Stolletz, Raik TD-07, TB-22 

Stonebraker, Jeffrey MC-44 

Storchi, Giovanni WC-27 

Stosic, Biljana TB-11 

Stoyan, Yuri MA-20, MD-20 

Strahonja, Vjeran WA-30 

Strami, Stefano TD-12 

Strauss, Arne Karsten MA-41 

Strömberg, Ann-Brith WD-28 

Strbac, Goran TD-33 

Street, Alexandre TB-28, MB-33 

Strijov, Vadim WA-23 

Strode, Christopher TC-37 

Strusevich, Vitaly WA-28 

Studniarski, Marcin TF-44 

Stummer, Christian MA-30, MB-44, 

TC-46 

Sturt, Alexander TD-33 

Stygar, Anna MF-32 

Stylios, Giorgos TA-24 

Su, Chao-Ton WD-47 

Su, Shan-Yu TA-31 

Suarez-Vega, Rafael TF-13 

Subramaniam, Arjun MA-14 

Sucha, Premysl WD-07, WC-11 

Suclla Fernandez, Nicole TB-28 

Sufahani, Suliadi Firdaus MF-21 

Suhl, Leena WA-09, MD-16, TD-40 

Suhl, Uwe WA-09 

Summerauer, Wolfgang TA-08 

Sun, Hao WC-26 

Sun, Juo-Yi WA-42 

Sun, Li TC-31 

Sun, Shinn TA-07, TB-07, TC-07 

Sung, Ming-Chien WB-17 

Sural, Haldun WB-45 

Surikov, V. N. MC-18 

Suyabatmaz, A. Cetin TA-27 

Suykens, Johan TA-42 

Suzuki, Atsuo TB-38 

Suzuki, Keiji TD-08, WC-12 

Swarat, Elmar TF-16 

Swiercz, Aleksandra MD-24 

Szachniuk, Marta MD-24 

Szantai, Tamas WA-46 

Szczech, Izabela TD-29 

Sze, San Nah MF-16 

Szedmak, Sandor MD-26 

Szeto, Wai Yuen MF-03, TC-04 

Sznajder, Roman MC-09 

Szolgayova, Jana WC-32 

T 

Ta, Anh Son MD-45 

Tacettin, Mustafa TF-25 

Tadano, Yumi WC-12 

Tadross, Mark TF-47 

Tagaras, George TB-05 

Taheri, Seyyed Hassan WC-04 

Taillard, Eric WC-04 

Takada, Haruki WD-11 

Takahara, Shigeyuki MD-20 

Takahashi, Yoshitaka MB-11 

Takkula, Tuomo MF-21 

Tako, Antuela TD-35 

Taktak, Raouia TB-10 

Talasova, Jana MF-36 

Talasova, Zuzana MF-36 

Talbi, El-Ghazali TB-20, MC-23 

Talbi, El-ghazali WA-21 

Talebi, S. TA-34 

Talinli, Ilhan MC-12 

Talla Nobibon, Fabrice WA-07 

Tamarit, Jose MC-20 

Tambe, Milind TA-05 

Tambovceva, Tatjana WB-43 

Tambovcevs, Andrejs WB-43 

Tammer, Christiane TF-23 

Tamosaitiene, Jolanta WC-43 

Tanaka, Katsuaki TA-38 

Tanaka, Yuma TA-20 

Tanasescu, Cerasela TA-10 

Tancrez, Jean-Sébastien MB-13, MF-22 

Taner, Mustafa Egemen TD-17 

Tanfani, Elena TA-22 

Tang, Ling-Lang WB-12 

Tangil, Cem WD-40 

Tanino, Tetsuzo TC-09 

Tannert, Johannes TF-23 

Tanrioven, Emin Ahmet MA-46 

Tanyas, Mehmet WA-38, 39 

Tarantilis, Christos TD-15, WB-31

Tarapata, Zbigniew MC-23 

Tarashnina, Svetlana WA-26 

Tardella, Fabio MA-27 

Tarim, Armagan MB-34 

Tas, Duygu MB-15 

Tas, Engin MB-18 

Tas, Fatih WC-12 

Tatsumi, Keiji TC-09 

Tavares, Gabriel TF-20 

Taylan, Pakize WD-17, MA-26 

Taylor, Peter MD-18, MF-18 

Törnquist Krasemann, Johanna TA-27 

Ta¸skın, Z. Caner TC-22 

Tchemisova, Tatiana WC-21 

Tchung-Ming, Stéphane MC-11 

Teghem, Jacques MA-23, TA-23 

Teich, Jeffrey MD-43 

Teixeira, Ana Paula MA-22 

Teixeira, Senhorinha TA-21 

Tejeida Padilla, Ricardo WB-39 

Teles, João TF-35 

Telhada, Joao MD-02, WD-07, TF-24, 

TF-32 

Telhada, José MA-13, TF-15 

Tempelmeier, Horst MA-14 

Temur, Gül Tekin TB-36 

ten Hompel, Michael WA-45 

ten Raa, Thijs MF-06 

Tenorio, Janaina TF-06 

Teo, Kwong Meng WE-05 

Teramoto, Takeshi WD-11 

Tereso, Anabela PereiraTA-04, WD-08, 

MC-12, MD-12 

Terlaky, Tamas TC-09 

Tervonen, Tommi TB-18, MB-30 

Tesch, Christian MB-46 

Testi, Angela TA-22 

Teunter, Ruud WA-11 

Thanassoulis, EmmanuelTA-06, WE-06 

Thapalia, Biju K. TB-19 

Theofanis, Sotirios WA-40 

Thiao, Mamadou TA-45 

Thibault, Nicolas MD-10 

Thiele, Aurelie WB-10, MA-27 

Thiemann, Markus MB-46 

Thomas, Lyn WD-18, WA-20, WC-30 

Thommes, Edward WC-21 

Thonemann, Ulrich WC-24 

Thorburn, Peter TF-47 

Thornblad, Karin WD-28 

Thurnher, Philipp MD-34 

Tian, Wendi WA-07 

Tiddi, Daniele WA-16 

Tijs, Stef TB-26 

Tikniouine, Abdessadek TF-30 

Timmer, Marco WB-26 

Timonina, Elena MB-18 

Tirado, Gregorio MB-37 

Tirkel, Israel TC-08 

Tishler, Asher TA-33 

Tokareva, Julia TF-26 

Toklu, Nihat Engin WC-25 

Tokola, Henri WA-25 

Toledano-Kitai, Dvora MB-18 

Toledo, Franklina WB-05, MF-34, 

TB-34 


Tolstad Uggen, Kristin WC-22 

Tomasgard, Asgeir MB-28, MC-28, 

MC-42 

Tombarkiewicz, Barbara MF-32, TA-32 

Tomlin, John TB-28 

Tone, Kaoru MC-06 

Tong, Edward WA-20 

Toninelli, Daniele WC-29 

Toninelli, Roberta WA-33 

Topaloglu, Seyda TD-09 

Topcu, Y. Ilker MB-11, MC-12, MD-12 

Topuz, Emel MC-12 

Torjai, Laszlo TD-08 

Tormos, Pilar MA-04 

Torres, Delfim F. M. TC-48 

Tortosa-Ausina, Emili WC-06 

Toshimasa, Ozaki TD-12, MC-38 

Toste, Marisa TA-04 

Toth, Paolo MB-02, WB-02, MB-16 

Totic, Selena TA-25 

Touati Moungla, Nora WB-25, MC-48, 

MD-48 

Tralhão, Lino MB-23 

Tran, Dung Trung MF-40 

Tran, Quang Van WE-44 

Tranchevent, Léon-Charles MA-24 

Transchel, Sandra MF-34 

Trapero Arenas, Juan Ramon MC-28 

Trautmann, Norbert MA-08 

Trautsamwieser, Andrea TA-22 

Trenholm, Sven MA-22 

Tresoldi, Emanuele WD-15 

Tricoire, Fabien MC-03, TB-03, WB-15 

Trinh Hai, Le MA-31 

Triulzi, Umberto WB-29 

Triunfante Martins, Paulo MC-20 

Triviño, Maria TD-32 

Troczka-Pawelec, Katarzyna TF-34 

Troncoso, Juan José MB-32 

Trunfio, Roberto TD-17 

Trutnevyte, Evelina MA-33 

Tsai, Hsien-Tung WD-11 

Tsai, Jason TA-05 

Tsai, Ling-Chuan TA-06 

Tsantas, Nikolas MD-11 

Tseng, Fang-Mei TB-36, WA-42 

Tseng, Hwai-En MB-05 

Tsotsolas, Nikos MF-30, TD-30, TF-30 

Tsoukiàs, Alexis WB-37, MA-43, 

MB-43 

Tsutsui, Miki MC-06 

Tubilla, Fernando WC-28 

Tumanov, Oleg WB-47 

Tunali, Semra WA-04, MA-34 

Turan, Gyorgy TC-29 

Turan, Hasan H. TB-28 

Turgut, Zerin WD-40 

Turkay, Metin TD-24, MA-26, WD-36 

Turkensteen, Marcel WE-15 

Turksen, Ozlem MD-36 

Turri, Guillaume WD-25 

Turskis, Zenonas WC-43 

Tuyttens, Daniel WB-04, TA-23 

Tuzun Aksu, Dilek TB-03 

Tveit, Tor-Martin TD-23 

Tzeng, Gwo-Hshiung TC-12, TD-12, 

TF-12, MC-38 

Tzionas, Panagiotis TA-18 

U 

Uchoa, Eduardo MD-15 

Udding, Jan Tijmen TF-40 

Ueda, Tohru MA-06, WD-06 

Uhan, Nelson TC-02 

Ukai, Takamori WB-16 

Ukovich, Walter TC-27 

Uldry, Marc TC-17 

Ulengin, Fusun MA-21, TF-42 

Ulukan, Ziya WA-12 

Uluscu, Ozgecan S. WD-47 

Ulusoy, Gündüz TA-08, TF-42 

Ulutas, Sevan TC-25 

Unal, Alper TD-09 

Uney-Yuksektepe, Fadime MA-26 

Unterschultz, Jim WA-41 

Unuvar, Merve TF-46 

Unzueta, Aitziber MA-28 

Uphaus, Andreas MA-36 

Uratani, Tadashi TA-25 

Uriol, Juan TC-32 

Usó, Juan Fernando TC-32 

Usberti, Fábio MB-45 

Uylas, Nur WE-17 

Uzkent, Burak MF-26 

Uzunlu, Boran MA-31 

V 

Vaillancourt, Kathleen WB-32 

Vainiunas, Povilas WC-43 

Vairaktarakis, George WB-08 

Vajjala, Satyadeep WE-14 

Valecky, Jiri TD-33 

Valente, Jorge WB-28 

Vallada, Eva MF-04 

Vallat, Pierrick TF-16 

Valls, Vicente WA-39 

Valok, Tamara TC-05 

van Asperen, Eelco TF-40 

Van Delft, Christian MA-18, WA-24 

van den Akker, Marjan TB-16 

van den Brink, Rene TB-26 

Van den Broeck, Dennis TD-10 

van den Heuvel, Wilco WB-38 

van der Laan, Erwin WD-39 

van der Vorst, Jack WC-38 

Van Der Wal, Jan WC-38 

van der Westhuizen, Magderie MA-09 

van Dijk, Nico WC-38 

van Dinther, Clemens TF-28 

van Elst, Nicole TC-45 

Van Gestel, Tony WC-30 

Van Hentenryck, Pascal MC-48 

Van Hoeck, Ellen MA-12, WA-43 

van Hoeve, Willem-Jan TB-29 

van Houtum, Geert-Jan WB-11, WA-26 

van Kooten Niekerk, Marcel TB-16 

Van Laere, Elisabeth WC-30 

Van Lier, Tom MA-12 

Van Lokeren, Mark MD-18, WD-26 

Van Peteghem, Vincent MF-08 

van Valkenhoef, Gert TB-18, MB-30 

337


Van Volsem, Sofie TC-14 

Van Wassenhove, Luk WA-11 

Van Woensel, Tom MB-15, MC-15 

Vandaele, Arnaud WB-04 

Vandaele, Nele WD-14, WC-46 

Vanden Berghe, Greet WB-24 

Vanderbeck, François WB-09 

Vanhoucke, Mario MF-08, MC-30 

Vanmaele, Michèle WD-14 

Vannelli, Anthony MB-29 

Vanneschi, Leonardo TC-24 

Vanthienen, Jan WB-20 

Vashchenko, Mikhail WB-23 

Vasilakis, Christos TD-35 

Vasin, Alexander MC-25 

Vaz, Clara WD-06 

Vazquez Novoa, Nadia MF-06 

Vícha, Josef TA-36 

Vöcking, Berthold WA-02 

Veelken, Sonja TD-09, MB-42 

Vega, Pastora TC-48 

Velasco, Nubia MA-15 

Velazco, Marta MB-48 

Velazquez, Leticia WB-21 

Velichko, Andrey MF-09 

Veliov, Vladimir MF-39 

Venables, Harry WB-13 

Venckauskait, Jrat WA-43 

Ventura, Alecsandra TB-35 

Ventura, Juan TF-06 

Vera, Juan C. MB-29 

Verbeeck, Katja TF-28 

Verbeke, Wouter WB-20, WC-30 

Vercellis, Carlo TA-42 

Vercher, Enriqueta TF-21, TD-25 

Verlinde, Sara WA-43 

Verly, Céline TA-30 

Vermunt, Jeroen K. WC-29 

Verter, Vedat TA-13 

Vespucci, Maria Teresa MB-28 

Vetschera, Rudolf MA-44, MB-44, 

TC-46 

Viana, Ana TC-22, WB-29 

Vianna, WilliamMC-22, MD-33, TA-35 

Vidal, Thibaut MF-03 

Vidal-Puga, Juan TA-26 

Vidalis, Michael TC-14, TD-14 

Vidovic, Milorad WA-13 

Vieira Júnior, Milton WE-40 

Vieira Junior, Milton WD-40 

Vieira, Bruno MA-05 

Vieira, Douglas TD-48 

Vieira, João TF-47 

Vieira, Luís TC-19 

Vieira, Susana MD-36 

Viejo, Pablo TD-46 

Vigna, Elena WB-14, MD-39 

Vigo, Daniele TA-15 

Vilela, Sonia TB-39 

Vilkkumaa, Eeva MA-44 

Villa Monteiro, Daniel TB-18 

Villa, Gabriel MC-06, TA-16, MF-21 

Villagran de Leon, Claudia Margarita 

TC-23 

Villas-Boas, Sergio B. WC-40 

Villegas, Juan G. MA-15 

338 

Vilutiene, Tatjana WC-43 

Vincenzi Bortolotti, Silvana Ligia MF- 

18, MD-38 

Vinti, Cinzia TD-17 

Violi, Antonio WE-30 

Virtanen, Terhi TA-32 

Viry, Patrick MC-21 

Visona_Dalla_Pozza, Laura TB-47 

Vitoriano, Begoña MB-37 

Vizcaino, José Federico MB-45 

Vizvari, Bela TD-18, MC-29 

Vla, Katarina MD-28, MF-28 

Vlach, Milan TA-36 

Vocaturo, Francesca MC-04 

Voessner, Siegfried TB-20 

Vogt, Bodo MA-36 

Voigt, Ina MF-46 

Volkovich, Zeev (Vladimir) MB-18, 

MC-18 

von Lucken, Christian TB-32 

Voss, Stefan WD-15, MC-17, WD-22 

Vosvrda, Miloslav WE-44 

Voyvoda, Ebru MA-33 

Vrisagotis, Vassilios TC-14 

Vryzidis, Lazaros TD-37 

Vujosevic, Mirko TC-05, MD-18 

Vukicevic, Milan TF-18 

W 

Waaub, Jean-Philippe WB-32 

Wachowicz, Tomasz MD-30 

Waegeman, Willem TD-29 

Wagelmans, Albert WB-38, TD-45 

Wagler, Annegret Katrin MD-29 

Wagner, Michael TB-27 

Wagner, Neal MC-05 

Wakolbinger, Lea M. TC-46 

Waligora, Grzegorz WC-39 

Wallace, Mark TB-27 

Wallace, Stein W. TB-19 

Wallenius, Hannele MD-43 

Wallenius, Jyrki MD-43 

Walther, Ursula WB-41 

Walukiewicz, Stanislaw WD-08 

Wan Husain, Wan Rohaida TC-06 

Wang, Hsiao-Fan TA-43 

Wang, Jingbo MA-41 

Wang, Kuo-Hsiung WD-40 

Wang, Pei-Chun MB-20 

Wang, Sheng-Pen TB-14 

Wang, Xin MB-17 

Wang, Yi-Hsien MA-11, TA-25 

Wang, Yufei TA-40 

Wang, Zidong WD-46 

Wanka, Gert WD-09 

Watanabe, Takahiro MC-19 

Waterer, Hamish TB-27 

Wauters, Tony TF-28 

Weber, Gerhard-Wilhelm WB-10, WD- 

17, MD-25, MA-26, MB-26, 

MD-26, WD-44, MC-46 

Weber, Richard WE-18 

Weber, Valentin WC-20, WD-25 

Weglarz, Jan WC-39 

Weider, Steffen TB-16, TF-16 

Weigert, Gerald WB-07 

Weintraub, Andrés MB-32 

Weis, Stephan MF-46 

Weismantel, Robert WC-35 

Weissensteiner, Alex WA-29 

Weitschek, Emanuel TC-24 

Wen, Min WA-15 

Wendt, Oliver WA-15 

Wensing, Thomas WA-38 

Wenstøp, Fred TA-31 

Werfel, Mandy TF-23 

Werk, Sebastian WC-31 

Werner, Adrian WC-22 

Werner, Axel TF-31 

Werner, Frank MD-08, WD-39 

Werners, Brigitte TC-15 

Wesselmann, Franz WA-09 

Westerlund, Tapio WB-35 

Westermann, Lutz MC-21, MD-21 

Weyland, Dennis WD-15 

White, Kevin TB-08 

White, Leroy MA-35 

Widmer, Marino TC-17, MF-35 

Wiedersheim, Sabrina WC-16 

Wiegele, Angelika MB-29 

Wierzbicki, Andrzej MC-23 

Wiese, Jörg TD-40 

Wiesemann, Wolfram TF-36 

Willemse, Elias Jakobus TA-17 

Willert, Bernhard MF-09 

Williams, Janet TD-28 

Windolph, Melanie TC-35 

Winkelkotte, Tobias TC-31 

Winner, Mssip 2010 WE-02 

Winterfeld, Anton TD-23, MF-46 

Wirl, Franz MA-39 

Wirth, Fabian MC-14 

Wirth, Jerome TA-37 

Witlox, Frank WA-43 

Witteveen, Cees TF-28 

Woeginger, Gerhard J. MF-10 

Woerner, Stefan MB-09 

Wolenski, Peter MB-39 

Wolfermann, Axel TD-27 

Wolfler-Calvo, Roberto TA-02, MC-03, 

MD-15, WD-15, WD-25 

Wollenweber, Jens MB-13, TD-31 

Wolsey, Laurence MA-34 

Wolter, Kati TB-02 

Wong, Elaine TF-04 

Wong, S.c. TC-04 

Wood, Richard M TD-28 

Woods, Megan WD-23 

Woodsend, Kristian MC-21 

Worthington, Dave TB-05, TC-22 

Wozabal, David WC-29, WD-29 

Wrzaczek, Stefan MA-19 

Wu, Cheng-Ru TC-11, WC-11, WA-42 

Wu, Lixin MA-18 

Wu, Yongzhong MF-03, TC-04 

Wu, Yue TD-40, WD-45 

W˛eglarz, Andrzej MF-32 

X 

Xabadia, Àngels MA-39 

Xambre, Ana Raquel MF-19, TA-44 

Xanthopoulos, Spyros MF-25, MB-38

Xanthopoulos, Stylianos MC-19 

Xavier, Adilson Elias WC-40, TF-43, 

MD-47 

Xiao, Wenqiang MB-41 

Xie, Aijie MB-31 

Xu, Genjiu WC-26 

Xu, Yifan WB-17 

Xue, Fan WD-15 

Xufre, Patricia MA-11, TA-28 

Y 

Yaesh, Isaac MC-39 

Yagiura, Mutsunori TA-20 

Yakupo˘glu, Lütfü TB-36 

Yalaoui, Alice MB-03, WA-04 

Yalaoui, FaroukMF-03, MA-04, TB-04, 

WA-04, WC-04, MC-05 

Yalcindag, Semih MC-48 

Yaman, Hande MD-40 

Yaman, Kübra MD-26 

Yanasse, Horacio WD-04, MF-34 

Yang, Boting WC-42 

Yang, Chien-hsin MA-11 

Yang, Fu-Ju MA-11, TA-25 

Yang, Guangyuan WD-38 

Yang, Kum-Khiong TA-22 

Yang, Shih-Hao MB-12 

Yang, Shih-Ting TF-18 

Yang, Tsau-Tang TD-18 

Yang, Xinan MB-45 

Yang, Yi-Shu TA-28 

Yannacopoulos, Athanassios MC-19, 

MD-19 

Yannacopoulos, Denis TD-30, TF-30 

Yannez, Javier WC-21 

Yanovskaya, Elena TB-26 

Yao, Zhong MB-21 

Yüksel, Ihsan TD-12, TC-36 

Yarovaya, Elena WE-39 

Yatsenko, Yuri MA-39 

Yavuz, Emre WD-13 

Yavuztürk, Suzan MC-15 

Yazarli, Sevcen MF-11 

Yazici, Onur MA-31 

Ye, Mujing MC-32 

Yee, Julie MB-35 

Yeh, Tsai Lien MF-07 

Yeh, Tsu-Ming WA-10 

Yelbay, Belma WC-05 

Yelkenci, Simge WA-04 

Yemshanov, Denys WA-41 

Yenigun, Husnu WC-24 


Yerlikaya Ozkurt, Fatma WD-17, 

MA-26 

Yet, Barbaros TD-22 

Yeung, Thomas MA-15 

Yevseyeva, Iryna MF-04 

Yildiz, Abdurrahman TB-03 

Yilmaz, A. Egemen WB-10 

Yilmaz, Arif TF-11 

Yilmaz, Erdal WE-05 

Yilmaz, Osman TC-25 

Yoda, Kunikazu TF-46 

Yolmeh, A.m. WC-04 

Yong, Hsiu-Feng TB-07 

Yorke-Smith, Neil TF-06 

Yoshida, Yuichiro MC-06 

Yoshizaki, Hugo MF-35 

Young, David MB-28 

Yousfi, Naouel WA-30 

Yu, Fong Jung WB-11 

Yu, Jing-Rung TA-25, TB-36 

Yu, Yugang TF-40, WA-45 

Yuan, Benjamin J.C. WB-12 

Yucel, Erol MC-12 

Yurtkuran, Alkin MA-03 

Yıldız, Burhan TD-47 

Yılmaz, Mustafa Kerim TF-25 

Yılmaz, Volkan TB-36 

Z 

Zabolotskyy, Taras WD-29 

Zachary, Daniel MA-33, MA-47 

Zadnik Stirn, Lidija MA-45 

Zaerpour, Nima WA-45 

Zaidi, Ichraf WB-05 

Zak, Eugene TD-38 

Zakeri, Golbon MB-28 

Zambujal-Oliveira, Joao WC-14 

Zamorano, Gonzalo TB-05 

Zanjirani Farahani, Reza WA-05, 

MF-14 

Zanoni, Simone MC-34 

Zapata, Elizabeth MC-37 

Zarama Urdaneta, Roberto MD-37 

Zaras, Kazimierz TF-37 

Zaraté, Pascale TD-37, MD-43 

Zarghami, Mahdi MB-21, MF-44 

Zarrabi, Mohammad Reza TF-48 

Zarrazola, Edwin WC-21 

Zarzo, Alejandro TA-16 

Zavadskas, Edmundas Kazimieras WC- 

43 

Zazanis, Michael MA-18 

Zazgornik, Jan WA-15, MA-32 

Zeblah, Abdelkader WA-04 

Zednik, Eliseu MC-35 

Zeferino, João TF-47 

Zeimpekis, Vasileios TA-17 

Zemkoho, Alain B. MD-42 

Zenklusen, Rico MB-09, WE-14 

Zey, Bernd MA-02 

Zhang, Dali TD-40 

Zhang, Hao TB-42 

Zhang, Huizhen TD-02 

Zhang, Jie WD-18 

Zhang, Xiang-Sun MA-24 

Zhang, Yan WC-31 

Zhao, Huiling MC-22 

Zheng, Jinhua MA-45 

Zheng, Zhichao TF-04 

Zhou, Li TB-11 

Zhou, Zhili WC-31 

Zhu, Dauw-Song TB-06 

Zhu, Joe MC-06 

Zhu, Xiang WA-11 

Zhuravlev, Vladimir WC-11, MC-18 

Ziarnetzky, Timm WB-07 

Zilinskas, Antanas TF-43 

Zilinskas, Julius TC-43, TF-43 

Zilinskas, Kestutis WC-43 

Zimányi, Esteban MF-23 

Zimmermann, Jürgen MA-08 

Zimmermann, Jörg TA-24 

Zimmermann, Karel MF-36 

Zografos, Konstantinos TB-08, WB-12, 

WA-42, MC-48 

Zohrehbandian, Majid MD-06 

Zonderland, Maartje MB-30 

Zouaghi, Iskander WA-10 

Zouhar, Jan TD-46 

Zuddas, Paola TA-15 

Zuglian, Sara TB-17 

Zuidwijk, Rob TD-14 

Zutt, Jonne TF-28 

Zwols, Yori MC-10 

Zylinski, Pawel WE-42 

¸Sahin, Halenur MA-03 

¸Seker, Merve TF-04 

¸Sen, Ceyda MF-33 

¸Sen, Halil WB-28 

Šarka, Vaidotas WC-43 

Šarkien, Edita WC-43 

Šelih, Jana WB-43 

Šetinc, Marko TD-44 

339

SESSION INDEX 


MA-01: Keynote Talk 1 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 

MA-02: Combinatorial Optimization I (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 

MA-03: TSP (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 

MA-04: Scheduling with metaheuristics (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 

MA-05: Theory (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 

MA-06: DEA Methodology I (8.2.30). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 

MA-07: Recent Developments in Mathematical Programming (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 

MA-08: Project Management Software and Applications (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

MA-09: Challenges of Mathematical Programming by Modern Applications (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

MA-10: Graphs and Networks I (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

MA-11: Emerging Data Mining Applications in Biomedics and Biotech (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 

MA-12: AHP 01 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 

MA-13: Location and GIS (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 

MA-14: Inventories in Supply Chains (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 

MA-15: Location-routing problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 

MA-16: Rolling stock and Re-scheduling (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 

MA-17: Collaborative Planning I (1.3.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 

MA-18: New Achievements in Stochastic Models and Optimization (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

MA-19: Game Theory and Economics (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

MA-20: Cutting and Packing 1 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 

MA-21: OR in Practice I (6.2.47). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 

MA-22: Teaching OR/MS (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 

MA-23: MOO: Algorithms for Multi-Objective Combinatorial Optimization I (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

MA-24: Bioinformatics I (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

MA-25: Financial Mathematics and OR 1 (6.2.48). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 

MA-26: Mathematical Programming Approaches for Classification Problems (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 

MA-27: Financial Optimization 1 (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 

MA-28: Stochastic Integer Programming (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 

MA-30: MCDA II: Multicriteria ranking and classification vs portfolio decision analysis (Panel) (8.2.13) . . . . . . . . . . . . . . . . . . . . . . 14 

MA-31: Societal Complexity and Climate (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 

MA-32: Transportation planning in forest products industry (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

MA-33: Energy Modeling (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

MA-34: Solution Approaches for Lot-sizing Problems I (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 

MA-35: Teaching Soft OR and PSMs (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 

MA-36: Fuzzy expert systems (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 

MA-37: OR Applications in the Health Field (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 

MA-38: Experimental Economics and Game Theory (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

MA-39: Optimal Control in Economics and Economic Demography (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

MA-40: Wireless and sensor networks (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

MA-41: Recent Advances in Revenue Management (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

MA-42: Bilevel Programming (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

MA-43: Algorithmic Decision Theory 1 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

MA-44: Portfolio Decision Analysis I (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

MA-45: Dynamic Programming I (8.2.12). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 

MA-46: OR Challenges Related to the Recent and Future Disasters I (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

MA-47: Various Aspects of Sustainable Living in Developing Countries (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 

MA-48: Ill-posed Variational Problems I (8.2.04). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 


MB-01: Keynote Talk 2 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 

MB-02: Advanced Combinatorial Optimization 1 (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 

MB-03: VRP (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 

MB-04: Scheduling with metaheuristics (2) (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 

MB-05: Genetic algorithms (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 

MB-06: DEA Methodology II (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 

MB-07: New Achievement in Mathematical Programming (8.2.47). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 

340

EURO 24 - Lisbon 2010 SESSION INDEX 

MB-08: Scheduling at Container Terminals (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 

MB-09: Optimization of Transport Problems on Networks I (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 

MB-10: Graphs and Networks II (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 

MB-11: Technical Communication (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

MB-12: AHP 02 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

MB-13: Facility Location and Supply Chain Management (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

MB-14: Lotsizing and Supply Chain Planning (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

MB-15: Rich routing problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 

MB-16: Robust optimization in public transport (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 

MB-17: Collaborative Planning II (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

MB-18: Stochastic Modeling and Simulation I (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

MB-19: Modelling the Human Decisions (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 

MB-20: Cutting and Packing 2 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 

MB-21: OR in Practice II (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 

MB-22: Teaching with cases (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 

MB-23: MOO: Algorithms for Multi-Objective Combinatorial Optimization II (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 

MB-24: Bioinformatics II (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 

MB-25: Financial Modelling and Risk Management (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 

MB-26: Bioinformatics Applications of Machine Learning (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

MB-27: Financial Optimization 2 (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

MB-28: Applications of stochastic programming to the energy sector - electricity (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 

MB-29: Global Optimization of Graph Partitioning (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 

MB-30: MCDA II: Health (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 

MB-31: Societal Complexity, City and Economy (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 

MB-32: OR in Forestry I (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 

MB-33: Energy Policy and Planning (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 

MB-34: Solution Approaches for Lot-sizing Problems II (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 

MB-35: Soft OR and Problem Structuring Methods I (6.2.46). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 

MB-36: Fuzzy Goal Programming 1 (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 

MB-37: Humanitarian Logistics (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

MB-38: Advances in Economical and Financial Theory (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

MB-39: Set-Valued Analysis for Control Problems (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

MB-40: Network design 1 (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

MB-41: Behavioral Models in Revenue Management (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 

MB-42: Mathematical Programs with Complementarity Constraints (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 

MB-43: Algorithmic Decision Theory 2 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

MB-44: Portfolio Decision Analysis II (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

MB-45: Dynamic Programming II (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 

MB-46: OR Challenges Related to the Recent and Future Disasters II (8.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 

MB-48: Ill-posed Variational Problems II (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 


MC-01: Keynote Talk 3 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

MC-02: Applications of Combinatorial Optimization (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

MC-03: Metaheuristics for the VRPTW (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 

MC-04: Parallel machine scheduling with metaheuristics (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 

MC-05: Multi-objective metaheuristics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 

MC-06: DEA Methodology III (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 

MC-07: Recent Advances in the Use of Mathematical Programming (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 

MC-08: Airside Airport Operations (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 

MC-09: Complementarity Problems, Variational Inequalities and Equilibrium (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

MC-10: Graphs and Applications (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

MC-11: Energy Market Modeling (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

MC-12: AHP 03 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

MC-13: Managing disruptions in facility location models (2.2.21). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 

MC-14: Price and Capacity Planning in Supply Chains (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 

MC-15: Green Vehicle Routing Problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 

MC-16: Rescheduling in railway operations (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 

MC-17: Online Planning in Vehicle Routing and Scheduling (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 

MC-18: Stochastic Modeling and Simulation II (1.3.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 

MC-19: Mathematical Finance (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

MC-20: Cutting and Packing 3 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

MC-21: Optimization Modeling I (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

341

SESSION INDEX EURO 24 - Lisbon 2010 

MC-22: OR in Education I (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 

MC-23: MOO: Algorithms for Multi-Objective Combinatorial Optimization III (6.2.49). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 

MC-24: Bioinformatics III (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 

MC-25: Risk Management in Operations (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 

MC-26: Machine Learning to help people with dissabilities (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 

MC-27: Financial Optimization 3 (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 

MC-28: Applications of stochastic programming to the energy sector - gas (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

MC-29: Algorithmic Applications of Boolean Functions (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

MC-30: MCDA II: Health and Environment (8.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 

MC-31: Societal Complexity and Stakeholders (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 

MC-32: OR in Forestry II (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 

MC-33: Electric Vehicles (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 

MC-34: Production and Inventory decisions with recycling (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 

MC-35: Soft OR and Problem Structuring Methods II (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 

MC-36: Genetic Fuzzy Systems; Fuzzy Goal Programming 2 (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 

MC-37: OR Applications for Public Policy Assesment in Developing Countries (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

MC-38: Recent Advances in the Economics Supported by OR I (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

MC-39: Recent Advances in Optimal Control Theory (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 

MC-40: Routing in telecommunication networks (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 

MC-41: Topics in Revenue Management (3.1.06). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 

MC-42: Solution Algorithms for Bilevel Problems (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 

MC-43: Algorithmic Decision Theory 3 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 

MC-44: Portfolio Decision Analysis III (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 

MC-46: OR on Environmental Networks and Management (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 

MC-47: OR in Oil Sector II (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 

MC-48: Green vehicle routing and scheduling (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 


MD-01: Keynote Talk 4 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 

MD-02: Algorithms for planning and scheduling problems (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 

MD-03: Dynamic routing problems (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 

MD-04: Job Shop scheduling with metaheuristics (3.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 

MD-05: Statistics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 

MD-06: DEA Methodology IV (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 

MD-08: Scheduling Problems - Approaches and Complexity (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 

MD-09: Nonlinear optimization for industrial applications (6.2.53). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 

MD-10: Graphs and Networks IV (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 

MD-11: Stochastic Models in Manpower Planning (8.2.38). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 

MD-12: AHP 04 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 

MD-13: Discrete Location I (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 

MD-14: Pricing Issues (2.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72 

MD-15: Exact Algorithms for Vehicle Routing (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 

MD-16: Robust planning and rescheduling (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 

MD-17: Breaks in Vehicle Routing and Scheduling (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 

MD-18: Queueing Systems (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 

MD-19: Industrial Organization (1.3.20). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 

MD-20: Cutting and Packing 4 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 

MD-21: Optimization Modeling II (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 

MD-22: OR in Education II (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 

MD-23: MOO: Multi-Objective Combinatorial Optimization (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 

MD-24: Bioinformatics IV (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 

MD-25: Financial Dynamics and Bubbles (6.2.48). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 

MD-26: Machine Learning for Multiple Sources (3.1.11). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 

MD-27: Financial Optimization 4 (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 

MD-28: Stochastic Programming Algorithms (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 

MD-29: Boolean Optimization in Graph Theory (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 

MD-30: MCDA II: Sorting Models, theoretical aspects and other issues. (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 

MD-31: Societal Complexity and Education (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 

MD-32: OR in Forestry III (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 

MD-33: Environmental Management I (8.2.19). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 

MD-34: Model Formulations and Real World Applications of Lot Sizing and Scheduling I (8.2.23). . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 

MD-35: Soft OR and Problem Structuring Methods III (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 

MD-36: Fuzzy Optimization (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 

342


MD-37: OR for Development and Developing Countries I (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 

MD-38: Recent Advances in the Economics Supported by OR II (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 

MD-39: Dynamic Programming Approach to Optimal Control Problems (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 

MD-40: Network design 2 (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 

MD-41: Advances in Revenue Management (3.1.06). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 

MD-42: Theory of Bilevel Programming (3.1.07). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

MD-43: Algorithmic Decision Theory 4 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

MD-44: Modelling complex systems (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 

MD-45: Novel opportunities of DC programming and DCA for Industry and Finance (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 

MD-46: Semi-Infinite Optimization I (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 

MD-47: OR in Oil Sector I (8.2.16). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87 

MD-48: Multi-objective optimization (8.2.04). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 


ME-01: Plenary Talk 1 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 


MF-02: Flexible shop scheduling by Metaheuristics and solutions for real problems (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 

MF-03: Population-based metaheuristics for routing problems (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 

MF-04: Shop scheduling with metaheuristics (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 

MF-05: EURO Doctoral Dissertation Award (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 

MF-06: DEA Methodology V (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 

MF-07: DEA Application IX (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 

MF-08: Scheduling: Algorithms and practical cases (6.1.36). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 

MF-09: Discrete and Continuous Optimization for Gas Networks (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 

MF-10: Multi-index Assignment Problems (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 

MF-11: Structural Equation Modelling Approach in User Acceptance of Information Technology I (8.2.38) . . . . . . . . . . . . . . . . . . . . 93 

MF-12: AHP 05 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 

MF-13: Discrete Location II (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 

MF-14: Supply Chain Planning (2.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94 

MF-15: Shortest Path Problems with Resource Constraints (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 

MF-16: Vehicle and crew rostering (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 

MF-17: Long-term Transportation Planning (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 

MF-18: Markov Chains (1.3.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 

MF-19: Game Theory and Statistics (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 

MF-20: Cutting and Packing 5 (1.3.33A). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 

MF-21: Optimization Modeling III (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 

MF-22: Quantitative Health Care Policy Decision Making (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 

MF-23: MOO: Network Territorial Partition Problems (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 

MF-24: Natural Computation in BioInformatics (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 

MF-25: Financial Risk Management (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 

MF-26: Neural Network Applications (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 

MF-28: Stochastic Programming - General Methodology (8.2.10). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 

MF-29: Boolean methods in system analysis, learning and circuit synthesis (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 

MF-30: New issues in aggregation-disaggregation philosophies (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 

MF-32: OR in Animal Production (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 

MF-33: Environmental Management II (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 

MF-34: Model Formulations and Real World Applications of Lot Sizing and Scheduling II (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . 102 

MF-35: Facilitated Decision Analysis I (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 

MF-36: Fuzzy Optimization and Decision Analysis 1 (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 

MF-37: OR for Development and Developing Countries II (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 

MF-39: Limit Behaviour and Approximations I (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 

MF-40: Telecommunications (6.2.52). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105 

MF-41: Pricing in networks (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 

MF-42: Optimization and Data Mining I (3.1.07). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 

MF-43: Revenue Management I (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 

MF-44: SD modelling of Scarcity and Sustainable Development (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 

MF-45: Recent developments from Nonconvex Programming (8.2.12). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 

MF-46: Semi-Infinite Optimization II (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 

343


MF-47: Advances in Discrete-Continuous Optimal Control 1 (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 

MF-48: OR/MS: Beyond Mathematics I (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 


TA-01: Keynote Talk 5 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 

TA-02: Advanced Combinatorial Optimization 2 (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 

TA-04: Project scheduling (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 

TA-05: EURO Excellent in Practice Award 2010 (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 

TA-06: DEA Methodology VI (8.2.30). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 

TA-07: DEA Application X (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 

TA-08: Project Scheduling: New Results and Applications (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 

TA-09: Mathematical Modeling on Timetabling Problems: Models and Analytic Network Process Approach (6.2.53) . . . . . . . . . . 111 

TA-10: Graphs and Networks VI (6.2.56). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 

TA-12: ANP 01 (8.2.39). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112 

TA-13: Continuous Location I (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 

TA-14: Closed-Loop Supply Chains (2.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 

TA-15: Vehicle Routing Problems with Pickups and Deliveries (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 

TA-16: OR models and algorithms (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 

TA-17: Contracts and Sub-contracting for Transportation (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 

TA-18: Supply Chain Management (1.3.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 

TA-19: Economical Models (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 

TA-20: Cutting and Packing 6 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 

TA-21: Optimization Modeling IV (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 

TA-22: Health Care Scheduling I (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 

TA-23: MOO: Scheduling Problems (6.2.49). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 

TA-24: Applications of OR in Life Science Informatics (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 

TA-25: Risk Management and Portfolio Optimization I (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 

TA-26: Operation research games (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 

TA-27: FREIGHT TRANSPORT AND LOGISTICS (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 

TA-28: Stochastic Programming Models 1 (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 

TA-29: Satisfiability: Structures and Complexities (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 

TA-30: MCDA II: Theoretical contributions (8.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 

TA-31: OR and Ethics I (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 

TA-32: OR in Animal Science (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 

TA-33: Energy Pricing Models (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 

TA-34: Model Formulations and Real World Applications of Lot Sizing and Scheduling III (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . 123 

TA-35: Facilitated Decision Analysis II (6.2.46). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 

TA-36: Fuzzy Optimization and Decision Analysis 2 (3.1.05). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 

TA-37: Mathematical models for energy and environment (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 

TA-38: Rating and Valuation of Credits (6.2.44). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 

TA-39: Limit Behaviour and Approximations II (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 

TA-40: Network design 3 (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 

TA-41: Methodological Aspects of System Dynamics Modeling (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 

TA-42: Optimization and Data Mining II (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 

TA-43: Revenue Management II (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 

TA-44: Modelling energy systems (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 

TA-45: Nonconvex Programming approaches for Machine Learning and Data Mining (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 

TA-46: OR in Fisheries, Maritime Sciences and Related Aspects (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 

TA-47: Advances in Discrete-Continuous Optimal Control 2 (8.2.16). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 

TA-48: Heuristics 1 (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 


TB-01: Keynote Talk 6 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 

TB-02: Advanced Combinatorial Optimization 3 (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 

TB-03: Routing problems (3.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130 

TB-04: Multi-objective scheduling with metaheuristics (3.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 

TB-05: EURO Excellent in Practice Award 2010 (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 

TB-06: DEA Application I - Banking (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

TB-07: DEA - General topics I (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

344


TB-08: Airport operations scheduling (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

TB-09: Recent Advances in the Theory of Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 

TB-10: Graphs and Networks VII (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 

TB-11: Advances in the Use of Information Technology I (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 

TB-12: ANP 02 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 

TB-13: Continuous Location II (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 

TB-14: Supply Chain Coordination (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 

TB-15: Polyhedral Approaches to Routing Problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 

TB-16: Public Bus Transportation (2.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 

TB-17: Container Terminal Planning I (1.3.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 

TB-18: Simulation and Optimization of Networks under Uncertainty (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 

TB-19: Networks and Industrial Organization (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 

TB-20: Cutting and Packing 7 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 

TB-21: Optimization Modeling V (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 

TB-22: Health Care Scheduling II (3.1.10). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 

TB-23: MOO: Network Optimization and Transportation (6.2.49). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 

TB-24: Algorithms in Computational Biology (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 

TB-25: Risk Management and Portfolio Optimization II (6.2.48). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 

TB-26: Games solutions (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 

TB-27: AIR TRANSPORTATION AND LOGISTICS (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 

TB-28: Stochastic Programming Models 2 (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 

TB-29: Applications of Boolean Functions (8.2.11). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 

TB-30: MCDA II: Group Decision (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 

TB-31: OR and Ethics II (8.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 

TB-32: OR in Agriculture (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 

TB-33: Renewable Energy Production (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 

TB-34: Model Formulations and Real World Applications of Lot Sizing and Scheduling IV (8.2.23). . . . . . . . . . . . . . . . . . . . . . . . . . 144 

TB-35: Facilitated Modelling in Action Research (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 

TB-36: Forecast based on fuzzy logic or neural networks (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 

TB-37: Meeting the targets: policy and measures (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 

TB-38: Stochastic Valuation of Derivatives and Commodities I (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 

TB-39: Advances in Control Problem (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 

TB-41: Applications of System Dynamics Modeling I (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 

TB-42: Data Analysis and Decision Making (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 

TB-43: Revenue Management III (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 

TB-44: Simulation Decision Support in Enterprises (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 

TB-45: Restricted and Unrestricted Clustering (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 

TB-46: Agent-Based Modeling of Electricity Markets (8.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 

TB-47: Stochastic Models for Service Operations I (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 

TB-48: Heuristics 2 (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 


TC-01: Keynote Talk 7 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 

TC-02: Cooperative Games and Combinatorial Optimization (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 

TC-03: Inventory and routing problems (3.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 

TC-04: Industrial and city problems (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 

TC-05: Location under uncertainty (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 

TC-06: DEA Application II - Education (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 

TC-07: DEA - General topics II (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 

TC-08: Production Scheduling (6.1.36). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 

TC-09: Mathematical Programming, Machine and Statistical Learning (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 

TC-10: Graphs and Networks VIII (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 

TC-11: Advances in the Use of Information Technology II (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 

TC-12: ANP 03 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 

TC-13: Reliability in Location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 

TC-14: Vendor Managed Inventory (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 

TC-15: Dynamic and Stochastic Vehicle Routing (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 

TC-16: OR Applications in Railways (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 

TC-17: Models for Vehicle Routing (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 

TC-18: Complex Systems under Uncertainty: Networks and Data Mining (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 

TC-19: Universality in complex systems (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 

TC-21: Software for OR/MS I - Optimization (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 

TC-22: Health Care Policy Making I (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 

345


TC-23: MOO: Facility Location Problems (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 

TC-24: Data Mining in Bioinformatics (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 

TC-25: Volatility Spillover and Liquidity Risk (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 

TC-26: Cooperative Games and Applications (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 

TC-27: FREIGHT TRANSPORTATION (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 

TC-29: Theory of Boolean Functions (8.2.11). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 

TC-30: PROMETHEE: Axiomatic basis and other issues (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 

TC-31: Network Planning in Postal Logistics (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 

TC-32: OR in Agriculture and Forest Management (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 

TC-33: Energy Planning Models (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 

TC-34: Generalized Convexity I (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 

TC-35: Researching Facilitated Modelling (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 

TC-36: Fuzzy Decision Making and Applications (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 

TC-37: Models for Decision Making & Decision Analysis (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 

TC-38: Stochastic Valuation of Derivatives and Commodities II (6.2.44). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 

TC-39: Optimal Control: Recent Advances I (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 

TC-41: Applications of System Dynamics Modeling II (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 

TC-42: Data Mining and Forecasting (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 

TC-43: Global Optimization 1 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 

TC-44: Interregional Security Work (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 

TC-45: OR in Military I (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 

TC-46: Agent-based Modeling I (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 

TC-48: Nonlinear Optimization and Applications 1 (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 


TD-01: Keynote Talk 8 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 

TD-02: Topics in Combinatorial Optimization (3.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 

TD-03: Transportation and logistics (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 

TD-04: Agriculture, forestry and environmental problems (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 

TD-05: Parameter tuning and interactive metaheuristics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 

TD-06: DEA Application III - Transportation (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 

TD-07: Scheduling problems in production and service (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 

TD-08: Scheduling Applications (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 

TD-09: Optimization of Transport Problems on Networks II (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 

TD-10: OR in Sports 1 (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 

TD-11: Various New OR Tools and Technologies I (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 

TD-12: ANP 04 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 

TD-13: Emergency facilities location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 

TD-14: Supply Chain Design and Sustainability (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 

TD-15: Routing Optimization (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 

TD-16: Strategic decisions and infrastructure (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 

TD-17: Container Terminal Planning II (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 

TD-18: Decision Analysis in Marketing and Financial Modeling (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 

TD-19: Dynamical Systems (1.3.20). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179 

TD-20: Cutting and Packing 8 (1.3.33A). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 

TD-21: Software for OR/MS II - Open Source (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 

TD-22: Health Care Policy Making II (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 

TD-23: MOO: Nonlinear Multi-Objective Optimization Techniques in Action (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 

TD-24: Bioinformatics V (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 

TD-25: Risk Management and Portfolio Optimization III (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 

TD-26: Dynamics, statistical inference and algorithms of cooperative game theory (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 

TD-27: URBAN TRAFFIC CONTROL (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 

TD-28: Scheduling in Health Care (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 

TD-29: Preference Learning I (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 

TD-30: MCDM 1 (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 

TD-31: OR based Decision Support by Fraunhofer Society (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 

TD-32: OR in Forest Management (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 

TD-33: Modelling of uncertainties in the energy sector (short-term planning) (8.2.19). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 

TD-34: Generalized Convexity II (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 

TD-35: Facilitated Discrete-Event Simulation (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 

TD-36: Fuzzy Decision Making (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 

TD-37: Applications in Decision Making & Decision Analysis (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 

TD-38: Stochastic Valuation of Energy Prices and Derivatives (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 

346


TD-39: Optimal Control: Recent Advances II (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 

TD-40: Port Simulation and Optimization (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 

TD-41: Applications of System Dynamics Modeling III (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 

TD-42: Decison Making 1 (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 

TD-43: Global Optimization 2 (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 

TD-44: Business Process Modelling and Simulation (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 

TD-45: OR in Military II (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 

TD-46: Agent-based Modeling II (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 

TD-47: Optimization in Water Systems I (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 

TD-48: Multi-Objective Optimization (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 


TE-01: Plenary Talk 2 (Aula Magna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 


TF-03: Location problems (3.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 

TF-04: Airline applications (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 

TF-05: Meet the Editors of EJOR (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 

TF-06: DEA Application IV — Quality of Life and development (8.2.30). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 

TF-07: Scheduling Applications (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 

TF-08: Scheduling with Uncertainties (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 

TF-09: Various Aspects of Modern Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 

TF-10: OR in Sports 2 (6.2.56). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .196 

TF-11: Structural Equation Modelling Approach in User Acceptance of Information Technology II (8.2.38) . . . . . . . . . . . . . . . . . . . 196 

TF-12: ANP 05 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 

TF-13: Competitive Location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 

TF-15: Vehicle Routing Applications I (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 

TF-16: Optimization Methods for Railway Freight Transportation (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 

TF-17: Technologies for Collaborative Planning (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 

TF-18: Data Mining and Knowledge Representation (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 

TF-20: Cutting and Packing 9 (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 

TF-21: Software for OR/MS III (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 

TF-22: Health Care Policy Making III (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 

TF-23: MOO: Multiple Criteria Approaches in Mathematical Finance (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 

TF-24: Workforce Scheduling 1 (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 

TF-25: ROADEF/EURO challenge junior session 1 (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 

TF-26: Cooperative situations on networks: algorithms and applications (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 

TF-27: LOGISTICS 1 (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 

TF-28: Decentralized scheduling (8.2.10). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 

TF-29: Risk measurement and management (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 

TF-30: MCDM 2 (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 

TF-31: Communication Network Design (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 

TF-32: OR & Water Management (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 

TF-33: Optimality of alternative policy instruments for climate and energy policies (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 

TF-34: Generalized Convexity and Related Topics (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 

TF-35: Recent advances in mixed-integer nonlinear and global optimization (6.2.46). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 

TF-36: Robust Optimization (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 

TF-37: ’MCDA & uncertainty (3.1.09). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .207 

TF-40: Operation Planning and Control in Container Terminals (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 

TF-42: Decision Making 2 (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 

TF-43: Simplicial methods in Global Optimization (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 

TF-44: Vector and Set-Valued Optimization I (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 

TF-46: Probabilistic Constrained Stochastic Programming (8.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 

TF-47: Optimization in Water Systems II (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 

TF-48: Nonlinear Optimization and Applications 2 (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 


347


WA-02: Keynote Talk 9 (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 

WA-04: Resource sizing and allocation (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 

WA-05: Networks (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 

WA-06: DEA Application V — Industry and natural resources (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 

WA-07: Project scheduling (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 

WA-08: Shop Scheduling (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 

WA-09: Recent Developments and Applications in Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 

WA-10: OR in Supply Chain Management I (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 

WA-11: Various New OR Tools and Technologies II: Quality Management Emphasized (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 

WA-12: ANP 06 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 

WA-13: Applications of Location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 

WA-15: Vehicle Routing Applications II (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 

WA-16: Case studies in planning and operations (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 

WA-17: Network Design for Road Transportation (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 

WA-18: Graph Theory and Combinatorial Optimization (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 

WA-20: Data Mining and Credit Risk (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 

WA-21: Optimization Algorithms I (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 

WA-23: Model Selection in Regression Analysis (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 

WA-24: Workforce Scheduling 2 (6.2.50). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 

WA-25: ROADEF/EURO challenge junior session 2 (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 

WA-26: Special classes of cooperative games and allocation rules (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 

WA-27: TRAFFIC AND ENVIRONMENT (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 

WA-28: Scheduling with Transportation (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 

WA-29: Simulation and Optimization Modeling in Finance (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 

WA-30: MCDM 3 (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 

WA-31: Industrial applications of scheduling and routing I (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 

WA-32: Meeting EURO-WG OR in Agriculture and Forest Management (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 

WA-33: Optimal policy in the energy markets (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 

WA-34: Computational Methods (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 

WA-35: MINLP: new developments and applications (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 

WA-36: OR in the Public Sector (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 

WA-37: MCDA applications in agricultural and environmental management (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 

WA-38: Issues in Inventory Management Applications I (6.2.44). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 

WA-39: Real-time scheduling of logistic warehouse operations (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 

WA-40: Berth Allocation in Maritime Container Terminals (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 

WA-41: Long Term Decisions in Forestry (3.1.06). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 

WA-42: Decision Making 3 (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 

WA-43: Operational Research in Sustainable Urban Development (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 

WA-44: Vector and Set-Valued Optimization II (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 

WA-45: New technologies in facility Logistics (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 

WA-46: Stochastic Programming Tools (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 

WA-47: OR in Educational Problems and Systems (8.2.16). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 

WA-48: Variational inequalities and applications to economic market models (8.2.04). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 


WB-02: Keynote Talk 10 (3.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 

WB-04: Assignment and clustering problems (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 

WB-05: Graph problems (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 

WB-06: DEA Application VI — Utilities (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 

WB-07: Scheduling Approaches for Complex Manufacturing Systems (8.2.47). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 

WB-08: Approximation and Competition (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 

WB-09: New Frontiers in the Application of Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 

WB-10: Emerging Nonlinear Optimization Applications of OR (6.2.56) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 

WB-11: Recent Advances in Quality Management and OR in Reliability Engineering (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 

WB-12: AHP 06 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 

WB-13: Heuristics in location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 

WB-14: Pension funds (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 

WB-15: Vehicle Routing Applications III (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 

WB-16: OR models in Public Transport (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 

WB-17: Financial Mathematics and Stochastic Modelling (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 

WB-18: Stochastic Models and Queueing Systems (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 

WB-19: Nonsmooth Optimization and Its Applications (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 

348


WB-20: Data Mining and Decision Making II (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 

WB-21: Optimization Algorithms II (6.2.47). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 

WB-22: Maritime transportation (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 

WB-23: Price and Risk Forecasting in the Financial Sector (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 

WB-24: Automated Nurse Rostering (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 

WB-25: ROADEF/EURO challenge senior session 1 (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 

WB-26: New Achievements in Game Theory I (cooperative and noncooperative) (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 

WB-27: URBAN TRANSPORT SYSTEMS (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 

WB-28: Scheduling with Due Dates (8.2.10). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 

WB-29: Energy and Environmental Finance (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 

WB-30: Risk Models in Finance (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 

WB-31: Industrial applications of scheduling and routing II (8.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 

WB-32: Energy and technological system issues (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 

WB-33: Carbon markets (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 

WB-34: Fast algorithms for large matrix optimization problems (8.2.23). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 

WB-35: MINLP Reformulations and Applications (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 

WB-36: News from Commercial MIP Solvers (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 

WB-37: MCDA and Public Administration (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 

WB-38: Lot-sizing models (6.2.44). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246 

WB-39: Scheduling and Pricing (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 

WB-40: Optimal Control and Design in Applications (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 

WB-41: Long Term Financial Decisions (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 

WB-42: Decision Making 4 (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 

WB-43: Sustainable Construction Processes (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 

WB-44: Vector and Set-Valued Optimization III (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 

WB-45: New Trends of Facility Logistics (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 

WB-46: Stochastic Optimization Models (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 

WB-47: Activities for Popularization of Science (8.2.16). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 

WB-48: Iterative Methods for Economic Models: Related Topics (8.2.04) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 


WC-02: Keynote Talk 11 (3.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 

WC-04: Facilities planning, design and management (3.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 

WC-05: Matheuristics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 

WC-06: DEA Applications XI (8.2.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 

WC-07: Machine Scheduling (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 

WC-08: Various Advances on Management and Scheduling I (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 

WC-09: Advanced Applications in Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 

WC-11: Advances in the Use of Information Technology III (8.2.38) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 

WC-12: AHP 07 (8.2.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 

WC-13: Location and Network Design (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 

WC-14: Stochastic Methods in Finance and Economics (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 

WC-15: Arc Routing Problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 

WC-16: Improving Real-time Railway Operations (2.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 

WC-17: Regression and Its Application (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 

WC-18: Stochastic Models and Optimization (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 

WC-19: Large-scale Mixed Optimization Problems (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 

WC-20: Discrete and Global Optimization (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 

WC-21: Optimization Algorithms III (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 

WC-22: LNG transportation (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 

WC-23: Data Mining in Portfolio Analysis 1 (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 

WC-24: Crew Scheduling (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 

WC-25: ROADEF/EURO challenge senior session 2 (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 

WC-26: Models of Cooperative Games: Theory and Applications (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 

WC-27: TRANSPORTATION PLANNING (8.2.06). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 

WC-28: Scheduling with Lags and Setups (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 

WC-29: Portfolio Selection (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 

WC-30: Advances in Quantitative Credit Risk Modeling: Change We Need (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 

WC-31: Industrial applications of scheduling and routing III (8.2.15). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 

WC-32: Long-term modelling for agriculture and forestry (8.2.17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 

WC-33: Risk and Uncertainty in Energy Models (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 

WC-34: Duality Problems (8.2.23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 

WC-35: Integer Nonlinear Programming (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 

349


WC-36: Container Terminal Applications (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 

WC-37: Territorial Decision Making (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 

WC-38: Perishable item inventory management (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 

WC-39: Discrete-continuous scheduling (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 

WC-40: Recent Advances in Engineering Optimization (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 

WC-41: Forestry Management and Long Term Financial Decisions (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 

WC-42: Graph Cleaning (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 

WC-43: Sustainable Development in Civil Engineering and Multiattribute (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 

WC-44: Dynamical Systems and Mathematical Modelling in OR I (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 

WC-45: Category and Shelf Space Management (8.2.12). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 

WC-46: Numerical methods for derivatives pricing and hedging (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 

WC-47: Data Mining in Early Warning Systems I (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 


WD-02: Keynote Talk 12 (3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 

WD-04: Constructive and local search methods (3.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 

WD-05: Electronics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 

WD-06: DEA Application VII - Retailing and health (8.2.30). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 

WD-07: Project Scheduling (8.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 

WD-08: Various Advances on Management and Scheduling II (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 

WD-09: Convex Duality in Mathematical Programming (6.2.53) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 

WD-11: Advances in the Use of Information Technology IV (8.2.38). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 

WD-12: AHP 08 (8.2.39). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275 

WD-13: Decision support for practical logistics problems (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 

WD-14: Stochastic methods in actuarial sciences (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 

WD-15: Traveling salesman problems (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 

WD-17: Data Mining Tools and Improvements (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 

WD-18: Data Mining for Credit Scoring (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 

WD-19: Nonsmooth and Nonconvex Optimization Methods (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 

WD-21: Social Policy and Education (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 

WD-22: Maritime Vessel Routing and Deployment (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 

WD-23: Data Mining in Portfolio Analysis 2 (6.2.49) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 

WD-24: Practical Issues in Timetabling (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 

WD-25: ROADEF/EURO challenge senior session 3 (6.2.48). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 

WD-26: New Achievements in Game Theory II (cooperative and noncooperative) (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 

WD-27: TRAFFIC MANAGEMENT (8.2.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 

WD-28: Scheduling in Production and Communication (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 

WD-29: Information and Ambiguity in Financial Modeling (8.2.11). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 

WD-30: OR methods in portfolio management and asset allocation (8.2.13). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 

WD-33: Decision making under model uncertainty (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 

WD-35: MINLP - Problem-specific Approaches (6.2.46). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 

WD-36: Linear and Conic Programming I (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 

WD-37: Application of outranking approach for sustainable development (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 

WD-38: Recent Advances in Spare Parts Inventory Management (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 

WD-39: Scheduling and lotsizing under uncertainties I (6.2.45) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 

WD-40: Recent Advances in Industrial and Engineering Optimization I (6.2.52). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 

WD-41: Monte Carlo and Malliavin Calculus (3.1.06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 

WD-42: Cops and Robber Games (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 

WD-43: Life Insurance, Risk Management & OR (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 

WD-44: Dynamical Systems and Mathematical Modelling in OR II (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 

WD-45: Logistics and Promotions Management (8.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 

WD-46: Approximation of Probabilities for Exotic and Compound Options (8.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 

WD-47: Data Mining in Early Warning Systems II (8.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 


WE-02: EURO Management Science Strategic Innovation Prize (MSSIP 2010) on the topic of Optimization in Telecommunications 

(3.2.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 

WE-05: Tools for metaheuristics (3.2.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 

WE-06: DEA Application VIII — Software (8.2.30). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 

350


WE-08: Various Advances on Management and Scheduling III (6.1.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 

WE-13: Hub Location (2.2.21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 

WE-14: Industrial Applications in Risk Management (2.2.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 

WE-15: Vehicle Routing and Set Covering Models (2.2.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 

WE-17: Recent OR Advances by Statistics, Probability and Performance Measures (1.3.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 

WE-18: Data Mining Applications in Business Intelligence (1.3.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 

WE-19: Nonsmooth Global Optimization (1.3.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 

WE-20: Social Networks (1.3.33A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 

WE-21: Education and Sustainable Development (6.2.47) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 

WE-22: Maritime Logistics: Theory and Practice (3.1.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 

WE-24: Educational Timetabling (6.2.50) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 

WE-25: ROADEF/EURO challenge senior session 4 (6.2.48) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 

WE-26: Game Theory and Social Choice (3.1.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 

WE-28: Scheduling under Resource Constraints (8.2.10) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 

WE-29: Financial Modeling (8.2.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 

WE-30: Risk measurement and control in volatile financial markets (8.2.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 

WE-33: Realistic Production Scheduling I (8.2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 

WE-35: Advances in Mixed-Integer Linear and Nonlinear Programming (6.2.46) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 

WE-36: Linear and Conic Programming II (3.1.05) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 

WE-37: Robustness concerns and multiple criteria decision aid (3.1.09) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 

WE-38: Issues in Inventory Management Applications II (6.2.44) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 

WE-39: Scheduling and lot sizing under uncertainties II (6.2.45). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 

WE-40: Recent Advances in Industrial and Engineering Optimization II (6.2.52) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 

WE-41: Financial Mathematics and Simulation (3.1.06). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 

WE-42: Graph Guarding (3.1.07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 

WE-43: Research Aspects Related to Life - Risk and Insurance (8.2.02) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 

WE-44: Systems and Mathematical Modelling in OR III (8.2.03) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 

WE-46: AIMMS User Meeting (8.2.14). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301 

351

TECHNICAL SESSIONS Monday, 9:00-10:20 - Euro 2010 Lisbon

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?