Pedro is a PhD student at the Mechanical Engineering department at the Technical University of Delft, The Netherlands.

In Inverse Optimization problems, a learning agent aims to learn how to mimic the behavior of an expert agent, which given an exogenous signal, returns an action. The underlying assumption is that in order to compute its action, the expert agent solves an optimization problem parametric in the exogenous signal. Therefore, given examples of exogenous signals and corresponding expert actions, the goal of the learner is to learn the cost function being optimized by the expert. In this talk, we will present fundamental concepts of Inverse Optimization and novel results, as well as applications to static and dynamic routing problems for the 2021 Amazon Last Mile Routing Research Challenge and the EURO Meets NeurIPS 2022 Vehicle Routing Competition.

Yousef Maknoon is a faculty member at the Faculty of Technology, Policy, and Management (TPM) at TU Delft. He is also the director of Orbit Lab, a research group specializing in Operations Research and Behavioral Informatics in Transportation. His research takes a multidisciplinary approach, firmly rooted in operations research, to tackle emerging challenges in the transport and logistics domain. In recent years, his primary focus has been on the design and operational strategies for on-demand and instant logistics services, driving the evolution of this dynamic field.

This study delves into the role of context effects in assortment optimization for platform-based urban mobility services, challenging the assumptions of classical Random Utility Maximization (RUM) models. We employ the Random Regret Minimization (RRM) model and introduce a novel 'Marginal Decoy Policy' specifically designed to capture behavioral anomalies arising from context effects. Through comparative analyses with established models such as Multinomial Logit (MNL) and Generalized Random Regret Minimization (G-RRM), we demonstrate the efficacy of our approach. The findings elucidate new strategies for dynamic service menu optimization, significantly enhancing customer satisfaction and retention by taking context effects into account.

Dongyang Xia is a PhD candidate at School of Transportation at Beijing Jiaotong University. He is currently enrolled as a visiting PhD student at Delft University of Technology under the supervision of Dr. Shadi Sharif Azadeh. His research focuses on the optimization of public transport operations under uncertainty. Dongyang's research aims to improve the operational efficiency and service quality of urban public transport. His PhD project was carried out in collaboration with an industry partner the Beijing Public Transportation Corporation, where he applies operations research optimization techniques to support their decision-making. As of 2024, he will start a postdoc position in the SUM lab for SINERGI project (ERANET funded).

Current metropolitan public transportation systems employ fixed-capacity vehicles; however, this makes it challenging to match uneven passenger demand over time and space. Fortunately, emerging modular vehicle technology makes it practical for vehicles to dynamically adjust their capabilities during operation, which holds the promise of being one of the solutions to the above problem. This technology introduces the concept of dynamic capacity, which refers to the number of modular units comprising an MV can be potentially changed at different times and stops. For the purpose of better applying the features of modularized buses and reducing the operating costs without compromising the quality of service, we have done the following researches: (1) we develop a data-driven distributionally robust optimization model at the line operation level that considers time-varying capacity over all possible demand distributions within the ambiguity set. (2) Further, we formalize and analyze the integrated problem of timetables, vehicle schedules, and dynamic allocations of capacity with MVs in a network under uncertainty. To efficiently solve the above models, we first design customized benders decomposition algorithms and then incorporate them into a rolling horizon optimization framework to solve large-scale problems. Finally, we conduct a series of real-world case studies based on the operational data of the Beijing Public Transportation Corporation and validate the effectiveness of the proposed approaches.

With increasing congestion of high-value orbit regions, effective and efficient space domain awareness (SDA) is becoming essential. Usual approaches to space situational awareness (SSA) involves application of mathematical techniques to propagate positions of satellites and their orbits to understand their positioning and predict conjunctions/collisions. This project attempts to develop a new wholistic approach to SDA by integrating a systems view. Hereby SDA would be achieved by analysing Space Systems including Space, Ground, User and Launch segments as well as their respective interactions. Subsequently the goal is to utilise multidisciplinary approaches and techniques to understand and predict the complex system that is the near-earth space environment with its satellites. Applications would be the development of models, simulations and algorithms for real-time predictions and control of space systems.

Nadia Lahrichi holds a PhD in applied mathematics from Polytechnique Montréal. She is currently a full professor at the department of Mathematics and industrial engineering at Polytechnique Montreal. She is also a member of CIRRELT and IVADO. Her research is mainly focused towards applying modeling and operational research tools to improve patient flow in the healthcare system. She uses exact, metaheuristics and discrete event simulation approaches to tackle patient and resource scheduling problems. She has received the award for outstanding application of operational research (from the Canadian Operational research society) for solving the home health care routing and scheduling problem.

Many approaches are used to handle uncertainty in stochastic combinatorial optimization problems. In this talk, we describe the application of a tabu search approach in a stochastic environment together with a real application in physician scheduling in a radiotherapy center. The goal is to determine a weekly cyclic schedule that improves the patient flow and shortens the pretreatment duration. High uncertainty is associated with the arrival day, profile and type of cancer of each patient. Additionally, two approaches to improve the efficiency of the method are introduced, both are based on leveraging methods that originate outside the field of metaheuristics. The first one discusses hyperparameters tuning. Research shows that it is a nontrivial task and efficient methods are required to obtain the best possible results. We present how blackbox optimization can help choose the tabu search parameters efficiently. We are solving this problem through a Mesh Adaptive Direct Search (MADS) algorithm with no derivative information. The second one presents a learning algorithm for improving tabu search by reducing its search space and evaluation effort. The learning tabu search algorithm uses classification methods in order to better motivate moves through the search space.

Dr. Andre L. Carrel is an associate professor of transportation at the Ohio State University (OSU) and the director of the OSU Travel Behavior Research Group. Dr. Carrel’s areas of expertise are travel demand forecasting, novel travel data collection technologies, and public transportation. Dr. Carrel’s current research focuses on modeling the dynamics of travel behavior, travelers' adaptation to experienced travel times and trip quality, the impact of vehicle electrification and automation on travel behavior, and interactions between information and communication technologies (e.g., online shopping and teleworking) and travel demand. His research leads to insights regarding factors that contribute to more sustainable travel choices, which allow decision-makers to plan and design cities, transportation services, and travel-related products that meet travelers’ needs while improving the sustainability of the transportation system. Dr. Carrel is jointly appointed in Civil Engineering and City and Regional Planning and is a core faculty member of the OSU Translational Data Analytics Institute. His academic experience spans both passenger transportation and logistics. Dr. Carrel holds a Dipl.-Ing. from ETH Zurich, an MS degree from MIT, and a PhD degree from the University of California, Berkeley, and was a postdoctoral associate at the MIT Center for Transportation and Logistics.

Strategies to achieve durable shifts toward more sustainable travel behavior should seek to both reduce car-centric lifestyles and to shift remaining car use to cleaner vehicle technologies. A key step toward the first goal is to incentivize reductions in household car ownership, which recognizes the interconnected nature of vehicle ownership with daily activity and travel patterns. In this presentation, Dr. Carrel will discuss how new forms of virtual working and shopping may not only lead to reduced household car ownership, but also, how technology-driven vehicle subscription services that cater to low-vehicle households may promote the use of electric vehicles. The findings are based on two separate studies conducted in the United States. The first study finds a possible association between travelers’ levels of teleworking and online shopping on the one hand and mode use patterns on the other hand. A latent class analysis reveals a segment of the population that engages in high levels of such online activities while also being characterized by a comparatively low-car – but not car-free – lifestyle, thereby suggesting possible sustainability benefits of teleworking that go beyond the elimination of commute travel. The second study finds that short-term car subscription services, which allow households to reduce the number of permanently owned vehicles, can promote the use of battery electric vehicles (BEVs). Analyses of data from a stated preference survey reveal that such services appeal to a segment of the population that is unwilling to commit to purchasing a Battery Electric Vehicle (BEV) but is open to subscribing to one, suggesting that subscription services could increase the number of BEVs on the road. In summary, this presentation argues that with appropriate policy guidelines, technology-based models of working, shopping, and owning vehicles can be leveraged to promote more sustainable travel behavior.

Léa Ricard is a Ph.D. candidate at the Canada Excellence Research Chair in “Data Science for Real-Time Decision-Making”. They completed a Bachelor degree in Industrial Engineering at Polytechnique Montréal in 2017. Between their undergraduate and graduate studies, Léa worked as a production planning coordinator at Rolls Royce Canada. In 2018, they started a Master degree in Computer Science at the University of Montréal and followed a fast track from the Master’s to the Ph.D. in the same program. In 2021, Léa was awarded the Alexander Graham Bell Canada Graduate Scholarships-Doctoral Program from the Natural Sciences and Engineering Research Council of Canada and the HORizon Alma Mater Arts and Sciences fellowship from the Faculty of Arts and Sciences of the University of Montréal. Their research focuses on mixed integer programming, column generation, stochastic optimization, and machine learning applied to public transportation. Their doctoral research project is carried out under the supervision of Andrea Lodi, Guy Desaulniers, and Louis-Martin Rousseau and in collaboration with GIRO Inc., a world leader in the development of optimization-based software for public transportation.

Logistics and transport problems, for example routing and scheduling problems, are typically challenging to solve and many of them have been proven to be NP-hard. Moreover, these problems are subject to many operational uncertainties that further increase their complexity. In this talk, I will first introduce decomposition methods, and in particular the column generation algorithm, as well as optimization frameworks to tackle uncertain problems. Then, I will present a recent work where we combined the latter for a classical public transportation scheduling problem, the multi-depot vehicle scheduling problem (MDVSP). In this work, we formulated the reliable MDVSP with stochastic travel time (R-MDVSP-STT) as a bi-objective problem where the objective is to build vehicle schedules that are both cost- and delay-efficient. We modeled the R-MDVSP-STT using a path-based formulation and proposed a heuristic branch-and-price algorithm to solve it. The reliability of a schedule is assessed according to the discrete probability mass function of its trips’ departure time. We have developed a method to compute the exact convolution of the latter distributions using the probability mass functions of the travel time and have integrated this information into the stochastic pricing problems. In a test with real data collected by buses running in Montréal (Canada), we showed that the R-MDVSP-STT provides schedules that are more tolerant to delays for a negligible increase in planned costs.

Evangelos Paschalidis is a Research Fellow at the Institute for Transport Studies, University of Leeds. He completed his PhD in 2019 on "Developing driving behaviour models incorporating the effects of stress" at the same department. Currently, he is working on the EU funded project PAsCAL that focuses on the public acceptance of automated vehicles. His main research interests are in the areas of driving behaviour modelling, choice modelling, human factors and traffic psychology.

Models used in microscopic simulation are mostly approximating driving behaviour as a function of the surrounding traffic conditions, without considering drivers’ individual characteristics. In this talk a series of traditional driving behaviour models (car-following and gap-acceptance) is presented, that incorporate the effect of drivers’ stress and other individual characteristics. The models were estimated via econometrics techniques, using data collected at the University of Leeds Driving Simulator (UoLDS) in two different road environments, while stress levels were measured via physiological responses as heart rate and skin conductance. Issues of transferability between driving simulator and field traffic are also discussed.

Pavel is a Ph.D. student at CERGE-EI, Prague, Czech Republic. His main research interest is Theoretical Behavioral Economics, mainly the theory of Rational Inattention. Pavel received an MA degree in Economics from European University at Saint Petersburg in 2017 and an MSc degree in Applied Mathematics from St Petersburg State University in 2014. He published a paper about discrete choice which connects rational inattention and conformity theories.

A principal needs to delegate a decision under uncertainty to an agent and chooses among candidates who differ in their prior beliefs. Prior to making the decision, the agents can flexibly acquire costly information. We show that the principal can generally benefit from delegation when the agent's belief differs from the belief of the principal. We also show that in a class of problems, hiring an agent with misaligned beliefs performs as well as action-contingent transfers, and is better than restricting the action set.

Rosemarie Santa earned a Bachelor and a Master’s in Industrial Engineering at the University of Puerto Rico at Mayagüez. Her master’s thesis was in sports logistics, titled Reducing Participants’ Interferences at the World’s Best 10K. She is currently a PhD student at the University of Quebec in Montréal. Her thesis is titled: Strategic and Tactical of Planing of Mobile Clinic Deployments and will be defended on August 30th, 2022. Before venturing into her PhD she worked as an engineering consultant, a marketing director, and as a health and safety instructor for the Red Cross. She has served as a lecturer for the introductory course of Operations Management at multiple institutions, teaching in English, French, and Spanish. She is a researcher at the CIRRELT and GERAD research centers in Montréal, Canada. She is also a plant lover

This research focuses on the tactical planning phase of mobile clinics deployments. Mobile clinics are vehicles that are often implemented in humanitarian contexts, as they allow healthcare practitioners to offer their services in underserved areas (i.e., areas that have limited access to healthcare facilities). There is no comprehensive tool or guide that addresses the tactical challenges faced by mobile clinic practitioners when planning their operations. This research studies these challenges and captures them with mathematical models to provide tools and managerial insights for practitioners and researchers. We will discuss in this presentation the existing literature pertinent to mobile clinics and we will focus on their deployment for humanitarian relief in complex emergencies or conflict zones. Additionally, we will present how to model and formulate the deployment of mobile clinics in a deterministic and stochastic humanitarian context. Specifically, we model the deterministic mobile clinics deployment as a multiperiod location routing problem (MLRP) to capture the time dependency nature of mobile clinic deployments for humanitarian relief. Moreover, we model mobile clinics deployment subject to different sources of uncertainty, and formulate the problem as a two-stage stochastic prize collection problem. We also present four different recourse policies and study their impact on the tactical plan.

Hajime Watanabe received a Ph.D. in engineering from Kumamoto University, Japan, in March 2022, with a thesis on “Bayesian approaches for handling and identifying endogeneity in discrete choice modeling.” He is currently a Japan Society for the Promotion of Science (JSPS) Research Fellow at the University of Tokyo and working with Prof. Eiji Hato. His research interests include activity-travel behavior analysis using discrete choice modeling, causal inference, and machine learning.

The sample selection modeling approach has been applied to describe unobserved dependency between residential location choice and travel behavior due to residential self-selection. Sample selection models describe a missing data mechanism in which expected travel behavior outcomes in residential locations are incidentally truncated by actual residential location choice, referred to as endogenous switching. A limitation of existing sample selection models in the literature is to assume a simple binary endogenous switching that describes people’s residential choices by a binary choice, e.g., choosing between urban or suburban residential areas. This study proposes a general sample selection modeling framework with a multinomial endogenous switching. The proposed model has a large-scale open-form structure, and this study thus employs an efficient Markov chain Monte Carlo algorithm according to the model structure for the parameter estimation. The proposed model can be a useful tool for providing insights into coordination between land use and transportation planning.

Moha Ghaderi is an Assistant Professor at Pompeu Fabra University and a BSE Affiliated Professor. He obtained his PhD (cum laude) from ESADE in 2017. His doctoral dissertation has received the ESADE best thesis award and the URL Extraordinary Doctoral Dissertation Award, Doctoral Dissertation Award from the European Doctoral Association in Management and Business Administration, and has been named among the top 3 doctoral dissertations for the biennial award by the MCDM International Society. His publications have appeared in several journals, including EJOR, Omega, and Computers & OR, and received Omega best paper award and INFORMS junior researcher best paper award. His research focuses on quantitative methods, marketing analytics, and decision sciences.

We develop a nonparametric model of discrete choice from a limited number of choice observations from different individuals. Choice elements have a multiattribute representation from the analyst perspective. However, preferences are not constructed from utility functions. Instead, we regard preferences as strict linear orders over the choice elements, paired with probability distributions that represent inferential uncertainty due to limited observations. While utility distributions are not identifiable from observed choices, the treatment of preferences in our model considerably alleviates the identification concern. Moreover, it allows capturing complex substitution patterns among options without making any specification assumption. We proceed by a robust formulation of heterogeneity and construct a high-dimensional joint distribution over the individual-level preference orders. Finally, an approximation method for the estimation's computational tractability is discussed.

Nejc started working on his doctoral research in March 2019. He holds a Master’s Degree in Transport, Infrastructure and Logistics from Delft University of Technology and a Bachelor’s Degree in Traffic Technology from the University of Ljubljana. His research interests are investigating travel behaviour, discrete choice modelling and market segmentation, in relation to both urban and long-distance travel. He is curious about how to design public transport networks and policies that encourage travellers to choose more sustainable alternatives.

Ridesourcing services (i.e. Uber, Lyft,…), microtransit and flexible public transport (collectively referred to as on-demand services) have come to be a popular transport mode around the world and are becoming increasingly interesting to public transport authorities and policy-makers as an alternative for fixed public transport services in low-demand areas. The main goal of the PhD is to advance the understanding of travellers’ preferences, attitudes and perceptions towards flexible on-demand services. Analysing how (1) these services fit in the wider mobility landscape and also how (2) travellers perceive the specific characteristics related to using an on-demand service with no fixed timetable or route, with variable waiting times, travel times and a potential to be denied service. The PhD is part of the larger CriticalMaaS project, investigating the interactions in online market places, focusing specifically on MaaS and Ridesourcing (on-demand) services: http://smartptlab.tudelft.nl/projects/criticalmaas

Fabian Torres is a Ph.D. student in the Department of Mathematics and Industrial Engineering at Polytechnique Montreal. He received his B.S. in Chemical Engineering at the Catholic University of Cuenca and an MBA from the University of Azuay, Ecuador. His research interests include stochastic programming, integer programming, dynamic programming, and combinatorial optimization, with applications in transportation and crowd-shipping.

E-commerce continues to grow all over the world. The recent pandemic caused by COVID-19 has increased this trend. Concurrently, crowd-shipping is emerging as a viable solution to fulfill last-mile deliveries, with AmazonFlex taking the lead in implementing such distribution models. We look at a problem of crowd-shipping were a crowd-shipping platform must fulfill delivery requests from a central depot with a fleet of professional vehicles and a pool of crowd-drivers. The latter can accept or reject routes based on their preference. The probability of route acceptance is dependent on the set of routes that are offered to crowd-drivers. The best compensated route is the most likely to be accepted. We develop a large neighborhood search heuristic to solve this routing problem. To investigate the practical viability of such distribution models, we show the market equilibrium when no fluctuation in supply is considered, versus the market equilibrium when the stochastic supply of crowd-drivers is considered. The best compensation for crowd-drivers that minimizes the total expected cost of the routing problem is determined. We show in our numerical experiments that a 6% cost reduction can be achieved by adjusting the compensation when we consider stochastic route acceptance.

Mengyi Wang is a third year Ph.D. student from Tongji University(Shanghai,China). At present, she is working with professor Xin Ye on the research of travel behavior and travel demand modeling. Her research interests include travel behavior and activity-based modeling. She served as the principal of the project“Feasibility Study on Development and Application of Activity-based Travel Demand Model in Shanghai” for Shanghai Urban Planning and Design Research Institute. Her doctoral thesis focuses on the modeling method of activity-travel generation under the optimal allocation of time resources.

Mengyi Wang will give a seminar which focuses on the introduction of her research about the activity-based modeling in Shanghai, including the feasibility study on the development and application of activity-based travel demand model in Shanghai and some related work in her doctoral thesis.

José Ángel Martín Baos is a predoctoral researcher in computer science in the School of Computer Science of the University of Castilla-La Mancha. There, he graduated with honors on both a BSc. in computer science in the year 2018, and a MSc. in computer science in the year 2019. He is currently studying his PhD at the University of Castilla-La Mancha with a FPU grant from the Spanish Ministry of Science, Innovation and Universities. He is also working as research assistant in the MAT (Modelos y Algoritmos en Sistemas de Transportes) research group since 2016.

During the last years, machine learning methods has gained great popularity due to its success in applications such as autonomous vehicles, intelligent robots, image and voice recognition, etc. This has led to an increased use of these methods and a growing interest in expanding the domain of application of machine learning methods, such as in the field of transport modelling. This seminar presents one of these methods, the Kernel Logistic Regression (KLR), from the point of view of Random Utility Models (RUM). It is presented how KLR can be used to specify the utilities in RUM, freeing the modeler from the need to postulate a functional relation between the features beforehand. A Monte Carlo simulation study is conducted to compare KLR with the Multinomial Logit model, and two of the most promising machine learning methods: the Support Vector Machines and the Random Forests. We have shown that on the simulated data, KLR is the only method that achieves maximum accuracy and leads to an unbiased willingness-to-pay estimator for non-linear phenomena. We have also carried an experiment with a real travel mode choice problem, where Random Forests achieved the highest predictive accuracy, followed by KLR.

Silvia Varotto is a researcher in human factors in vehicle automation at SWOV Institute for Road Safety Research in The Hague (NL). Her main research interests are developing advanced mathematical models that capture the decision making of individuals realistically and understanding the impact of cutting edge technologies on users' behaviour. In 2018, she received her PhD in Transportation Engineering from Delft University of Technology (NL) under the supervision of Dr. Haneen Farah, prof. Bart van Arem and prof. Serge Hoogendoorn. She was visiting scholar at Technion-Israel Institute of Technology (IL), Universität de Bundeswehr München (DE) and Ecole Polytechnique Fédérale de Lausanne (CH). She received her BSc and her MSc in Civil Engineering from University of Trieste (IT).

Intelligent vehicle systems and automated vehicles can contribute to reduce traffic congestion and accidents. Empirical findings have shown that intelligent vehicle systems might have both intended and unintended impacts on driver behaviour. Notwithstanding the potential effects on traffic operations, most microscopic traffic flow models currently used to evaluate the impact of intelligent vehicle systems do not describe driver responses accurately. To increase the realism and predictive ability of these models, my research aims to incorporate human behaviour elements based on empirical data and theories developed in the fields of human factors and traffic psychology. This seminar presents findings based on on-road experiments with full-range adaptive cruise control (ACC) and with an automatic incident detection (AID) system. A continuous-discrete choice model was developed to describe the underlying decision-making process of drivers with ACC at an operational level. Linear mixed-effects models were used to capture adaptations in driver deceleration behaviour while approaching traffic congestions with and without AID. These models can be used to predict driver response to intelligent vehicles systems and can be implemented into microscopic traffic flow simulations to evaluate the impact of these systems on traffic operations.

Michel Beine has been a professor of economics at the University of Luxembourg for more than 10 years and is a research fellow of IZA (Bonn), CES-Ifo (Munich), IRES (Louvain) and CREAM (London). His current research focuses on the economic issues related to international migration, such as brain drain, international mobility of students, climate change and migration and determinants of international migration. He has published more than 50 papers in academic journals, such as Economic Journal, Journal of Development Economics, European Economic Review, Review of Urban and Regional Economics, Canadian Journal of Economics or Scandinavian Journal of Economics or the Journal of Economic Geography. His work has received more than 7500 citations with a h-index of 39 (Google scholar). He is currently the coordinator of the PRIDE MINLAB doctoral program on topics involving migration, labour and inequality. He has been acting as an adviser of several international institutions such as the World bank, the European Commission, Industry Canada, the Walloon Region in Belgium or the Banque de France.

This paper studies the intended and unintended effects of a specific policy conducted by the French Government around 2006 aiming at boosting the number of foreign students admitted in French universities. The Campus France program aimed at facilitating the application process of foreign candidates from some particular countries and applying in specific universities. We develop a small theoretical model that allows for the existence of capacity constraints in order to analyse the potential effects of such a policy in terms of student inflows and in terms of selection. Using a Diff-in-Diff-in-Diff approach, we test the impact of Campus France on the magnitude of inflows. We find that the Campus France policy led to a global increase of inflows of foreign students around 8%. The increase is concentrated on universities outside the top 150 of the Shanghai Ranking. We also use the CF policy as a way to test the potential crowding-out effects on native students while taking care of the usual endogeneity concerns. We find some evide,ce in favour of crowding-in effects, either on native students or on foreign students coming through the traditional channel (joint work with Lionel Ragot, Univ Paris X).

Melvin Wong is a Postdoctoral Research Fellow at the Laboratory of Innovations of Transportation (LITrans) at Ryerson University, Canada. He completed his PhD at Ryerson University (Civil Engineering) in 2019 under the supervision of Dr Bilal Farooq. He specializes in using data-driven deep learning techniques for model optimization and developing interpretable deep learning based behavioural models in the context of smart mobility services. He holds a Bachelor degree in Electrical and Electronics Engineering from Nanyang Technological University (Singapore).

One area of rapid growth in transportation systems is smart mobility services and connected autonomous vehicles. These technological achievements have led to the development of new modes of travel and data sources which are decentralized, on-demand and dynamic – they stand to benefit the most from data-driven analysis and deep learning. This seminar presentation explores how we can capture and model changes in travel behaviour by leveraging on recent deep learning optimization and modelling techniques. We developed a new framework connecting deep learning to discrete choice models by representing the unobserved heterogeneity using deep residual layers - shortcut connections that capture the correlation patterns in behavioural choice models. This lends the model the ability to estimate increasingly complex models while retaining key information about the model parameters and individual behaviour. We further discuss how an implementation of deep learning optimization in discrete choice modelling might have an effect on economic interpretability.

Dr. Melvyn Sim is Professor and Provost's Chair at the Department of Analytics & Operations, NUS Business school. His research interests fall broadly under the categories of decision making and optimization under uncertainty with applications ranging from finance, supply chain management, healthcare to engineered systems. He is one of the active proponents of Robust Optimization and has given invited talks in this field at international conferences. Dr. Sim won second places in the 2002 and 2004 George Nicholson best student paper competition and first place in the 2007 Junior Faculty Interest Group (JFIG) best paper competition. He is also the recipient of the 2009 NUS outstanding young researcher award. Dr. Sim serves as an associate editor for Operations Research, Management Science and Mathematical Programming Computations.

Optimal routing solutions in deterministic models usually fail to deliver promised on-time services in the real world of uncertainty, causing potential loss of customers and revenue. In this study, we propose a new formulation for the data-driven Vehicle Routing Problem with Time Windows (vrptw) under uncertain travel times that is compatible with the paradigm of distributionally robust optimization. To mitigate the lateness as much as possible, our model minimizes an innovative decision criterion on the delays, termed the Service Fulfillment Risk Index (sri), while limiting the travel cost within a budget. The sri accounts for both the late arrival probability and its magnitude, captures the risk and the Wasserstein ambiguity in travel times, and is efficiently evaluable in closed form. In particular, the closed-form solution reduces the vrptw under the Wasserstein ambiguity of interest to the problem under the empirical distribution with advanced deadlines. To solve the problem, we develop a Benders decomposition algorithm and a variable neighborhood search heuristic, and explore their speedup strategies. We demonstrate their effectiveness through extensive computational studies. In particular, our solution greatly improves on-time arrival per- formance with slightly increased expenditure than the deterministic solution. Our sri also outperforms the canonical decision criteria, lateness probability and expected lateness duration, in out-of-sample simulations. This is a joint work with with Yu Zhang, Zhenzhen Zhang and Andrew Lim.

Ricardo Hurtubia is Assistant Professor at Pontificia Universidad Católica de Chile, with a dual appointment to the School of Architecture and the Department of Transport Engineering and Logistics. He is Associate Researcher at the Centre for Sustainable Urban Development (CEDEUS) and Invited Researcher at the Complex Engineering Systems Institute (ISCI). He completed his PhD at EPFL (Mathematics) in 2012, under the supervision of Professor Michel Bierlaire. His research is focused on location choice models, integrated transport and land use models, accessibility indicators as tools for project and policy evaluation and the use of discrete choice models to analyze and improve the design of public spaces and infrastructure through the understanding of user behavior.

Understanding how individuals define their walking environment is key to measure the role played by the built environment in (dis)encouraging walking and other active transport modes. This presentation discusses how walkability has been traditionally studied in the scientific literature and proposes a model for the "Perceived Walking Neighborhood". The model is estimated over data from Santiago, Chile. Results show how socioeconomic characteristics and land use attributes influence the shape of the walking neighborhood, opening the possibility to estimate individual-specific accessibility curves. Further research in this subject, such as the inclusion of factors accounting for residential self-selection and spatial dependency, is proposed and discussed.

Damon Wischik is a lecturer in the Computer Science department at Cambridge University, and a fellow of the Alan Turing Institute. He specializes in data science and machine learning. He spent five years as chief data scientist at Urban Engines, a startup which worked on urban analytics and behaviour nudging. Before that, he was a researcher in probability theory, computer networking, and statistics, in maths at Cambridge University and in computer science at University College London.

How should Uber set its prices and allocate its drivers? There are two separate types of modelling that are needed. First, we need to model how users make choices and how their choices depend on price; this is the field of discrete choice modelling, a staple of transport modelling. Second, we need to model traffic flows and how capacity should be allocated in a network; this is the field of network utility maximization, pioneered by Frank Kelly in the context of Internet congestion control. I will describe how these can be linked, via a surprising application of Shannon's information theory. I will also suggest avenues of research on how machine learning models of human behaviour might be integrated with models of urban or economic systems. Pictures: https://www.cl.cam.ac.uk/~djw1005/AboutMe/damon__bigsur.jpg

Marcel Mongeau received his BSc (1985) and MSc (1987) degrees in Mathematics from Universite de Montreal, and his PhD (1991) in Combinatorics & Optimization from the University of Waterloo (Canada). He was then a post-doctoral researcher at CRM (Universite de Montreal), at INRIA (France) and at the University of Edinburgh. From 1994 to 2011, he was at IMT, Universite Paul Sabatier (France), where he received a Habilitation a Diriger des Recherches in 2003. He is currently Professor in Operations Research at ENAC in Toulouse (France). His research interests include Global Optimization, Numerical Optimization, and Operations Research with applications to aeronautics.

We propose two approaches to address a challenging problem arising in Air Traffic Management, that of keeping at all times a distance between any pair of aircraft throughout their flight trajectory above a threshold value. We address the problem by adjusting both aircraft speeds and heading angles simultaneously. Both the mixed-integer nonlinear programming model and the penalty continuous optimization model we are introducing deal with the complex aircraft separation constraints through reformulations. Numerical results validate the proposed approaches.

Ricardo Daziano, choice modeler and PhD in economics (Universite Laval), is an associate professor of Civil and Environmental Engineering at Cornell University. Daziano's research focuses on engineering decision making, specifically on non-market valuation and choice microeconometrics applied to technological innovation in transportation and energy transitions. One of his goals is to better understand the interplay of consumer behavior and engineering, investment, and policy choices for broad adoption of energy efficiency.

Bayes estimators of the parameters of choice models offer several advantages over the dominant maximum likelihood approach. Although Bayesian techniques are the norm in some choice modeling fields, such as marketing, in other fields there has been some resistance to the use of Bayesian econometrics. In this talk, the fundamental principles behind computational Bayesian statistics will be reviewed before stressing associated benefits of Bayesian tools such as the use of predictive posteriors and credible intervals, estimators that are integral, gradient, and Hessian free, treatment of latent variables, and direct inference on transformation of the parameters of interest via post-processing Monte Carlo Markov chains. In addition to discussing implementation with empirical case studies, common misunderstandings will be clarified.

Nick Caros recently completed an M.S. in Transportation Planning and Engineering at New York University. During his studies he worked as a Research Assistant at the federally-funded C2SMART Tier 1 University Research Center, conducting research into simulation of flexible mobility services and network optimization in an autonomous vehicle environment. He holds a Bachelor of Applied Science in Integrated Engineering from the University of British Columbia and is currently working as a Transportation Planner for Stantec in New York City.

Emerging automotive technology is creating opportunities for more flexible ride-sharing services with shorter travel times and increased operational efficiency. A modular autonomous mobility service with the ability to transfer passengers between vehicles allows for new ride-matching and routing algorithms that reduce passenger disutility and operating cost. Our research, in partnership with NEXT Future Transportation, develops an insertion heuristic to assign trips to a fleet of modular autonomous vehicles and determine whether to engage in an en-route passenger transfer. The effectiveness of this algorithm is tested using a multi-day simulation with variable demand and an endogenous operator profit maximization algorithm. The simulation is tested using transit ridership data from the Dubai region in three scenarios: door-to-door service within an urban core, commuter first/last mile service and long-distance commuter service.

Pierre Attard is an engineer specialized in energy systems modeling and statistics. He holds the Engineering diploma from Mines Saint-Etienne, and a master's degree in energy economics/market from IFP (French institute for oil and energy). Pierre co-authored prospective studies and policy assessments for actors such as the European Commission, the European Climate Foundation as well as ministries and regulators. From his previous experiences, he has also worked on network-specific topics such as pricing or investing in production assets. His expertise in energy systems has led him to contribute to multi-energy modelling for local energy systems (Grand Lyon, Pays de Gex). He is proficient in R as well as in Python - including Machine Learning Libraries such as Scikit-Learn. He is also familiar with Machine Learning environments such as ElasticSearch, Kibana and Docker. Pierre participates in numerous projects requiring the implementation of forecasting tools and their industrialization.

As the European power system is progressing towards more liberalization and while climate policy targets are increasingly ambitious, the power sector is facing radical changes. Intermittent and decentralised renewable energies are developing at a fast pace, innovations in energy efficiency and power demand management are being deployed at a large-scale, and increased interconnections between market zones result in an unprecedented level of system integration, bringing new challenges for planning and operations. System level studies are necessary for decision-makers as to understand the issues and opportunities raised by these changes. In particular, major difficulties lie in making the right decisions for long-term planning while taking into account operational constraints representing the system's behaviour at a local level. In this sense, mathematical optimisation represents a powerful tool to model and simulate this type of complex problems. The first part of this presentation will give an overview of the actors involved in the European power system strategic decisions and will illustrate the type of studies typically conducted. The second part will detail a mathematical framework used for modelling the problems aforementioned with an emphasis on flexibility, providing means to avoid or defer unnecessary capital-intensive investments. A particular focus will be given on demand side flexibility with an illustration of demand-response schemes contributing to system's reliability as well as electric vehicles and heat pumps driven by price signals. The third part will illustrate a dedicated methodology for optimal investment with two studies conducted for the European Commission and the ADEME (French Environment and Energy Management Agency), respectively: 'Mainstreaming RES, Design of flexibility portfolios at Member State level to facilitate a cost-efficient integration of high shares of renewables' and 'The evolution of the French electricity mix between 2020 and 2060'. In the last part, the shortcomings of system-level studies regarding local issues will be discussed. The contrast between smooth aggregated and highly-variable local demand profiles will be used as to illustrate the challenges of transposing a system-level approach at local levels. Hence, typical questions such as 'how to predict the energy behaviour of a neighbourhood by 2040?' or 'how to define a market or regulatory mechanism to encourage households to invest in a particular type of heating equipment?' require a different approach and call for new methodologies.

Rafal did his PhD in real-time Dynamic Traffic Assignment with La Sapienza University of Rome (prof. Guido Gentile). Currently works at Department of Transportation Systems of Krakow University of Technology (Poland). He gained strong experienced as IT R&D developer (PTV SISTeMA), data-scientist (NorthGravity) and transport modeller with a rich portfolio of strategic models. Now focused on academic career, researching in the field of dynamic processes in traffic and transit networks especially non-equilibrium, unexpected and adaptation states of the network.

Traffic events (vehicle accidents, road closures, demand peaks, etc.) happen on a daily basis on road networks. Most typical effects are related to local capacity reductions, or global flow increases, which may both cause oversaturation, congestion and then delays. Forecasting how events affect the traffic pattern is of particular importance, since proper monitoring (to assess impacts) and management (to take actions) rely on predictions. Due to events, both travel times and traffic flows change with respect to typical values, not only for their direct effects, but also as a consequence of their indirect effects. More specifically, the event impacts may propagate on the network both upstream because of queue spillback and sideways because of rerouting, causing further delays. Modelling these effects on the supply side is relatively straightforward; if the route choice pattern is assumed fixed, this can be accomplished through a dynamic network loading model (Corthout et al., 2011). It is instead much more challenging to represent the effects on the demand side, namely how drivers are first informed and can then react to unexpected events by changing en-trip their pre-trip route choice.

Francesco Corman holds the chair of Transport Systems at the Institute of Transport Planning and Systems, Swiss Federal Institute of Technology, ETH Zurich. He has a PhD in Transport Sciences from TUDelft, the Netherlands, on operations research techniques for realtime railway traffic control. He has academic experience at KU Leuven, Belgium and TUDelft as research associate in transportation and logistics. Main research interests are in the application of quantitative methods and operations research to transport sciences, especially on the operational perspective, public transport, railways and logistics.

A major problem of public transport, and railways in particular, is to improve quality of operations by updating an offline timetable to the ever changing delays situation, in order to improve performance of the transport system. In railway systems, this relates to reduce train delays by reordering retiming or rerouting trains, and/or change connection plans and route advised to passengers, to improve their traveltime. Key point of research is the interaction between the microscopic characteristics of the problem (of the infrastructure manager) to reschedule trains and the macrosopic, multi-decision maker problem (of the travellers) to find the optimal route in the network. In fact, changing passenger flows, respectively delaying trains and/or dropping passenger connections, varies the setting under which the two decision makers respectively interact, and disruptions make the situation even more complex. The interaction of the two decisions makers is mediated by the information one decision maker has about the other, and the service which is offered/used. We analyse some solutions from the literature and some direct and indirect effects which are available in disruption situations, for railwaysn and public transport networks in general, where multiple stakeholders or processes interact with each other in multiple possible ways.

Francisco Pereira is Professor at the Technical University of Denmark (DTU) since August 2015, where he leads the Machine Learning for Mobility (MLM) group. MLM works on real-time traffic prediction, behavior modeling, advanced data collection technologies, big data and transport modelling. Previously, he was Senior Research Scientist at the MIT ITS Lab, based in both Boston and Singapore, as part of the Singapore-MIT Alliance for Research and Technology, Future Urban Mobility project (SMART/FM). He has a Masters and Phd from University of Coimbra, Portugal, on Computer Engineering and Artificial Intelligence. His research focus is on applying machine learning and pattern recognition to the context of transportation systems with the purpose of understanding and predicting mobility behavior, and optimizing the transportation system as a whole, but also on using concepts and methodologies from transportation (e.g. behavior modeling) to develop new machine learning research.

For multiple reasons, including little data and computing power, and a traditionalist community, representations of models and people (or, "agents") in transport demand behavior have changed very little through the decades. Each agent is represented as a vector of characteristics, and each model is a function that combines such characteristics with contextual information (e.g. travel time and cost for different transport options, in a mode choice model). However, new paradigms exist that have been under-explored, including networked representations, logic programming, deep representations, natural language based. It is well-known that many scientific breakthroughs in history have come up from "seeing things in a different perspective". It is my belief that strong opportunities in re-representing travel behavior, particularly considering current grand challenges of causality, transferability, new data types, social interactions, and many more. In this talk, I give some examples of the above from earlier work of mine and others, in order to present a few ideas and on-going work. Ultimately, I want to stimulate an exciting discussion towards collaboration between DTU and EPFL on these pressing and, I will argue, very promising opportunities!

Dr. Hewitt is an Associate Professor in the Information Systems and Supply Chain Management Department in the Quinlan School of Business at Loyola University Chicago, where he also serves as the Director of graduate programs in Supply Chain Management. His research includes developing quantitative models of decisions found in the transportation and supply chain management domains, particularly in freight transportation and home delivery. His work has assisted the decision-making of companies such as Exxon Mobil, Saia Motor Freight, and Yellow Roadway. He has expanded his area of expertise to include workforce planning, including working on multi-disciplinary projects at the intersection of operations management and cognitive psychology. His research has been funded by agencies such as the National Science Foundation, the Material Handling Institute, and the New York State Health Foundation. Before entering the PhD program at Georgia Tech, Dr. Hewitt worked as a software engineer, contributing to the development of software to support consumer set-top boxes and LED signs in mass transit stations.

Time-expanded networks are a useful tool from both modeling and computational perspectives. In terms of modeling, they enable a natural method for representing decisions that have both a geographic and temporal component. In terms of computation, they yield stronger integer programming formulations than those that represent time with continuous variables, which in turn require less time to solve. A drawback to the use of time-expanded networks is that they require time to be discretized. While finer discretizations yield more precise representations of time, they also lead to larger optimization models which may then require too much time to solve. However, this trade-off is primarily a function of choosing a discretization in a static and a priori manner. In this talk, we will present a method that generates time expanded networks in an iterative and dynamic fashion in the context of solving an optimization model that prescribe actions in both time and space. We will illustrate the use of this method on two classical problems seen in transportation and logistics: (1) the Service Network Design problem, which can be used to model the routing of goods between cities, and, (2) the Traveling Salesman Problem with Time Windows, which can be used to model the routing of goods within a city.

Nikola Obrenovic is an assistant professor at the Department of Power, Electronic and Telecommunication Engineering, Faculty of Technical Sciences, University of Novi Sad, and teaches BSc courses about databases and database design. He received his PhD degree in Electrical and Computer Engineering - Computing and Control Engineering from the University of Novi Sad, in October 2015. His doctoral studies and thesis incorporated areas of model-driven database development, rapid prototyping and automatic design verification. Currently, his research interests include databases, database design and data science, with applications to smart grid management systems. Additionally, Nikola Obrenovic is employed as a software architect at Schneider Electric DMS NS, Novi Sad, Serbia. His primary duties are design and development of various software solutions, which are part of Schneider Electric's Advanced Distribution Management System. The developed solutions belong to the fields of data mining, data warehouses, business intelligence and historian databases.

An important part of any power distribution management system data model is a model of load type. A load type represents typical load behaviour of a group of similar consumers, e.g. a group of residential, industrial or commercial consumers. A common method for creation of load types is the clustering of individual energy consumers based on their yearly consumption behaviour. To reach the satisfactory level of load type quality, the crucial decision is a choice of proper clustering similarity measure. In this talk, a comparison of different metrics, used as similarity measures in our process of load type creation, will be presented. Additionally, a novel metric, also included in the comparison, will be introduced. The metrics and the quality of load types created therewith are assessed by using a real data set obtained from the distribution network smart meters.

Moshe Ben-Akiva is the Edmund K. Turner Professor of Civil and Environmental Engineering and Director of the Intelligent Transportation Systems (ITS) Lab at the Massachusetts Institute of Technology (MIT) and at the Singapore-MIT Alliance for Research and Technology. He holds a BSc degree in Civil Engineering from the Technion and MSc and PhD degrees in Transportation Systems from MIT. He received honorary degrees from the University of the Aegean, the Universite Lumiere Lyon, the Stockholm Royal Institute of Technology, and the University of Antwerp. His awards include the Lifetime Achievement Award of the International Association for Travel Behavior Research, the Jules Dupuit prize from the World Conference on Transport Research Society, and the Institute of Electrical and Electronics Engineers ITS Society Outstanding Application Award for DynaMIT, a simulation-based dynamic traffic assignment. Ben-Akiva has coauthored two books, including the textbook Discrete Choice Analysis, and over 200 refereed papers. He has worked as a consultant in industries such as transportation, energy, telecommunications, financial services and marketing.

Effective Smart Mobility solutions incorporate optimization and behavioral modeling techniques. Real-time optimization provides efficiency while advanced behavioral modeling enables the design of personalized solutions. An app-based behavior laboratory, Future Mobility Sensing, and a computer simulation laboratory, SimMobility, are the two platforms used to design and evaluate Smart Mobility solutions. Four solutions are presented: 1. real-time toll optimization, 2. Autonomous Mobility On-Demand, 3. Flexible Mobility On-Demand, and 4. Tripod, a system of sustainable travel incentives with prediction, optimization and personalization capabilities. Simulation case studies demonstrate their potential benefits. The global long term effects on urban transportation of Smart Mobility, vehicle and fuel technologies, and energy and environmental policies are the subject of future research.

Michal Tzur is a professor in the Industrial Engineering Department of the Faculty of Engineering, Tel Aviv University, Israel. She received her Ph.D. in Management Science from the Graduate School of Business at Columbia University in the city of New York. She was a faculty member in the business school of the University of Pennsylvania (Wharton School) and a visiting faculty member in the department of Industrial Engineering and Management Sciences (IEMS) at Northwestern University. Her main research interests are in the areas of humanitarian logistics, vehicle sharing systems, vehicle routing and supply chain management. Since 2015 she serves as the president of the Operations Research Society of Israel (ORSIS).

We study the logistic challenges of a food bank that, on a daily basis, uses vehicles of limited capacity to collect donated food from suppliers in the food industry and distribute it to welfare agencies. We model this problem as a routing- resource allocation problem, with the aim of maintaining equitable allocations to the different agencies, while delivering overall as much food as possible. We introduce an innovative objective function that satisfies desired properties of the allocation, that is easy to compute and implement within a mathematical formulation, and that balances effectiveness and equity acceptably. We present an exact solution method, upper bounds, and a heuristic approach. Numerical experiments on several real-life and randomly generated datasets confirm that high-quality solutions may be obtained. Some implementation issues will be discussed, concerning the integration of our solution method into the information systems of the Israeli food bank with whom we cooperate (Joint work with Ohad Eisenhandler).

Heather Twaddle is a working on her PhD with Professor Fritz Busch at the Technical University of Munich (TUM). She has a BSc. In Civil Engineering from the University of Calgary and a MSc. in Transportation Systems from TUM.

Bicycling offers an inexpensive, environmentally friendly and compact option for urban mobility. However, issues with bicyclist safety persist in most countries. Driver assistance systems that warn drivers of the presence and potential actions of bicyclists are one possibility for mitigating these issues. In order to develop and evaluate such systems, microscopic traffic simulation software that is capable of realistically simulating bicycle traffic is necessary. However, the flexible behaviour of bicyclists makes it difficult to realistically simulate bicycle traffic using currently available simulation software. Here, a method for extracting behaviour patterns of bicyclists at intersections from video data is presented. These observed patterns are integrated with a microscopic traffic simulation, increasing the realism of simulated bicyclist behaviour and enabling the development and evaluation of driver assistance systems.

Dr. Justin Dauwels is an Associate Professor with School of Electrical and Electronic Engineering at the Nanyang Technological University (NTU) in Singapore. He serves as Deputy Director of the ST Engineering NTU corporate lab, which comprises 100+ PhD students, research staff and engineers, developing novel autonomous systems for airport operations and transportation. He is also involved as project PI in the Centre of Excellence for Testing and Research of Autonomous Vehicles - NTU (CENTRAN), which will lead the development of testing requirements for such vehicles, and was launched by the Land Transport Authority (LTA) and JTC, in partnership with NTU. Moreover, he serves as project PI in the BMW-NTU lab on Future Mobility, and the NXP-NTU lab on vehicle-to-vehicle communications. His research interests are in data analytics with applications to intelligent transportation systems, autonomous systems, and analysis of human behavior and physiology. He obtained the PhD degree in electrical engineering at the Swiss Polytechnical Institute of Technology (ETH) in Zurich in December 2005. He was a postdoctoral fellow at the RIKEN Brain Science Institute (2006-2007) and a research scientist at the Massachusetts Institute of Technology (2008-2010). He has been a JSPS postdoctoral fellow (2007), a BAEF fellow (2008), a Henri-Benedictus Fellow of the King Baudouin Foundation (2008), and a JSPS invited fellow (2010, 2011). His research on intelligent transportation systems has been featured by the BBC, Straits Times, Lianhe Zaobao, Channel 5, and numerous technology websites. His research team has won several best paper awards at international conferences. Besides his academic efforts, the team of Dr. Justin Dauwels also collaborates intensely with local start-ups, SMEs, and agencies, in addition to MNCs, in the field of data-driven transportation and logistics.

An estimated 64% of all travel today is made within urban environments. By 2050 the total amount of urban kilometres travelled worldwide is expected to triple, with traffic congestion potentially bringing major cities to a standstill. In Singapore, a small island with a population of 5.4 million, there are approximately 1 million cars on the roads. At the same time, roads take up 12% of land space. With the limited land space in Singapore, it is unrealistic to further increase the number of vehicles or add more roads. To address these challenges, the Singapore government plans to implement an intelligent and adaptable transport system which uses data to empower commuters and adjusts to their needs. Sensor networks are being deployed that collect data from busy areas such as traffic junctions, bus stops and taxi queues, then relay it back to the relevant agencies for analysis through data analytics and real-world applications. Besides transportation systems powered by big data analytics, driverless vehicles are also a major focus so far for the Singapore government. More than six kilometres of public roads have been opened this year for AV trials, currently in use for trials with a small fleet of public self-driving taxis. Various stakeholders are aiming for full-scale commercial autonomous taxi service in 2018 in Singapore. It is within this context that our research group has developed various data analytics and simulation tools for transportation applications. In the seminar, I will give an overview of our research efforts. Over the last years, we have been working towards scalable real-time algorithms for predicting traffic speed and travel time. The prediction systems designed by our team is able to perform accurate real-time predictions in large networks consisting of 10,000 - 100,000 links, by exploiting the correlations in traffic data. The sensing and prediction can be performed in a distributed fashion, e.g., on smartphones, as alternative to high-cost centralized systems. In recent work, we are investigating the effect of rainfall and road incidents on road traffic, in an attempt to further improve traffic predictions by incorporating information about traffic incidents and weather. We are also working towards traffic-and weather-aware online stochastic routing algorithms that are able to adapt the routes of vehicles based on real-time information about the condition of the transportation networks. Besides macro-scale data analytics, our team is designing machine learning algorithms for micro-scale transportation applications. Specifically, currently we are creating algorithms for scene understanding in urban and off-road scenarios. In collaboration with our local industry partner ST Engineering, we are integrating these technologies into autonomous vehicles (AVs) for urban mobility and airport automation. In parallel efforts, we have created a simulation platform for exploring emerging transportation paradigms. One of these technologies is vehicle-to-vehicle (V2V) and vehicle-to-infrastructure communications systems (V2X). Our simulation platform allows researchers to explore various use cases of V2V/V2X technologies at a high level of realism, including smart traffic signals and vehicle platooning. As part of the recently established Centre of Excellence for Testing and Research of Autonomous Vehicles - NTU (CENTRAN), the team is currently incorporating realistic models of AVs into the simulation platform, which will yield a sophisticated simulation tool for studying and testing AVs and designing the required infrastructure for supporting AVs. This simulation tool will be instrumental for the certification of AVs to be deployed in Singapore. The tool will also allow us to simulate and design various approaches to collect, communicate, and analyse transportation data through networks of V2V/V2X enabled AVs, providing real-time macro-scale analytics about transportation networks.

Michel Gendreau is Department Chair and Professor of Operations Research at the Department of Mathematics and Industrial Engineering of Polytechnique Montreal (Canada). He received both his M.Sc. and his Ph.D. degrees from University of Montreal. His main research area is the application of operations research methods to transportation and logistics systems planning and operation. Dr. Gendreau has published around 300 papers in peer-reviewed journals and conference proceedings. He is also the co-editor of six books dealing with transportation planning and scheduling, as well as with metaheuristics. Dr. Gendreau was the Director of the Centre for Research on Transportation (formerly CRT and now CIRRELT) from 1999 to 2007. He completed his 6-year term as Editor in chief of Transportation Science at the end of 2014. In 2001, he received the Merit Award of the Canadian Operational Research Society in recognition of his contributions to the development of O.R. in Canada. He was elected Fellow of INFORMS in 2010. In 2015, Dr. Gendreau received the prestigious Robert Herman Lifetime Achievement Award of the Transportation Science & Logistics Society of INFORMS.

The term "Dynamic Vehicle Routing Problems" (DVRP) refers to the large class of vehicle routing problems in which problem data is not completely available when the solution process is initiated and where solution determination (i.e., computation) and solution execution by the vehicles are (at least, partially) concurrent. This class of problems covers, in particular, routing problems in which customer demands arrive over a long period of time during which the vehicles are already under way to serve some requests. A typical example of this situation occurs in the area of express courier services. While classical, static Vehicle Routing Problems have now been studied for more than 55 years, the interest for DVRP's started in the late 1970's and has been steadily growing since then, largely due to the emergence of technological innovations, such as cellular phones, on-board computers, global positioning systems, etc. This has led to the development of various models and solution approaches that are able to solve effectively dynamic problems in a large variety of settings. In this talk, we will first review the main concepts relevant to the definition, analysis, and solution of DVRP's. Among other things, we will explain the differences and similarities between DVRP's and Stochastic Vehicle Routing Problems. We will then survey the most important application areas and the main solution methods that have been proposed for DVRP's. The last part of the talk will be devoted to a discussion of the research avenues that the recent developments in Big Data technologies are opening.

Virginie LURKIN is a Research Fellow at the Belgian National Fund for Scientific Research (FNRS) and a member of the Research Centre for Quantitative Methods and Operations Management (QuantOM) at HEC Management School of the University of Liege (ULg).Her research interests lie in the application of operations research to airline related business problems and the development and application of advanced models of air travel demand based on discrete choice methods. She holds a Bachelor degree in Business Engineering from HEC Management School (University of Liege, Belgium) and a Master Degree in Business Engineering from Solvay Business School (University of Brussels, Belgium).

This study formulates an itinerary choice model that is consistent with those used by industry and corrects for price endogeneity using a control function that uses several types of instrumental variables. We estimate our models using database of more than 3 million tickets provided by the Airlines Reporting Corporation. Results based on Continental U.S. markets for May 2013 departures show that models that fail to account for price endogeneity overestimate customers' value of time and result in biased price estimates and incorrect pricing recommendations. Extensions to advanced discrete choice models show the importance of accounting for inter-alternative substitution for products that share similar departure times.

Dr. Stefan Seer is leading the “Dynamic Crowd Solutions” research group at the AIT Austrian Institute of Technology in Vienna. He was a visiting researcher with the SENSEable City Lab at the Massachusetts Institute of Technology. He is the coordinator of the research collaboration between the AIT and the MIT, where he currently leads a project on “Persuasive Urban Mobility” together with the MIT Media Lab and a project on “Perception Based Modeling” together with the MIT SENSEable City Lab. He was awarded the “National Award for Traffic 2008” by the Austrian Ministry for Transport, Innovation and Technology in the category “Logistic Traffic Solutions at Major Events” for an innovative computer-aided crowd control system to optimize pedestrian flows in public transport infrastructures. He has specialized for more than 10 years in the development of models and algorithms for pedestrian flow simulation also including technologies to collect and analyze data on crowd behavior in the context of urban transportation. He has a Master's Degree in Electronics Engineering and holds a Doctoral Degree in Computer Science from Vienna University of Technology.

Microscopic simulation models are used in many applications for predicting pedestrian flows with high granularity. Current simulators do not allow for easy and quick switching between models. Moreover, reliable human movement data is still sparse, which is a prerequisite for model calibration and validation. These shortcomings inhibit to evaluate the capabilities of different models. This talk presents a unified framework for the structured investigation on strengths and weaknesses of microscopic pedestrian simulation models. The empirical baseline is a highly accurate benchmark data set measured under real life conditions in a bidirectional corridor with a novel data collection approach using the Microsoft Kinect. The proposed simulation framework is built on a scalable and flexible system architecture to easily integrate different models. The results highlight individual capabilities of seven different modeling approaches to represent microscopic and macroscopic characteristics of human movement behavior.

Bernard Gendron is a Professor at the Département d'informatique et de recherche opérationelle, Université de Montréal. He was the Director of CIRRELT, the Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (2008-2015). His research interests focus on the optimization of logistics and transportation networks. He has held positions of Principal Scientist at ILOG, Paris, and of Visiting Professor at MIT, EPFL, Pisa, Nice-Sophia-Antipolis, Blaise-Pascal, Versailles and Valenciennes. He has served as Chair of the Canadian Operational Research Society (CORS), Chair of the Montreal Chapter of CORS, and Chair of the Section on Transportation Science and Logistics of INFORMS (the Institute for Operations Research and the Management Sciences). He was awarded the CORS Practice Prize (2004), the CORS Service Award (2006) and the CORS Merit Award (2010).

In this talk, I will give an overview of the applications in transportation of network design and facility location models. In particular, I will present three research projects I am involved in that illustrate typical applications. The first one comes from the forest industry and concerns the location of logging camps for workers. The second one focuses on the location of vehicle inspection facilities, taking into account the allocation of patrols. The third one is a facility location problem for an express package delivery company.

Marc Demange holds a PhD in Computer Science from Paris I - Pantheon Sorbonne University (1994) and an Habilitation Thesis in Computer Science from Paris Dauphine University (2000). After his PhD he has held a position of Assistant Professor in Computer Science at Paris 1 Pantheon Sorbonne University. In 2001 he was appointed Associate Professor in Operational Research at ESSEC Business School (Paris - Singapore) and has held a position of full Professor from 2005 to 2014. Meanwhile he has also held several management positions at the same institution: Vice Dean of the Faculty, Associate Dean for Research and Director of ESSEC Romania Centre (in Bucharest). Since 2014 he is Associate Professor in Mathematical Sciences at RMIT University, Melbourne, Australia and in charge of the Bachelor of Science in Mathematics. During his career he has taught a large range of topics in Computer Science, Operational Research and Discrete Mathematics. His research interests, in Combinatorial Optimisation, are centred around the notion of efficient algorithms with performance guarantees, mainly polynomial approximation, complexity theory, algorithmic graph theory, online algorithms and inverse combinatorial optimisation. He is (co)author of more than 50 papers in international journals or book chapters.

In this talk I will present the minimum <em>Selective Graph Coloring Problem</em>, a generalization of the standard graph coloring problem as well as several of its possible applications and related complexity results. Given a graph with a partition of its vertex set into several <em>clusters</em>, one wants to select one vertex per cluster such that the chromatic number of the subgraph induced by the selected vertices is minimum. This problem appeared in the literature under different names for specific models. Here, I will describe different models -- some already discussed in previous papers and some new ones -- in very different contexts under a unified framework based on this graph problem. Each model motivates the problem in some graph classes and I will discuss the related complexity in these classes. I will also conclude by introducing the maximum version of the problem. This talk covers recent papers co-authored with T. Ekim (Bogazici University, Istanbul), J. Monnot (CNRS, France), B. Ries (University of Fribourg), C. Tanasescu (RMIT University, Australia) and P. Pop (University of Baia Mare, Romania).

Olivier Huber is a post-doctoral researcher at the University of Wisconsin-Madison, USA. He obtained his PhD degree in automatic control from Grenoble Université, France in 2015. He graduated from École Normale Supérieure de Cachan, France in 2011 with a M.S. degree in Electrical Engineering. His research interests include the applications of complementarity theory and variational inequalities, in particular for the simulation and control of physical systems.

We consider a convex optimization problem where the objective function is the sum f(x) + g(y) and the coupling between the variables x and y is at the constraint level. We focus on the case where g is not available in closed form and can only be evaluated at a given point by running a long simulation process. The results of interest are prices formed from the gradient of g. It is assumed that the function g is convex or can be (reasonably) approximated by a convex one. We choose to use an approximation of g defined as a pointwise supremum over a family of piecewise affine functions. This part of the procedure is carried out offline, and uses evaluations of g to define the approximation from its epigraph. We report on using the Moreau-Yosida regularization on our approximation function to return a smoothed value of the gradient that reduces the volatility in the prices. We outline some results in the context of a reserve energy market planning problem.

Wei-Shiuen Ng joined the Precourt Energy Efficiency Center at Stanford University as a postdoctoral scholar in 2014, after completing her Ph.D. in City and Regional Planning from the University of California, Berkeley. Her recent research focuses on the evaluation of transportation demand management measures, including parking pricing, and their impact on travel demand and behavior. Wei-Shiuen is currently on a leave of absence from Stanford to develop a series of policy scenarios for reducing transportation carbon emissions in Chinese and Indian cities for the International Transport Forum (ITF) at the Organisation for Economic Cooperation and Development (OECD) in Paris. Wei-Shiuen holds a Masters in Environmental Science from the Yale University School of Forestry and Environmental Studies and a Bachelor of Science in Environmental Economics and Environmental Management from the University of York. She is also on the Transportation Research Board of the National Academies Committee on Transportation in the Developing Countries.

Parking pricing plays an important role in shifting transportation demand by changing mode choice and reducing distance traveled. The price of parking is a direct cost of driving and a market based pricing policy that can efficiently manage transportation demand. In particular, variable pricing when applied to parking can have great potential, just as it has been successfully shown to manage demand through congestion pricing or peak period tolling. Understanding the impact of parking pricing on transportation and parking behavior is critical for determining the type of pricing structure that is most effective for managing demand and scarce land resources, reducing externalities, yet at the same time, generating adequate economic revenue. This study examines parking pricing impact with the consideration of various payment type, parking location, transit incentives, flexibility of work schedule, income, and walking time at the University of California, Berkeley. A discrete choice experiment was designed to analyze changes in transportation and parking behavior under various parking pricing scenarios using revealed and stated preferences data.

Elisabetta Cherchi is Associate Professor at the Department of Transport, Technical University of Denmark, where she is also Deputy Head of the Ph.D. school in Transport. She is Area Editor of Transportation, and member of the editorial board of Transportation Research part B, Journal of Choice Modelling and Transport Policy. She is also Secretary and Treasurer of the International Association for Travel Behaviour Research (IATBR). Her research interest is in data collection, in the behavioural background of demand modelling and in how to use and expand it to study emerging problems such as understanding what drives sustainable transport behaviour and how it can be promoted.

According to Crutchfield (1955) individuals consciously or unconsciously tend to "yield to group pressures" and consequently to act in agreement to the majority position. Social conformity has been extensively studied in psychology with also several applications to transport problems. Field experiments are typically used to evaluate the impact of social influence on self-reported changes toward environmentally sustainable transport behaviours. In this research, we discuss various aspects of social conformity and present a stated preference experiment set up to measure their effect on individual preferences. The choice of electric cars is used as an illustrative example. In particular, we explicitly measure how individuals' preference change before and after they have received social information on other's experience about driving range, about the need to adapt the activity schedule and about the benefit of parking policies. The effect of descriptive norm and other-signalling concern are also measured as part of the stated preference experiment, while injunctive norms are measured with typical statements on a 7-point Likert scale. Results from the estimation of mixed logit model and hybrid choice models, clearly confirms that the experience (especially negative) of other people has a powerful effect on individual preferences for range and parking policies. Results also confirm that individuals' behaviour is affected by the image they want other people to have of them, making them "more honest" in their answers.

C. Angelo Guevara is associate professor at Universidad de los Andes in Chile; research affiliate of the Intelligent Transportation Systems (ITS) laboratory at the Massachusetts Institute of Technology (MIT); and external affiliate of the Choice Modelling Centre (CMC) at the University of Leeds. He holds an MSc in transportation from Universidad de Chile, as well as an MSc and a PhD in the same area from MIT. He has been awarded the Fulbright and the Martin-Family fellowships, as well as the honorable mention of IATBR's Eric Pas dissertation prize. His main research interest is in the modeling of choice behavior, with recent contributions on endogeneity, sampling of alternatives, behavioral economics.

Empirical evidence suggests that, under some circumstances, the introduction of a new option in a choice-set can increase the choice probability of other alternatives. This result, known as the decoy effect, defies the basic regularity assumption, which is at the root of standard models of choice that are based on a compensatory approach under the Random Utility Maximization (RUM) framework. The goal of this research was threefold. First, we worked toward the development of a practical probabilistic choice-model that could account for the decoy effect, building upon various types of choice behaviors that that been described in cognitive psychology. Then, we used the proposed choice model to study, with Monte-Carlo simulation, the power of different statistical tests for detecting the presence of this phenomenon. Finally, we designed and applied a Stated Preferences (SP) survey to detect and to characterize the decoy effect in route choice. Results of this research showed first that all the decoy effect types that have been described in the literature, can be replicated by the Random Regret Minimization (RRM) discrete-choice model. Regarding statistical testing for the presence of the decoy effect, we found that McNemar and Proportions tests showed larger power when the effect size was modeled as RRM. Finally, four conclusions were driven from the application of the SP survey. The first was that the decoy effect was present in route choice, but that it was hard to detect it in the context of commuting trips or when alternatives were far from the true trade-off line. The second result of the SP experiment was that the magnitude of the average sample effect obtained from it was coherent with a data generation process based on the RRM model. Third, the SP survey showed that the larger decoys found were of the compromise type, and that the more robust ones were those of the range type. Finally, the SP survey indicated that, although an emergent-values Logit model showed slightly better fit, the RRM had substantially superior performance in outer-sample forecasting. This final result suggests that the RRM does capture, to some extent, the underlying behavior that is causing the decoy effect, but that this choice-model may still be somehow incomplete for this purpose. Four future steps of this line of research can be identified. The first is to improve the RRM model. The second step corresponds to the design and application of a Revealed Preference (RP) experiment to detect the decoy effect in real transportation behavior. The next, is to deepen the analysis of the circumstances under which the decoy effect occurs. The final step corresponds to the study of possible transportation public policies that can benefit from the decoy effect, such as seated-only buses to favor the use of public transportation or different pricing strategies.

Nelson Maculan is full Professor of Optimization at the Dept. of Systems Engineering and Computer Science, Graduate School of Engineering, COPPE, Federal University of Rio de Janeiro, Brazil. He is the President of the International Federation of Operational Research Societies, IFORS (since Jan 2013), was the State Secretary for Education, Rio de Janeiro State, Brazil (Jan 2007-Feb 2008), the National Secretary of Higher Education, Ministry of Education (SESu-MEC), Brasilia, Brazil (Feb 2004-Dec2006) and the President of the Federal University of Rio de Janeiro (UFRJ), Brazil (Jul 1990-Jul 1994).

The Euclidean Steiner tree problem (ESTP) in Rn consists of finding a tree of minimal Euclidean length that spans a given set of points in Rn, using or not additional points. Only a few papers consider the exact solution for the ESTP in Rn (n>2) and there are just two works that considered a mathematical programming formulation for the ESTP. One of them presented a convex mixed integer formulation that could be implemented in a Branch and Bound algorithm. This work presents techniques to improve the performance of the algorithm in order to implement this formulation. After, we present a mathematical programming model for the problem of covering solids by spheres of different radii. Given a set of spheres, possibly with different diameters, and a solid, the goal is to locate the spheres in such a way their union forms a coverage for this solid, using the smallest possible number of spheres of this set. This problem has an application in the radio-surgical treatment planning known as Gamma Knife and can be formulated as a non-convex optimization problem with quadratic constraints and a linear objective function.

Yoram Shiftan is a Professor of Civil and Environmental Engineering in the Technion, where he teaches and conducts research in travel behavior with a focus on activity-based modeling and response to policies, the complex relationships between transport, the environment and land use, transport economics and project evaluation. Prof. Shiftan is the editor of Transport Policy and the chair of the International Association of Travel Behavior Research (IATBR). He is the co-chair of the Network on European Communications and Transport Activities Research (NECTAR) cluster on Environment and Policy, member of the World Conference Transportation Research (WCTR) scientific committee, and chair of its Transport Security Special Interest Group. Prof. Shiftan received his Ph.D. from MIT and since then has published dozens of papers and co-edited the books "Transportation Planning" in the series of Classics in Planning, and "Transition towards Sustainable Mobility, The Role of Instruments, Individuals and Institutions."

Activity-based models are the new generation of travel demand models. These models treat travel as being derived from the demand for personal activities. The explicit modelling of activities and the consequent tours and trips enables a better understanding of travel behaviour and more credible analysis of response to policies and their effect on traffic and air quality. These models have various advantages in support of transport project evaluation by being able to provide detailed disaggregate individual and vehicle activity output that can improve our analysis of emissions and provide various accessibility measures important for equity and other economic evaluation. This presentation will demonstrate these various advantages for policy-making and discuss future challenges in using these models to forecast the impact of new transportation services and technologies promoting sustainable.

Professor at the department of computer science and operations research of the University of Montreal, Patrice Marcotte has published over 80 articles related to network design, variational inequalities, bilevel programming, network pricing and revenue management, traffic assignment, and badminton. He his also the author of two guides and a large website devoted to cycling, and currently sits on the editorial board of the following journals: Operations Research, Transportation Science, Journal of Optimization Theory and Applications, Operations Research Letters, European Journal of Combinatorial Optimization.

We consider the problem of setting profit-maximizing tolls on a subset of arcs of a multicommodity transportation network. The case where users are assigned to cheapest paths, which is naturally formulated as an NP-hard bilevel program, has been extensively studied and will serve as the background for an extension where user assignment is performed according to a discrete choice model of the logit family. Following a description of the model and its theoretical properties, we develop an algorithmic framework for determining a near-optimal solution of this nonconvex problem, based on a variety of approximations involving mixed integer programs, either linear or quadratic. Through a battery of tests performed on a variety of network topologies, we reach the conclusion that very crude approximations (that scale well) perform surprisingly well.

Asad Lesani completed his M.Sc and B.Sc. in Electrical Engineering at University of Tehran, Iran in 2012 and 2009 respectively. He is doing his PhD in Civil Engineering with the filed of Transportation Engineering in Department of Civil Engineering, McGill University, Montreal, Canada. He is a member of CIRRELT (Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation).

Intelligent transportation systems depend on technologies to obtain valuable road metrics, such as travel times, speeds, and volumes. Novel ways of collecting anonymous data from road users across multiple modes are becoming more recognized in literature and industry. Bluetooth detectors have been widely researched as a way of detecting smartphones and vehicles while maintaining anonymous identity across multiple detection sites. This paper proposes a smartphone detection system using wireless Internet (WIFI) signatures from mobile devices in a similar way to Bluetooth, but with a higher detection rate due to the higher usage of WIFI over Bluetooth. The system is tested on mixed-mode and pedestrian-only facilities with 9-20% accuracy for vehicular traffic and greater than 20% accuracy for pedestrian-only routes with multiple sensors. These initial findings look promising, making the possibility of building a combined WiFi/Bluetooth system that take advantages of both sources of data.

Luis is an Associate Professor in the Department of Civil Engineering, McGill University. His specialty is in transportation safety, data collection and monitoring methods and technologies, and sustainable transport strategies. His research interests include the development of crash-risk analysis methods, the development and integration of systems for traffic monitoring, the impact of climate on mobility, energy efficiency measures and non-motorized transportation.

The seminar presents the current development of a smartphone-based system for collecting and analyzing route GPS data coming from cyclists and drivers routes. Using GPS functionality on Android and iOS smartphones to log route data, a large database containing thousands of trips is collected in a short period of time after a mass-media campaign in Canadian cities. For motorists, a platform is built for mapping traffic congestion using link-level indicators such as speeds or travel times, speed differentials, etc. For cyclists, a platform is built for mapping bike activity and injury risks in the entire network, which is used for identifying routes and/or intersections at high risk and identifying bicycle infrastructure needs. Work in progress is discussed at the end of this presentation.

Vinayak is the Deputy Director of the Research Center for Integrated Transportation Innovation (rCITI) and a Senior Lecturer at the School of Civil Engineering, UNSW Australia. His key area of interest is in studying the risks in transport systems from the point of view of linking the physical risks from the point of view of traffic dynamics with peoples’ choices in response to these risks, and how these interactions impact traffic safety, emergency management and traffic congestion.

Transport systems are inherently risky, ranging from risk in travel time to risk in crashing. This presentation provides an overview of my current work on the use of Experimental Economics in understanding behaviour with respect to interactions of risk attitudes and subjective beliefs on crash propensity, route choice and information, and finally evaluating policies in emergency management. Experimental economics has been used intermittently since the 1960s to conduct controlled experiments to evaluate policies, but recently has gained significant interest to test theories. The presentation will also present trends, strengths and pitfalls of these methods

Mohammad Yousef Maknoon is a postdoctoral fellow at Polytechnique Montreal - CIRRELT (interuniversity Research center on Enterprise Networks, Logistics and Transportation). His research focuses on scheduling and optimization in transportation network. He completed his Master and Ph.D. in Industrial engineering (Operations research) at Polytechnique Montreal.

Cross-dock is a center that transships freight between trucks with minimal usage of storage in between. Cross-docking reclaims transportation efficiency by bundling arriving freight into full truckloads. Cross docks are beneficial as long as they are well coordinated with transportation plans when their operational cost does not overwhelm their savings in inventory. In this talk, I present scheduling problems at macro and micro level and show variety of models to address their issues.

Song Gao is an associate professor of Civil and Environmental Engineering at the University of Massachusetts Amherst. Dr. Gao's research focuses on optimization in stochastic networks, econometric and psychological models of travel behavior, equilibrium analysis of stochastic networks with traveler information, with applications in intelligent transportation systems (ITS), transportation planning under both normal and emergency conditions, and sustainable transportation systems. Prior to joining the faculty of the University of Massachusetts Amherst in 2007, Dr. Gao worked as a transportation engineer at Caliper Corporation, Newton, MA for three years, and developed advanced traffic assignment modules for TransCAD, a GIS-based transportation planning software and provided consultancy to transportation demand forecasting projects of state, regional and local planning agencies. She has published in scientific journals such as Transportation Research Parts A, B, and C, IEEE Transactions on ITS and Transportation Research Record. Dr. Gao has obtained over $1M in external research funds as the principle investigator (PI), and over $1.8M as a co-PI from federal and state governments, regional consortia, planning agencies, and private foundations. Dr. Gao was a member of the winning team of the 2010 MacArthur Digital Media and Learning Competition. She received an honorable mention (second place) in the INFORMS (Institute for Operations Research and Management Science) Transportation Science and Logistics Dissertation Prize Competition in 2005. Dr. Gao is a member of the Transportation Research Board (TRB) Committees on Travel Behavior and Values (ADB10) and Transportation Network Modeling (ADB30), and Chair of the TRB Route Choice and Spatio-Temporal Behavior Subcommittee (ADB10(2), ADB30(3)). She is on the editorial board of the Journal of Intelligent Transportation Systems. Dr. Gao received her Ph.D. and M.S. in Transportation from Massachusetts Institute of Technology in 2005 and 2002 respectively. She received her B.S. in Civil Engineering from Tsinghua University of China in 1999.

Transportation systems are inherently uncertain due to disruptions such as bad weather and incident, and the randomness of traveler' choices. Real-time information allows travelers to adapt to actual traffic conditions and potentially mitigate the adverse effect of uncertainty. Both algorithmic and behavioral studies of adaptive routing are presented. A series of optimal adaptive routing problems are investigated, where time-dependent travel times are modeled as correlated random variables and various assumptions on the real-time information accessibility are made. Behavioral studies of adaptive route choice in both one-shot and day-to-day learning contexts based on stated preferences data show that travelers can plan ahead for traffic information not yet available. Two modeling paradigms for route choice under unreliable travel times, utility maximization based on the prospect theory and non-compensatory heuristic, are compared. The non-compensatory heuristic is found to be potentially a suitable alternative to the conventional utility maximization approach. The ongoing work of developing a history-dependent route choice learning model for realistic networks is also discussed.

Sofia is a third year PhD student of the MIT-Portugal program in Lisbon (Instituto Superior Tecnico -IST). Her research project deals with flexibility of airport passenger buildings. She holds a Diploma in Civil Engineering from National Technical University of Athens (NTUA)and an MSc from IST in Complex Transport Infrastructure Systems. In her thesis projects she evaluated the performance of Greek airports with Data Envelopment Analysis (NTUA) and she developed a pedestrian model to simulate passenger movements in Lisbon Portela airport and evaluate the airport processes (IST). Her research interests focus on transport terminal planning and management, non-motorized transportation, behavior and activity models and operations research.

In spite of the transport mode, all the travellers are primarily pedestrians. Transport terminals are considered as infrastructure planned for pedestrians and capable of employing tools that will let them respond with flexibility and efficiency to future challenges. It is suggested that pedestrian flows should have an active role in the definition of the flexibility of a transport building configuration and that pedestrian behavior should be integrated in this process at an early stage. This talk aims to present some preliminary thoughts on the structure of passenger route choice models for an airport building and the way that they can be used in airport flexibility analysis. Travel time, wayfinding, space characteristics and available free time are important for passenger route choices and pillar sources for the specification of passenger route choice models. This combination allows us to explore any latent relationships between space characteristics, route choices and flows of passengers, to model more efficiently pedestrian route choices and to designate the properties of efficient terminal configurations. In terms of planning policy and flexibility, it gives some indications for the value of each area as derived from the way the passengers perceive it and the value they add to it. In this way route preferences can be incorporated in flexibility analysis as inputs that have the potential to indicate areas that are often preferred.

Shadi Sharif Azadeh completed her Masters and Ph.D. in Mathematics (Operations Research) at Polytechnique Montreal in 2013. She is a member of GERAD (Group for Research in Decision Analysis) and CIRRELT (Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation). She is presently a postdoctoral fellow at Polytechnique Montreal.

In transportation companies, demand forecasting is crucial. To maximize revenue, these systems use historical data. However, due to booking limits, registered reservations do not represent the real demand. We first present a comprehensive review on different aspects of demand modeling in the context of revenue management systems. Then, we propose a new non-parametric global optimization approach which is able to model demand by using choice probabilities. Our proposed model is able to extract seasonal features of demand and customer utilities for a given product. Finally, in a comparative study, we investigate the impact of different methods of customer preference estimation on revenue.

Gunnar completed his PhD at TU Berlin in 2008. He was a post-doc at TRANSP-OR from 2009 until 2011. Since 2011, he is on a junior faculty position in transport modeling at KTH.

We derive an utmost simple microscopic queueing model of bidirectional pedestrian flow through long channels. The model is to the extent possible consistent with the Kinematic Wave Model and requires the calibration of only four parameters. Its fundamental diagram is derived and compared to real data (joint work with Gregor Laemmel, FZ Juelich).

Iliya Markov is an assistant HES at the Haute école de gestion, Geneva. He has an MSc in Operational Research with Finance from the University of Edinburgh and a BA in Mathematics and Economics from the American University in Bulgaria. In recent years, Iliya has published several book chapters and journal articles in the fields of finance and optimization. He is currently working in the field of operational research, and in particular vehicle routing, and would like to pursue an academic career in this domain.

An eco-logistical company manages a set of several hundred eco-points, with a number of waste containers at each eco-point. Certain containers from one eco-point may need to be moved to another one or to a repair facility if they have been damaged. Moreover, every container needs to be cleaned on site over a planning horizon. Both problems need to be solved by taking into consideration important constraints such as service/cleaning time at the eco-point and vehicle capacity for the movement problem, so that the tour lengths are minimized as much as possible and their durations respect the driver's daily working time. The container movement problem is solved as a dial-a-ride problem. A genetic algorithm is used to find a good assignment of movements to vehicles. The quality of the assignments is determined by a routing heuristic which considers the specified problem constraints and provides the resulting route's cost. The cleaning problem is solved with minimum changes to the above methodology. The algorithm has been tested on real data and is currently being integrated by the client company. The algorithmic part is complemented by a graphical user interface with route plotting capabilities.

Dr. Weihua Gu is the Deputy Director of the UC Berkeley Center for Future Urban Transport, who also served as a lecturer at the Department of Civil & Environmental Engineering, UC Berkeley. His research interests include public transit systems, multimodal urban transportation systems, freeway traffic operations, queueing models, and infrastructure management. He received his B.S. degree and M.Eng. degree in Civil Engineering from Tsinghua University (Beijing, China) in 2002 and 2005; and his Ph.D. degree in Transportation Engineering in 2012, the M.A. degree in Economics in 2011, and the M.Sc. degree in Industrial Engineering & Operations Research in 2010, all from UC Berkeley. His awards include a Gordon F. Newell Award for Excellence in Transportation Science (given by UC Berkeley Transportation Engineering Faculty), and a Chinese Government Award for Outstanding Self-Financing Students Abroad.

The global-wide trend toward urbanization, with its attendant traffic congestion and environmental degradation, calls for enhancements to public transit and other green modes of travel. In the near term, this means that greener modes must compete more effectively with, and coexist more harmoniously with, automobile traffic. The presentation explores ideas in this realm, beginning at the scale of a city block. I first focus on busy, multi-berth bus stops where multiple bus lines converge and on the bus queues that form when these stops have insufficient numbers of berths. Queueing models are developed to predict the bus-carrying capacities of these stops, and analytical solutions are derived by exploiting renewal processes that are embedded in the unique operating features of serial bus berths. The resulting models can be used to estimate a stop’s number of berths needed to achieve a target capacity and to determine the conditions when bus maneuvers into or out of berths should be regulated. I then examine bus stops that reside near signalized intersections, where dwelling buses can impede cars from discharging into or out of the intersections. Using kinematic wave theory, analytical models are formulated, both to quantify negative impacts and design mitigation strategies. The presentation ends with brief discussion of ongoing work, e.g., to improve travel along corridors by dispatching buses in platoons; and to design greener city-wide transport networks by integrating bicycle-sharing systems with public transit.

Riccardo Scarinci is a PhD student at the Centre for Transport Studies at University College London. His current research on Cooperative Intelligent Transport System is part of the European Commission FP7 project NEARCTIS, and it has been developed in collaboration with the Technical University of Delft – The Netherlands. Before joining UCL he worked as research assistant at the Laboratory for Mobility and Transport at Politecnico di Milano – Italy, investigating ex-ante and ex-post evaluation of ITS for the European Commission projects EasyWay and 2DECIDE.

In the area of Active Traffic Management, Intelligent Transport Systems (ITS) are widely used to reduce congestion on motorways. In this vein, emerging technologies continually offer new communication capabilities that can be used to improve performance. The aim of this study is to develop a novel Ramp Metering strategy, exploiting communication capabilities to further reduce congestion at motorway junctions. This new system rearranges gaps present in the motorway traffic by requesting cooperation from participating vehicles in order to facilitate merging. Macroscopic traffic flow theory is used to develop the control algorithm, and microscopic simulation of individual vehicles is used to evaluate the system traffic performance. Results indicate reduction of congestion occurrence, improvement in merging capacity and increase in travel time reliability. This shows how the use of new communication technologies can improve the performance of current ITS and lead to a reduction of congestion on motorways.

Joan Walker's research focus is behavioral modeling, with an expertise in discrete choice analysis and travel behavior. She works to improve the models that are used for transportation planning, policy, and operations. Professor Walker joined UC Berkeley in 2008 in the Department of Civil and Environmental Engineering and as a member of the interdisciplinary Global Metropolitan Studies initiative. She received her Bachelor's degree from UC Berkeley and her Master's and PhD degrees from MIT. Prior to joining UC Berkeley, she was Director of Demand Modeling at Caliper Corporation and an Assistant Professor at Boston University. She is Chair of the Transportation Research Board's Committee on Transportation Demand Forecasting, an Associate Editor of Transportation Science, and a recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE).

<p>Latent modal preferences, or modality styles, are defined as lifestyles built around the use of a particular travel mode or set of travel modes. Traditional models of travel mode choice assume that (1) that all individuals are aware of the full range of travel modes at their disposal and make a rational mode choice based on level-of-service and (2) individual modal preferences are characteristics of the individuals that are exogenous to the choice situation and stable over time. Though these assumptions simplify the model, they risk overlooking the impact of more deeply entrenched individual variations in modal preferences. </p><p> This talk presents a latent class choice model (LCCM) that allows modal preferences to be endogenous to the choice situation and variable across individuals. The model is applied to analyze modality styles and travel mode choice behavior of 25,000 people in the San Francisco Bay Area. The study identifies six distinct modality styles in the sample population that differ in terms of their taste parameters and choice sets. Most notably, nearly a third of the sample is found not to consider any mode other than auto. Results show that an individual's value of time is sensitive to the level-of-service, and an increase in congestion can induce decision-makers to lower their value of time. Findings further reveal that incremental improvements in the transportation system result in far smaller changes in travel behavior than predicted by traditional models; what is needed is a dramatic change to the transportation system that forces individuals to reconsider their modality styles. </p><p> This is joint work with UC Berkeley doctoral candidate Akshay Vij.</p>

Ricardo Hurtubia is assistant professor at the Urbanism Department of the University of Chile. He received his PhD (Mathematics) from EPFL in 2012 and a Master's degree (Transportation Engineering) from the University of Chile in 2006. His research interests include modeling behavior in the urban context and simulation of land use and transportation systems.

A method to use psychometric indicators in the estimation of latent classes for discrete choice models is presented. An extension of the method is proposed in order to model the preferences for elements of the city that can't be completely described in terms of quantitative attributes, like open/public spaces or street profiles. The method attempts to use latent constructs accounting for qualitative or hard to measure attributes of the alternatives by relating them with the perception of users, captured through psychometric indicators.

Franziska Borer Blindenbacher is an economist and works as an international consultant and teacher in the fields of transport, spatial planning and the environment. She also works part-time as a scientific adviser for the Swiss Federal Office for Spatial Development (ARE). Among her clients in the public and private sector are the Canadian, the US as well as the Swiss Ministry of Transportation, the Swiss Association for Public Transportation, PostAuto Switzerland, the Institute for Transportation and Development Policy (ITDP) in New York and Washington DC, and the European Institute for Sustainable Transport (EURIST) in Hamburg. Ms Borer Blindenbacher was part of the team that developed and implemented the distance- and emission-related heavy vehicle fee (HVF) in Switzerland as well as the federal “Agglomerationprogram” coordinating transportation infrastructure and spatial development in metropolitan areas. Ms Borer Blindenbacher is the author of various publications in the field. Among them the case study From an Integral Transport Concept to Financing Urban Transit - The Swiss Approach (2008). Further the Study of Methods of Road Capital Cost Estimation and Allocation by Class of User in Austria, Germany and Switzerland (2007). She teaches at different national and international academic institutions such as the George Washington University in Washington DC and the University of Applied Sciences of Eastern Switzerland (HSR) in Rapperswil.

Various countries have meanwhile adopted market incentives to meet transport needs in an economically efficient way. The seminar will touch on different types of approaches for financing passenger and freight transportation, and for addressing transportation-related externalities. There will be a special emphasis on the unique Swiss Heavy Vehicle Fee (HVF), which will serve as an example of the practical and political issues that need to be addressed in order to implement sustainable national transportation policies. The objectives of this seminar are to encourage students to familiarize themselves with a range of transportation and related environmental, social and economic policy issues and to develop an understanding of the complexity, inter-connection and potential resolution of some of these issues.

Luis Miranda-Moreno is an assistant professor in the Department of Civil Engineering at McGill University. His research interests include road safety, non-motorized transportation, and the relationship between transportation and the environment. More specifically, his research interests include developing methods and tools for crash risk analysis, monitoring and modeling pedestrian and bicycle flows, as well as identifying strategies to reduce fuel consumption and greenhouse gases. He is currently collaborating on several projects for transportation agencies in Canada, the US and Mexico, including Transport Canada, the Ministries of Transportation of Ontario and Quebec, the National Cooperative Highway Research Program, and the City of Montreal.

Monitoring and analyzing non-motorized (pedestrians and bicycle) flows over time and space in a urban road network is essential for various reasons, including i) evaluation of the impacts of new infrastructure, programs and policies to encourage cycling and walking, ii) identification of current traffic patterns and prediction of future demand for the planning, design and operation of facilities, iii) mapping injury risk for the identification of dangerous facilities and inappropriate designs, etc. This work provides a discussion of the non-motorized traffic data needs and the emerging technologies for traffic monitoring and data collection. Using the City of Montreal as an application environment, this work also presents a framework for monitoring and modeling safety and mobility of non-motorized flows in an urban environment. As an input data, automatic long-term and manual sort-term traffic counts are combined. This framework allows identifying factors affecting non-motorized traffic flows and safety, such as weather, built environment and road designs. The proposed framework will also help monitoring changes in non-motorized mobility and safety in the study network.

Bruno F. Santos is an Assistant Professor at the University of Coimbra, Portugal. He finished is PhD in 2009 on the topic of Transportation and Spatial Planning, in a collaboration between the University of Coimbra and the University of Toronto, Canada. He had published several works on transportation planning topics and has been involved, besides other research initiatives, in the MIT-Portugal research and teaching program.

The strategic decisions with regard to the investments in transportation infrastructures usually involve large amounts of money. Therefore, decisions should be carefully analyzed. Cost-benefit analysis, usually based on trial-and-error approaches, does not allow the full exploration of the solution space. This can only be done using optimization techniques. This talk will give an overview of the work developed by the author on the usage of optimization approaches to solve strategic transportation investment planning problems. The different approaches will be illustrated with the application to several case studies, including the definition of national/regional road network plans and the location of freight intermodal terminals in Belgium.

Dick Ettema is associate professor in the Department of Human Geography and Planning in Utrecht University. He received a PhD from Eindhoven University of Technology with a thesis “Activity-based Modeling” and edited the book “Activity-based Approaches to Travel Analysis” with Harry Timmermans. Since then, he has published widely on activity based analysis, activity based modeling and time-use studies. Dick’s recent work has focuses on the use of active travel modes and the impact of travel and activities on well-being.

Studies of walking behavior have gained momentum over the past years, due to improved data collection techniques and further development of modeling approaches. In most cases, such studies emphasize the detailed movement of pedestrians in relation to aggregate pedestrian flows, and decisions regarding route or destinations during a walking trip. This presentation aims to look upon walking behavior from a broader perspective, by discussing the options of applying an activity based approach to walking behavior. In particular, it will be discussed to what extent decisions made regarding the daily (or longer term) activity pattern influence decisions regarding the walking trip (such as where to walk, for how long, with whom), decisions during the walking trip (such as routes, places to visit and activities to pursue while walking) and the experience of the walking trip (how pleasant, stressful etc.). It will be argued that looking upon walking from this broader perspective gives rise to the inclusion of addition variables in our analyses, including functional characteristics of places and routes (amenities, functions) as well as ambient conditions and aesthetics.

Evanthia Kazagli holds a Diploma of Rural and Surveying Engineering from the National Technical University of Athens (NTUA). She is currently a Master student in the division of Traffic and Logistics at the Royal Institute of Technology (KTH), where she is elaborating her thesis under the supervision of Prof. Haris Koutsopoulos. Her work deals with the estimation of arterial travel time distributions. Her research interests include among others intelligent transportation systems and travel time estimation and reliability.

<p>Automatic Number Plate Recognition (ANPR) data have been widely used for estimation of travel time and travel time distributions, mainly in the case of freeways. The objective of this work is to formulate a finite mixture model for the estimation of arterial travel time distributions based on ANPR data. A black spot when extracting arterial travel times from ANPR data concerns vehicles that do not traverse the monitored section directly, but stop in between for various reasons (loading/ unloading, buses stopping at bus stops etc), resulting in higher than the usual travel times (invalid observations). Assuming that the population of ANPR travel times is generated by two different subpopulations (components) -one deriving from non-stopped (valid) vehicles and one from the stopped- finite mixture models can be used at a first level as clustering technique to separate these two components. In an attempt to reinforce the model, explanatory variables such as weather conditions are included in the estimation. In arterial networks, route travel times are likely to vary among the valid observations, even over small time intervals. This is due to such factors as traffic lights, buses stopping, vehicles turning mid-link delaying following vehicles, etc (Robinson, 2005), resulting in a multimodal probability density function of travel time. In this context mixture models can be used - at a second level - for the estimation of route travel time distribution. A very important aspect is the assumption for the underlying distribution. The common assumption of normal distribution of travel time is replaced by log-normal (in this case mixture of log-normal distributions). </p><p> Robinson, S. (2005). The development and application of an urban link travel time model using data derived from inductive loop detectors, PhD Thesis, Centre for Transport Studies, Department of Civil and Environmental Engineering, Imperial College London, United Kingdom. </p>

Florent Hernandez is a Postdoctoral reseacher in Operation Research and transportation science at Ecole polytechnique de Monréal and CIRRELT, Canada, since october 2011. He recieved his Ph.D thesis in Computer science from the Université Montpellier 2, France, in november 2010. Between these dates, he was teaching assistant at the Université Montpellier 2. Even if his PhD research topic was focused on designing a Branch and Price algorithm to solve a spraying agricultural problem, his main research topic is more generally expanded to the combinatorial optimization methods used in transportation science.

The multi-trip vehicle routing problem with time windows is a variant of the classical vehicle routing problem with time windows, where a vehicle can perform several trips during the planing time horizon. Multiple trips are beneficial to the carrier by limiting the number of vehicles and drivers necessary to the deliveries, especially in the cases where vehicle capacities, distances or time constraints naturally imply short distribution routes. However, this feature generate many combinatorial issues due to the mutual exclusion constraints that appear between two trip performed by the same vehicle. We propose a set covering formulation, where variables represent trips, and an exact method based on Branch and Price scheme to solve it. Computational results on Solomon's benchmarks evaluate the method on small to middle-sized instances.

Carlo L. Bottasso is a Professor of flight mechanics at the Department of Aerospace Engineering of the Politecnico di Milano in Milano, Italy, where he directs the Poli-Rotorcraft and Poli-Wind research labs. Dr. Bottasso earned a Ph.D. in Aerospace Engineering from the Politecnico di Milano in 1993, and did postdoctoral work at the Rensselaer Polytechnic Institute in the USA, before returning to the Politecnico in 1996. In 2003-2005 was on the faculty of the Georgia Institute of Technology, and he has held visiting positions at numerous institutions, including Aalborg University in Denmark, the National Renewable Energy Laboratory (NREL), the Lawrence Livermore National Laboratory, NASA Langley and others. His research interests are in multibody dynamics, aero-servo-elasticity and control. Specific interests are in flexible multibody dynamics with application to the modeling of rotary wing vehicles and wind turbines, and corollary modeling and numerical technologies, including system identification, model reduction, methods for the solution of algebraic-differential equations, non-linear finite elements, adaptive and optimal control. On these topics he has co-authored over 250 publications, including 90 peer review journal papers.

Helicopters and wind turbines are very different complex engineering systems, that however share some common physical processes. For both, rotating flexible blades interact in a highly unsteady manner with air to make a vehicle fly in one case and to generate energy from wind in the other. In this talk we will first review some mathematical models that are used for the simulation of both systems. Then, we will describe some challenging applications, ranging from the simulation of helicopters flying extreme maneuvers at the boundaries of their flight envelope, to the passive and active mitigation of loads on large wind turbines. These examples will help highlight the differences between these two engineering systems and will illustrate their very different design drivers. In both cases, the ability to model the relevant coupled physical processes to a high level of fidelity is key for achieving the ever more ambitious design goals posed by industry.

Marija Nikolic received my MSc degree in Electrical and Computer Engineering from the Faculty of Technical Sciences, University of Novi Sad, Serbia, in 2010, where she got a prestige award for the best student at the promotion of the Faculty of Technical Sciences in 2010. She has been working for 2 years as a software engineer in “Model Server and CIM Integrations” team for TelventDMS company from Novi Sad, Serbia.

SCADA and utility management (SCADA-UMS) systems often use multiple, distributed data sources. Thereby, each data source usually needs to be read individually via its specific data access possibilities. This paper presents a solution for centralized and generic access to data from various data sources utilized by a SCADA-UMS system. The solution also offers a possibility of complex filtering data from different sources. The SCADA-UMS system, which this solution is implemented for, is a smart grid software system for monitoring and management of a power distribution network. The implementation of the solution is based on the Generic Data Access (GDA) specification, issued by IEC 61970-403 Ed.1: Energy management system application program interface (EMS-API) - Part 403: Generic data access, final draft, 2008. The main part of the solution is the GDA Smart Proxy component, which behaves as the central data source from the system's point of view. GDA Smart Proxy combines data from existing sources and implements a module for complex filtering.

Maya Abou-Zeid is an Assistant Professor of Civil and Environmental Engineering at the American University of Beirut and a research affiliate of the Intelligent Transportation Systems Program at the Massachusetts Institute of Technology. Her research interests include travel behavior modeling, urban transportation planning, market research, and road safety.

<p>Travel behavior models predict individual decisions related to various travel dimensions, such as the choice of auto ownership, activity patterns, destinations visited, modes of transportation, routes, and departure times. They are used as part of decision support systems by transportation planners and policy makers to predict travelers' responses and traffic impacts resulting from infrastructure (e.g. road widening), operational (e.g. public transportation service improvement), or policy interventions (e.g. congestion pricing). The accuracy of these decision support tools depends on the richness embedded in the travel behavior models. </p><p>This talk will give an overview of some recent developments in the area of travel behavior modeling with applications to modeling time-of-travel choice accounting for the cyclicality of time-of-travel, modeling heterogeneity across individuals in the value of travel time savings arising from different attitudes towards travel modes, and modeling mode switching from car to public transportation in relation to travel happiness using recent experiments conducted with habitual car drivers in Switzerland and in Boston. The applications show how insights from behavioral theories can be used to enrich the travel demand models and make them more policy-sensitive.</p>

Mr. Chen Jiang Hang obtained B.Eng from the Department of Civil Engineering, Tsinghua University, PR China in 2007 and received his PhD from the Department of Civil and Environmental Engineering, National University of Singapore in 2012. Currently, he is a research engineer at The Logistics Institute - Asia Pacific, National University of Singapore. His PhD research topic focuses on port container terminal operation management especially for the quayside operation problem which includes the berth allocation problem, the quay crane assignment problem, and the quay crane scheduling problem.

The swift pace of globalization has significantly increased the demand for containerized maritime transport services. Under the atmosphere, the competition among port container terminals has become acute and such drives the managers in port container terminals to maintain seamless flows of containers through terminals while to keep the operational costs as low as possible. To this end, operational research methods have received considerable importance for the operations management in port container terminals. This seminar will cover Mr. Chen's studies on quayside operation problem in port container terminal including both mathematical modeling and algorithm development. Firstly, for mathematical modeling, there are two highlights in Mr. Chen's studies: 1) the technology updates and innovative implementations have been reflected; 2) the integration issues to synchronize the decision-making processes for each key component of the quayside operation problem have been stressed. Secondly, for problem solving, a spectrum of methods has been devised to handle the proposed models. In this seminar, Mr. Chen will also introduce the problem-oriented meta-heuristics, approximation algorithms, and exact algorithms (like Benders' Cut Method) developed for the quayside operation problem.

David Rey is a PhD candidate in Operations Research at the Transport and Traffic Engineering Laboratory (LICIT) at the French Institute of Science and Technology for Transport, Development and Networks (IFSTTAR) in Lyon (France). He obtained a Master degree in Applied Mathematics at the Pontifical Universidade Catolica (PUC) in Rio de Janeiro and started a PhD thesis in OR applied to air traffic management in September 2009. His research topic is focused on the minimization of potential air conflicts using speed control.

We address the speed regulation problem in air traffic management. As global air traffic volume is continuously increasing, it has become a priority to improve air traffic control in order to deal with future air traffic demand. During this decade, European and United States initiatives were launched to design the future of air traffic management. One of their objectives is to increase air traffic density and optimize flight route plans. This can be achieved through en-route deconfliction. Potentials air conflicts occur when two or more aircraft are predicted to be below a separation norm in a near future. Such situations highly affect air traffic controllers’ workload thus limiting their capacity to deal with great numbers of aircraft and inducing flight delay. Reducing air traffic controllers’ potential workload through speed regulation has been investigated in the En-Route Air traffic Soft Management Ultimate System (ERASMUS) project. Efficiency of the method has been validated through simulations including humanin- loop experiments thus opening the door to conflict resolution algorithms based on speed regulation. We propose an optimization-oriented formulation for the speed regulation problem. We start by presenting a potential conflict detection and resolution framework. Uncertainty is then introduced in the model, aiming at reproducing realistic air navigation conditions. We conclude with a case study on real air traffic instances and results are discussed to assess the potential of the proposed algorithm.

Emergency doors may be bottlenecks in the evacuation of a building. As neither the capacities of these doors are not known, nor details on the behaviour of people near these doors, large scale laboratory experiments have been performed. Using video images, we have identified the relation between the capacity and a number of factors, such as door width, population composition and stress level. In addition, the collected trajectory data have been used to calibrate our pedestrian simulation model Nomad and to compare the resulting behavioural parameters between evacuation conditions and normal conditions. To do this, an automated calibration procedure has been developed, which yields parameter estimates for individual pedestrians. The existing calibration procedure has been extended by including data from multiple pedestrians into a single estimate. The resulting parameter distributions provide insight into pedestrian behaviour. So far, dedicated parameter sets have been estimated for elderly, adults and children. It could be shown that not only the pedestrian behaviour changed between normal and evacuation conditions, but also between the different types of persons.

Cathy Macharis is Professor at the Vrije Universiteit Brussel. She teaches courses in operations and logistics management, as well as in transport and sustainable mobility. Her research group MOSI-Transport and logistics focuses on establishing linkages between advanced operations research methodologies and impact assessment. She has been involved in several national and European research projects dealing with topics such as the location of intermodal terminals, assessment of policy measures in the field of logistics and sustainable mobility, electric and hybrid vehicles, etc. She is the chairwoman of Brussels Mobility Commission.

The multi-actor multi-criteria analysis (MAMCA) method to evaluate transport projects will be presented. This evaluation methodology specifically focuses on the inclusion of the different actors that are involved in the project, the so called stakeholders. As the traditional multi criteria analysis, it allows to include qualitative as well as quantitative criteria with their relative importance, but within the MAMCA they represent the goals and objectives of the multiple stakeholders and by doing so allow to include the stakeholders into the decision process. The theoretical foundation of the MAMCA method will be shown together with several applications in the field of transport appraisal.

<p>Stef Proost is full professor at the Catholic University of Leuven. At the KULeuven he teaches transport, environmental and energy economics at the Faculty of Economics and Business and at the Engineering Faculty. He is director of a group of 10 researchers at the Center for Economic Studies that deals with environment, energy and transport topics. He is co-founder of the Energy Institute of the KULeuven and co-founder of the spin-off Transport Mobility Leuven (TML)). </p><p> He is specialised in using mathematical models to address public policy questions: optimal pricing and investment in transport, choice of policy instruments for environmental policy, energy pricing questions. He is co-author of the models TRENEN, TREMOVE, MOLINO, MARKAL and GEM-E3 that are used widely in the EU. He coordinated and participated in several European research consortia (TRENEN-II, FUNDING, GEM-E3, PRIMES, MARKAL, CAPRI, AUTO-OIL 2, UNITE, MC-ICAM, REVENUE, etc. ). He has served as expert for EU Administrations for Transport, Environment, Energy and Economic and Financial affairs, for ECMT-OECD, UIC, for the Federal and Regional governments of Belgium and for several other national governments as well as for private firms in the energy and transport sector. </p><p> Policy issues he studied over the last years include the deepening of the Scheldt, the Iron-Rhine, the Oosterweel bridge in Antwerp, the selection of TEN-T projects, the introduction of road pricing and the climate policy in the transport sector. </p>

<p>This article examines long-term trends and broad policy options and challenges related to the road transport sector and its congestion and environmental impacts. A brief review of long-term projections of demand for road transport suggests that problems related to road network congestion and greenhouse gas emissions are likely to become more pressing in the future than they are now. Next we review, from a macroscopic perspective, three policy measures aimed at addressing these problems: stimulating shifts in transport modes to decrease congestion and greenhouse gas emissions, boosting low carbon technology adoption to reduce greenhouse gas emissions from cars, and regulating land use to reduce road transport volumes. We find that although these policies can produce tangible results, they may also have unintended and costly consequences. </p> <p>Keywords: transport, environment, modal choice, climate change, car technology</p>

Caspar Chorus is associate professor at Delft University of Technology (Section of Transport and Logistics). In past years, he was visiting scientist at Cornell University, assistant professor at Eindhoven University of Technology, visiting doctoral student at MIT and doctoral student at Delft University of Technology. His research is concerned with increasing the behavioral realism of travel demand models. The Random Regret Minimization-approach he developed has been succesfully applied in various research groups around the world, and is being incorporated in the newest version of the NLOGIT-software package. His 2007 dissertation ‘Traveler response to information’ has been awarded, among other prizes, IATBR’s Eric Pas Prize.

This talk presents an overview of recent progress related to the recently introduced discrete choice-paradigm of Random Regret Minimization (RRM). The RRM-approach to discrete choice-modeling provides an alternative to the conventional, Random Utility Maximization (RUM)-based approach which has dominated the field since its inception. In contrast with RUM-theory, RRM-theory postulates that when choosing, decision-makers are concerned with avoiding the situation where one or more non-chosen alternatives perform better than a chosen one in terms of one or more attributes. From this central behavioral premise, semi-compensatory decision-making and choice set-composition effects like the compromise effect emerge as RRM-model features. Being as parsimonious as RUM’s linear-additive multinomial logit model, RRM features logit-choice-probabilities and is easily estimable using conventional discrete choice-software packages. This paper ties together the main insights and results from a number of recent studies that have explored RRM’s model properties and empirically tested RRM-based models vis-à-vis competing model forms. As such, the talk provides an assessment of RRM’s potential and its limitations as a discrete choice model.

Dr. Ricardo A. Daziano is a specialist in theoretical and applied microeconometrics of consumer behavior, specifically on discrete choice models applied to sustainable behavior, technological innovation, and transportation demand. Daziano's specific empirical research interests include the analysis of pro-environmental preferences toward low-emission vehicles, modeling the adoption of sustainable behavior, estimating willingness-to-pay for renewable energy, and forecasting consumers' response to environmentally-friendly energy sources. Daziano completed his undergraduate studies and a masters' degree in engineering in Santiago at the University of Chile. After working four years as a demand analyst consultant in the private sector, he decided to pursue an academic career in discrete choice modeling. In 2010 he completed a PhD in economics at Université Laval in Québec City, with a specialty in econometrics and industrial organization. In January 2011 he joined the faculty of the School of Civil and Environmental Engineering at Cornell University, adding a new dimension to the area of sustainable systems engineering in both teaching and research. Additionally, Daziano continues to serve as a consultant in consumer choice modeling in areas such as transportation and sustainable tourism.

Range anxiety - consumers' concerns about limited driving range - is generally considered an important barrier to the adoption of electric vehicles. If consumers cannot overcome these fears it is unlikely that they will consider purchasing an electric car. Hence, for planning a successful introduction of low emission vehicles in the market it becomes essential to fully understand consumer valuation of driving range. Analyzing experimental data on vehicle purchase decisions in California, in this paper I derive and study the statistical behavior of three Bayes estimates that summarize consumer concerns toward limited driving range. Independence Metropolis-Hastings appears as a well-behaved sampler for nonlinear transformations of the marginal utilities. One of the empirical results is the posterior distribution of the driving range that makes an electric vehicle equivalent to internal combustion vehicles. Interestingly, this posterior is centered at driving range parity. The credible interval for the willingness to pay for an increase in range is also analyzed.

Flurin Hänseler is a MSc candidate in Process Engineering at ETH Zürich, from which he holds a Bachelor in Mechanical Engineering. He is currently conducting his Master thesis at MIT's Department of Chemical Engineering, for which he receives financial support from the Ernest Solvay Foundation. He has been a member of the Swiss Study Foundation since 2008; during his studies he concentrated on modeling and optimization of both technical and non-technical systems.

As a Process Engineering student, I am highly interested in methodologies for improving the operation of a technical process. Such techniques typically involve the design of a process model, detailed simulations as well as the development of control systems leading to optimum plant operation. Many times the skills and techniques used by process engineers overlap those applied and developed in operations research. In this seminar, I would like to talk about a chemical transport process I was involved during an internship and about a model on which I am currently working in the framework of my Master thesis. The goal of the former project was to optimize an industrial plant with respect to several criteria by employing a purely statistical analysis; the aim of my current work is to analyze a micro reaction using a physical (non-parametric) model and subsequently to optimize the reaction yield. I will keep my focus on the philosophy of process engineering which can be appreciated without an engineering background.

Dr. Madanat is Xenel Professor in the Department of Civil and Environmental Engineering, at the University of California, Berkeley. He received his Ph.D. from MIT in 1991. He is the Editor in Chief of the ASCE Journal of Infrastructure Systems. Dr. Madanat's research interests are in the area of civil infrastructure systems management, with emphasis on the utilization of analytical methods in modeling facility performance and optimizing life-cycle management decisions. His research has been published in Transportation Science, Transportation Research, Journal of Infrastructure Systems, Journal of Transportation Engineering, Computer-Aided Civil and Infrastructure Engineering and Transportation Research Record.

<p>This research addresses the determination of optimal maintenance and replacement policies for a heterogeneous system of facilities. The problem of optimizing maintenance and replacement policies at the system level is formulated in a reliability-based framework, based on policies that are optimal at the facility level. The facility-level policies are determined using a finite-state, finite horizon Markov decision process in which the state of the Markov chain contains information on the history of maintenance and deterioration. </p><p>Optimality conditions for the continuous-case system-level problem are explained intuitively. A numerical study shows that the results obtained in the discrete-case implementation of the solution are valid approximations of the continuous-case results. The computational efficiency of the system-level solution makes the formulation suitable for systems of realistic sizes.</p>

Shoshana Anily holds her Ph.D. in Management Science from the Columbia Business School, an MA degree is statistics and B.Sc. degree in mathematics, both from Tel Aviv University. Upon graduation she joined the University of British Columbia, and in 1989 she returned to her home country and joined Tel Aviv University. She is a professor of decisions and Operations Research at the Leon Recanati Graduate School of Business Administration of Tel Aviv University, Israel. Her research interests include performance analysis of heuristics, routing problems, inventory control, production, scheduling, and cost allocation problems in a supply chains. She spent some periods as a visiting professor at Columbia University, EPFL and the CRT at the University of Montreal.

<p>We consider a <em>cooperative game with transferable utility</em>, defined on a set of players N, whose characteristic function V depends only on a vector of some quantitative properties of the players, and is otherwise independent of their identity. As a result, the game can be defined on any virtual coalition of players, not necessarily subsets of N, as long as each player is associated with a vector of properties. We call such games regular games. Many games considered in the literature are regular. We assume that the characteristic function of the game is sub-additive, i.e., the cost of two disjoint coalitions is at least as large as the cost of their union, implying economies of scope. Sub-additivity implies that a natural outcome of any bargaining process is the formation of a single coalition N called the grand coalition. </p> <p> A central question in cooperative games is how to fairly allocate the cost V(N) among the players of N. The literature refers to a few concepts of fairness. Here we focus on the known concept of the core: core allocations guarantee that no coalition has an incentive to quit the grand coalition. A fundamental problem is that the core may be empty. The literature describes just a couple of methods for proving the non-emptiness of the core. In this talk we present what we believe is a novel condition that guarantees the non-emptiness of the core. We refer to a property called homogeneity of degree one, which means that when k coalitions, associated with exactly the same set of vectors of properties, cooperate, the resulting cost is k times the cost of a single coalition. Homogeneity of degree one implies lack of economies of scale. We prove that if the characteristic function of a regular game is sub-additive and homogeneous of degree one, then the core of the game is non-empty. </p> <p> We then present a few examples for such games which naturally emerge when servers in queuing systems cooperate in order to outsource some of the activities, while the rest of the activities are optimally assigned to the servers in the system. The existing conditions for showing the non-emptiness of the core fail here, while the conditions stated above are satisfied, proving that the cores of these games are non-empty.</p>

<p>As more airports in the United States and in Europe become more congested, it also becomes increasingly likely that delays at one or more airports will spread to other parts of the network. We describe an analytical queuing and network decomposition model developed to study this complex phenomenon. The Airport Network Delays (AND) model computes the delays due to local congestion at individual airports and, more important, captures the "ripple effect" that leads to the propagation of these delays. The model operates by iterating between its two main components: a queuing engine (QE) that computes delays at individual airports and a delay propagation algorithm (DPA) that updates flight schedules and demand rates at all the airports in the model in response to the local delays computed by the QE. The QE is a stochastic and dynamic queuing model that treats each airport in the network as a M(t)/Ek(t)/1 queuing system. AND also includes an algorithm to model Ground Delay Programs.</p><p> The AND model is very fast computationally, thus making possible the exploration at a macroscopic level of the impacts of a large number of scenarios and policy alternatives on system-wide delays. AND has been implemented for two networks, one consisting of the 34 busiest airports in the continental United States and the other of the 19 busiest in Europe. It provides insights into the complex interactions through which delays propagate through the network and the often-counterintuitive consequences of these interactions. We are currently using AND to investigate a number of policy issues, such as the effect of implementing slot constrains at New York airports.</p>

When a region is threatened by a disaster, like a flood or a fire, people have to be evacuated to avoid as many casualties as possible. A plan (for example containing instructions for the people when to leave and where to go) has to be created to be able to evacuate these people in an efficient way. My research focuses on optimization methods to create these plans under uncertainty. In this seminar, I will introduce some results obtained so far, showing among others the doubled effectiveness of evacuation instructions if they are optimized instead of created by practical rules. Furthermore, I will discuss my objective of my stay at EPFL, which is to study the appropriateness of robust optimization methods to solve the evacuation problem under uncertainty and to apply one of these methods to the problem.

Elisabetta Cherchi has recently joined DTU Transport as Associated Professor. Previously she has been Lecturer in Econometrics Methods for Transportation at the Department of Territorial Engineering, Universita de Cagliari (Italy), Associate Lecturer at the Transport Research Centre, Universidad Politecnica de Madrid (Spain) and Researcher at CTS, Imperial College (UK) where she is still Honorary Researcher. She has collaborated for almost 15 years with Pontificia Universidad Catolica de Chile, and more recently with Maryland University. Her main research interest is in the demand modeling of consumer behavior, with particular reference to discrete choice analysis, microeconomic derivation of behavioral models and project evaluation. She is member of the Editorial Board of Transportation Research part B and Transportation, and Board Member of the International Association of Behavioral Research (IATBR). Elisabetta Cherchi has been also involved as consultant in more that 25 transport projects at local, national and international level.

A specific problem inherent in urban planning in areas of environmental interest is that problems can occur when ideas are effectively turned into action, and it is possible that eventually the implementation is not consistent with the plan. If the development of the project produces undesired effects, there might be an irreversible loss of environmental values. To evaluate the quasi-option values associated with environmental renewal projects in urban areas and to elicit inter-temporal preference over projects with different time spans. We adopt an attribute-based stated choice approach to elicit preferences about three planning alternatives for environmental renewal of the corridor along the beach front, including car access restrictions. Since such projects involve some uncertainty and irreversibility, the quasi-option values associated with project development is estimated. Results show a great sensitivity towards environmental renewal, and this varies significantly with the characteristics of each individual. However, individuals show a strong preference for a mild environmental renovation rather than an extreme ecological scenario. Results also show that a more prudent development strategy is valued at about four times greater than a procedure that provides a greater chance of an undesired outcome.

<p>Ennio Cascetta, born in Naples in 1953, is full professor Transportation Systems Planning at the University of Naples Federico II. He is the author of several books, including ''Transportation system engineering: theory and methods'' (Kluwer, 2001) and ''Transportation Systems Analysis: Models and Applications'' (Springer, 2009) and of more than one hundred papers published in International journals and proceedings. Main research areas include analysis, modelling and estimation of transport demand, static and dynamic models for traffic assignment, railway network planning and pricing, supply design models for transit and traffic light systems, traffic theory and the quality of circulation on motorways, analysis and control of traffic congestion in urban areas. He has taught in several international courses. Since 1995 teaches the yearly course on Modeling and Simulation of Transportation Networks at the Massachusetts Institute of Technology </p> <p> He has been director of the Second Special Project on Transportation Research of the Italian National Research Council (seven years duration, 150 million EURO value) from 1997 to 1999. He is editor of a series of books on Transportation published by Franco Angeli in Italian. He acted in the scientific committees of a number of international journals such as Transportation Research ,Transportation Science and Transportation Policy. He acted as scientific committee member of several international conferences such as World Conference on Transportation Research (Vice-Chairman) , Conference of the International Association of Travel Behaviour Research, Tristan. </p> <p> He is an appointed member of the Transport Research Board of National Research Council (USA). He has been the scientific responsible of research units in the context of several UE research programmes in DGXIII and DGVII such as DYNA, AIUTO and TRACE. Took part as on expert to the international committee for the design of the EURET 2 Project and 4th Framework program appointed by DGVII. </p> <p> Concerning his professional and political role, from 2000 till 2010, he was Ministry of Transport for the Campania Region. In the years 1999 and 2000, he was coordinator of the experts board of the Italian Ministry of Transport for the National General Transportation Plan. In the 1998, he was the coordinator of the experts board of the Italian Ministry for Public Works for the project about the main road infrastructures. From 1995 till 1997, he was the scientific coordinator of the Transportation Master Plan of Naples. In the 1997, he took part, in quality of expert of Italian Ministry of Transport, to the joint Transport - Environment commission for the Italian high-speed rail lines. From 1994 to 1995, he was a member of the board of the main Naples Public Transport Company (A.T.A.N.). From 1993 to 1999, he was president of the commission of the Italian Ministry of Transport for the project: ''The Italian Decision Support System for transportation policies and investments''. He acted as a professional consultant on various aspects of transport planning and management at urban and inter-city level and on technical-economic feasibility studies of transport infrastructures (i.e. Transport Plans of Genoa, Naples, Palermo, Verona) </p>

<p>Transportation systems engineering can be defined as a discipline aimed at the functional design of physical and/or organizational components of transportation systems. A set of coordinated, internally consistent actions on a transportation system are referred to as a project or plan. Transportation system modelling allows the prediction of the relevant impacts of possible projects in order to design them , assess their technical suitability and to support intermediate and final decision-makers through the process of projects evaluation. </p> <p> The range of transportation system projects is very diverse, and so are the points of view from which their consequences can be evaluated. Projects might involve transportation infrastructures, control systems, services, fares. Similarly, projects can be designed and evaluated from the perspective of the community they will serve, from the perspective of the service and/or facility operators, or from the perspective of potential investors (i.e. project financing). Decision-making for transportation systems is often more complex than for systems in other sectors of engineering. This is especially true when the decision-maker must consider, either directly or indirectly, the effects of proposed actions on the larger community .These effects include variations in travellers generalised costs, investment and operating costs for different companies; effects on land-use and activity location; effects on economic levels, variation on air pollution; number of accidents, energy consumption, etc. </p> <p> Over the last 50 years a wide body of models and computational tools have been developed and applied to solve countless problems ranging from large investments plans at national and international levels to traffic control schemes of urban intersections. </p> <p> Two applications of these principles are discussed in some detail. The first is related to the Campania Regional Metro System project, which is an example of integrated land-use, infrastructure and operational planning. The whole project is based on the idea of integrating the existing railway lines into a single physical network by building some new links, new stations and new modal interchange facilities. The RMS project integrates also operational components such as service lines, timetables and integrated pricing. In addition to the transit ''supply-side'' elements, the project includes relevant''demand-side'' aspects, including town planning based on the rail network, urban renewal around rail stations and upgraded and new stations with high architectural standards and a new symbolic value. The overall project is worth approximately 9 billion Euros and is currently in an advanced realization phase . Approximately 50 kms of new tracks and 40 new stations were opened to operation in the last ten years (3 billions worth) and investments for 3 more billions are currently underway. </p> <p> The second application of the principles of transportation systems engineering deals with the recent High Speed Rail investments in Italy. The recent opening of new sections and the oncoming entrance in the High Speed Rail market of a new private operator, competing with the national railways operator, create the conditions for a unique case study to investigate the behavior of long- distance passengers. The framework developed to forecast the national passenger demand for different Italian macroeconomic, transport supply, and High Speed Rail marketing scenarios is presented and the main results are discussed.</p>

Bilal Farooq is a PhD Candidate under the supervision of Professor Eric J. Miller at Urban Transportation Research and Advancement Centre (UTRAC), University of Toronto. He has a Master degree in Computer Science and Bachelor in Engineering. Before starting his PhD, he worked as a software engineer in the software industry. Bilal’s research expertise involves econometric modelling and microsimulation of decision making and interactions in the context of urban systems. He is also interested in demographic evolution, urban energy consumption, and GHG emission/dispersion modelling. Currently, he is working as the software architect of Integrated Land Use Transportation and Environment (ILUTE) modelling framework that is under development at UTRAC.

Spatial and temporal distribution of built-space supply plays an important role in shaping up the urban form and thus the general travel pattern in an urban area. Within integrated framework, we are interested in modelling the decisions of a builder in terms of when, where, what type, and how much of a built-space to build. This talk presents a discrete-continuous model formulation for the built-space supply decisions that are based on expected profit maximization. The framework is applied to estimate a model for supply of new office space in the Greater Toronto Area (GTA) for the duration of 1986 to 2006. The results indicates a risk taker behaviour on the builders’ part, while market conditions and supply of resources (labour, construction cost etc.) are also found to be important factors in decision making.

Erik Jenelius received the M.Sc. in Engineering Physics at The Royal Institute of Technology (KTH), Stockholm, in 2004. From June 2005 to January 2006 he worked for the City of Stockholm with the evaluation of the Stockholm congestion pricing trials. Since February 2006 he is a Ph.D. Candidate at the Division of Transport and Location Analysis, KTH. He is working in a project concerning road network vulnerability analysis, the main aim of which is to develop the methodology for vulnerability analyses with application to real large-scale road networks. In 2008 he received the Lic.Eng. degree in Infrastructure from KTH. His research interests include road network vulnerability, critical infrastructure protection, transport reliability, transport economics and complex networks.

As the recent volcanic blockade of European air travel shows, disruptions in the transport systems can have severe impacts for affected individuals, businesses and the society as a whole. In the research presented here, vulnerability is seen as a combination of the severity and the frequency of unplanned system disruptions, with a focus on large, rare events. I will describe model-based studies of different aspects of vulnerability, in particular the dichotomy of system efficiency and user equity, applied to the Swedish road network. The scenarios studied include both closures of single road links and disruptions of extended areas. I will also address the issue of how to value the delays that are incurred by network disruptions and, using an activity-based modelling approach, I illustrate that these delay costs may be considerably higher than the ordinary value of time, in particular the first few days after the event when travel conditions are uncertain.

Shlomo Bekhor is Associate Professor in the Transportation and Geo-Information Department, Faculty of Civil and Environmental Engineering at the Technion. He has a B.Sc. in Aeronautical Engineering at the Aeronautical Institute of Engineering, Sao Jose dos Campos, Brazil, and M.Sc. and Ph.D in Transportation Engineering at the Technion. He teaches and conducts research in transportation planning, discrete choice and network equilibrium models. His main areas of interest include route choice modeling, large-scale equilibrium models, and innovative transport systems. He has extensive experience in transportation master plans, economic evaluations, and traffic and transit assignments. Prof. Bekhor is currently spending 6 months in the IVT at ETH Zurich, hosted by Prof. Axhausen.

<p>Long-distance trips are generally under-reported in typical household surveys, because of relative low frequency of these trips. This paper proposes to utilize location data from cellular phone systems in order to study long-distance travel patterns. The proposed approach allows passive data collection on many travelers over a long period of time at low costs. The paper presents the results of a study that applies cellular phone technology to assess trips at the national level. </p><p> The method was specifically designed to capture long distance trips, as part of the development of a national demand model conducted for the Economics and Planning Department of the Israel Ministry of Transport. The method allows the construction of Origin-Destination tables directly from the cellular phone positions. The paper presents selected results to illustrate the potential of the method for transportation planning and analysis. </p>

<p>Hybrid choice models integrate many types of discrete choice modeling methods, including latent classes and latent variables, in order to capture concepts such as perceptions, attitudes, preferences, and motivation (Ben-Akiva et al., 2002). Although they provide an excellent framework to capture complex behavior patterns, their use in applications remains confidential in the literature due to the difficulty of estimating the models. </p><p> In this talk, we provide a short introduction to hybrid choice models, and explain how the new version of biogeme can be used to estimate their parameters. </p><p> Ben-Akiva, M., McFadden, D., Train, K., Walker, J., Bhat, C., Bierlaire, M., Bolduc, D., Boersch-Supan, A., Brownstone, D., Bunch, D., Daly, A., de Palma, A., Gopinath, D., Karlstrom, A., and Munizaga, M. A. (2002). Hybrid choice models: Progress and challenges, Marketing Letters 13(3):163-175</p>

Dr Philippe Canalda got M.Sc. and Ph.D. Degrees in computer science from the University of Orléans (France) in 1991 and 1997, respectively. He worked at INRIA Rocquencourt from 1991 to 1996 on the automatic generation of optimizing and parallel n-to-n cross-compilers. From 1996 to 1998, he worked as Research Engineer in the Associated Compiler Expert start-up factory at Amsterdam, The Netherlands. Then he worked 2 years at LORIA on the synchronization of cooperative process fragment, based on workflow model, and applied to ephemeral enterprise. Since 2001, he is an Associate Professor at the Computer Science Laboratory (LIFC, EA 4269) at the University of Franche-Comté in France. His research topics deal with, on the one hand mobility services and wireless positioning, and on the other hand on robust and flexible optimizing algorithms based on graph, automata and rewriting theories. His latest works are conduced in a trans-disciplinary way: geographic aspects, sociologic one’s, operational research and systemic approach help and propose complex system to various contexts among those of end-users, transportation operators, and authoritative organizers.

System key words - Trans-disciplinary studies and experimentations of dynamic and complex transportation systems: on demand transportation, car-pooling, concealing combinatory algorithm with ground realities, job realities, new hot topics (infomobility, multimodality, instantaneity of offers, immediately of operation, ICT). Combinatorial key words - Dynamic N-TSP/DARP split n PU-m D with TW, variable capacities of vehicle fleet and other criteria among dynamic depots, operational politics, QoS oriented incitative politics, minimizing call-backs, ecolometry, ... Others dimensions, being taken in account should be discussed as service guaranty, certification, implicit and contextual information. These different aspects will be describe through the 10 late years biography, animation of networks dedicated to transportation system, especially flexible transportation system and dynamic carpooling. I will end by presenting key points of interest for possible interreg project : sure and dynamic home to work cross-boarder displacements, integrating dynamic carpooling within a public and global transportation offer.

<p>The likelihood functions of many discrete ordered and unordered-response choice models entail the evaluation of analytically-intractable integrals. For instance, the use of a mixing mechanism to relax the independent and identically distributed (IID) error term distribution in the multinomial logit model is well documented in the discrete choice literature on unordered multinomial response models. In such an approach, the error term vector is effectively decomposed into an IID component vector and another vector of jointly distributed random coefficients that lends the non-IID structure. It is typical (though not always the case) to consider the joint distribution of the random coefficients to be normally distributed. A particular advantage of the mixing approach is that it can be used for both cross-sectional choice data as well as panel data without any substantial conceptual and coding difference. However, such mixed models also lead to intractable likelihood function expressions. Except in the case when the integration involves only 1-2 dimensions, maximum simulated likelihood (MSL) techniques are usually employed to estimate these models. Unfortunately, for many practical situations, the computational cost to ensure good asymptotic MSL estimator properties can be prohibitive and literally infeasible as the number of dimensions of integration rises. Besides, the accuracy of simulation techniques is known to degrade rapidly at medium-to-high dimensions, and the simulation noise increases substantially. This leads to convergence problems during estimation. In addition, such simulation-based approaches become impractical in terms of computation time, or even infeasible, as the number of mixing dimensions grows. </p><p> In this paper, we introduce a maximum composite marginal likelihood (CML) estimation approach for multinomial ordered-response and unordered-response models. The CML approach can be applied using simple optimization software for likelihood estimation. It also represents a conceptually and pedagogically simpler simulation-free procedure relative to simulation techniques, and has the advantage of reproducibility of the results. This new CML estimation approach for mixed cross-sectional/panel ordered and unordered multinomial models does not have the convergence and stability problems that can plague simulation-based techniques, while at the same time leading to substantial computational efficiency. Simulated datasets are used to empirically compare the performance of the maximum simulated likelihood (MSL) approach and the maximum CML approach introduced here, in terms of ability to recover model parameters, estimator efficiency, and computation time. The presentation will also discuss the CML estimation for a wide variety of cross-sectional and panel discrete choice model structures, some of which, to our knowledge, have not appeared in the econometric literature. </p>

António Pais Antunes holds a Diploma in Civil Engineering from the University of Coimbra, Portugal, and a PhD in Applied Sciences from the Catholic University of Louvain, Belgium. He currently is a professor at the Department of Civil Engineering of the University of Coimbra, where he teaches courses in urban and regional planning and transport engineering, and is the director of the Doctoral Program in Transport Systems offered jointly by the Technical University of Lisbon and the Universities of Coimbra and Porto and in collaboration with the MIT. Before joining the University of Coimbra, he was the director of planning of the Centro Region Development Agency. His main subjects of research are public facility planning and transport infrastructure planning. He is the author of over 150 scientific publications, including 35 articles in international journals (e.g., Environment and Planning A and B, European Journal of Operational Research, Journal of Transportation Engineering, Journal of Water Resources Planning and Management, and Transportation Research A).

: In the seminar, a multi-objective approach to long-term inter-urban multilevel road network planning is presented. In addition to the efficiency objectives dealt with in most of the literature where the subject is addressed, the approach takes into account robustness, equity, and environmental objectives. For achieving the objectives, two types of action can be performed: the construction of new road segments and the upgrading of existing road segments. The approach is consistent with the planning framework of the Highway Capacity Manual, using the concept of level of service to assess traffic flow conditions. The application of the approach is illustrated for case studies involving the road networks of Poland and Portugal.

María Francisca Yáñez received her B.Sc. degree in Civil Engineering from Pontificia Universidad Católica de Chile (PUC). She is currently Ph.D. candidate at PUC, and is enjoying a one-year stay at the German Institute for Economic Research (DIW-Berlin) financed by the Chilean Council for Scientific and Technological Research and the German Academic Exchange Service (DAAD). She has presented part of her thesis work at the most important conferences in America and Europe.

The mode choice process, especially in the case of commuter trips, reflects the strong tendency people have to simplify the assessment of their options when confronted with successive well-known decisions. Thus, it is common to repeat the “habitual” choice over time involving a potentially important inertia element. However, while inertia effects increase the probability of maintaining the same choice in a stable situation, in a changing environment i.e. one that is disrupted by a radical or significant policy intervention, user behaviour may be affected by a specific response to abrupt changes. Shock effects of this kind could increase the probability of individuals leaving their habitual choices. Temporal effects have been commonly ignored in practical studies, as most demand models to date have been based on cross-sectional data. A few recent studies dealing with panel data have managed to incorporate inertia effects, but there are no studies that have included both inertia and shock effects. To address this, I started by building a data panel around the introduction of a new and radical policy for the conurbation of Santiago de Chile. The final aim was to develop mode choice models incorporating the effects of three main forces involved in the choice process: (1) the relative values of the modal attributes, (2) the inertia effect, and (3) the shock resulting from and abrupt policy intervention. This research includes the formulation of an inertia-shock model and its application to each of simulated and real data. The results obtained provide empirical evidence that inertia-and-shock models are superior to the traditional approach (i.e. ignoring the consequences of not including temporal effects) both in terms of explaining a real phenomenon (estimation), and in terms of predictive capability. So, these findings reinforce the belief that real systems should be modelled with data which allows capturing the effect of new policies and habit/inertia effects in individuals’ choice processes. Apart from the research backbone summarised above, a related investigation has been conducted. The last wave of the Santiago Panel included a latent variable section that allowed estimating hybrid choice models in a panel data context, and to test empirically the performance of the preferred estimation methods for this type of models.

Amedeo R. Odoni is Professor of Aeronautics and Astronautics and of Civil and Environmental Engineering at MIT and one of the Co-Directors of the Airline Industry Program. He has also served as Co-Director of the FAA's National Center of Excellence in Aviation Education, Co-Director of MIT's Operations Research Center, Editor-in- Chief of Transportation Science, and consultant to numerous international airport and aviation-related organizations and projects. The author, co-author or co-editor of eight books and more than 100 other technical publications, he is a Fellow of the Institute for Operations Research and Management Science (INFORMS) and has received several distinctions, among them the INFORMS Lifetime Achievement Award for Contributions to Transportation Science, the T. Wilson Endowed Chair at MIT, the U.S. Federal Aviation Administration's (FAA) National Award for Excellence in Aviation Education, a Honorary Ph.D. from the Athens University of Economics and Business, and four MIT awards for excellence in teaching, mentoring and advising. His students have also received many prizes for their research and dissertations.

<p>As more airports in the United States and in Europe become congested, it also becomes increasingly likely that delays at one or more airports will spread to other parts of the network. We describe an analytical model, Airport Network Delays (AND), developed to study this complex phenomenon. It computes delays due to local congestion at individual airports and, more important, captures the "ripple effect" that leads to the propagation of local delays throughout the network. The model operates by iterating between a stochastic and dynamic queuing engine (QE) that computes delays at individual airports and a delay propagation algorithm that updates flight schedules at all the airports in the model in response to the local delays computed by the QE. </p><p> The AND model is fast computationally, making possible the exploration of the impacts of a large number of scenarios and policies on system-wide delays. It has been fully implemented for the network of the 34 busiest airports in the continental United States. An implementation for the network of the 34 busiest airports in Europe is in progress. The model provides insights into the complex interactions through which delays propagate through the network and the often-counterintuitive consequences of these interactions. (joint work with PhD student Nikolas Pyrgiotis)</p>

Prem is a Doctorate in Business Management with specialization in Operations Research from a premier Busi-ness school in India - The Indian Institute of Management. During his doctoral research, he was awarded the prestigious John Curtin International Institute Fellowship and visited Curtin University of Technology (Austra-lia) as a research scholar, received a scholarship from the Danish Government to undergo a training program on Geographical Information Systems (GIS) applications. Following Doctorate, Prem worked at a University level research center with NUS Singapore and manages the off-shoring and outsourcing of research and analytical projects. Prem has also undertaken special lecture sessions on industrial applications of Operations Research at NUS (Singapore), Delhi University (India), Curtin Univer-sity of Technology (Australia), and several other business schools around India. Currently, he is teaching one course on “Business Analytics and Optimization” at the Indian Institute of Management Bangalore. His research interests include (but not limited to): 1. Airline schedule planning and application of search strategies to solve large scale optimization problems 2. Using combinatorial optimization and graph theoretic techniques to solve route net design for shipping liners 3. Supply chain planning and design problems for multiple items.

In this paper, we consider the gate assignment for a large airline at its hub airport. The problem is considered in both modes – planning as well as operations. The first objective in planning mode assigns airport gates dynamically to scheduled flights based on daily origin and destination passenger flow data ensuring that maximum connection revenue is realized. The second objective aims to schedule flights to gates in as few manpower zones as possible. This would help reduce the operating costs for the airline. The third objective ensures that gate rest, i.e. the gate idle time between an outgoing and the next incoming flight, is kept to the maximum possible so that extent of regating is minimized in the eventuality of successive flight delays. One of the major contributions of this paper is gate assignment in the operations mode to ac-count for real-time delays and disruptions to the planned schedule. This would entail making deviations to the existing assignment plan and retiming flights marginally in extreme situations when the existing gate infrastructure cannot handle all the flights. It is important that our solution technique is fast enough to produce this operational solution within a few seconds. We formulate these problems as mixed 0-1 integer program with a linear objective function and constraints. Due to the complexity in the problem size and formulation, we have resorted to relaxation for certain instances when a reasonable solution is not obtained within the time limit. Problem formulation for flight retiming in the operations mode has been borrowed from the concepts of resource constrained project scheduling. Implementation is done using OPL and computational results for actual data sets are presented.

Lavanya Marla is currently a PhD candidate at the Massachusetts Institute of Technology. She holds Masters degrees in Transportation and Operations Research from MIT, and a Bachelors degree in Civil Engineering from the Indian Institute of Technology, Madras.

In this work, we consider large-scale resource allocation problems under uncertainty. Inherent real-world uncertainty guarantees that deterministic optimal solutions are rarely, if ever, executed. By proactively modeling uncertainty, robust methods attempt to generate solutions that are less vulnerable to uncertainty. We will examine the application of robust methods such as the Bertsimas and Sim approach and Chance-Constrained Programming, which represent two different paradigms: worst-case and probabilistic modeling. We discuss differences in models and solutions generated by these approaches, and propose extensions to these models that overcome some crucial limitations.

Design of Experiments (DOE) is widely used within Nestlé R&D in order to properly plan the data collection in numerous projects. This technique leads to significant time savings and cost savings in all steps of product development (renovation of existing products and innovation with new products). We first start by introducing the general framework of product development and by motivating the use of DOE in such context. Then, we briefly remind the basic concepts related to DOE… and the mistakes to avoid when collecting data! After a short literature review on existing techniques for mixture designs and constrained mixture designs, we introduce a novel approach called "Adjusted Mixture Designs". The way to build statistical models based on these designs is also discussed. Finally, the advantages and the flexibility of this approach are illustrated with concrete food examples."

The problem of allocating resources in proportion to some measure has been studied in various fields of science for a long time. The apportionment problem of allocating seats in a parliament in proportion to the number of votes obtained by political parties is one example. This presentation will show a number of other real-life problems, for instance Liu-Layland problem, stride scheduling, fair queuing which can be formulated and solved as the problems of proportional optimization and fairness.

I will describe our recently developed theory of n-fold integer programming, which enables polynomial time solution of fundamental linear and nonlinear integer programming problems in variable dimension. I will demonstrate the power of this theory by deriving a polynomial time algorithm for multicommodity flow problems where the cost of flow over a channel is a more realistic, nonlinear, function of the flow, accounting for channel congestion under heavy traffic or communication load.

Christina Fragouli is an Assistant Professor in the School of Computer and Communication Sciences, EPFL, Switzerland. She received the B.S. degree in Electrical Engineering from the National Technical University of Athens, Athens, Greece, in 1996, and the M.Sc. and Ph.D. degrees in electrical engineering from the University of California at Los Angeles in 1998 and 2000, respectively. She has worked at the Information Sciences Center, AT&T Labs, Florham Park New Jersey, and the National University of Athens. She also visited Bell Laboratories, Murray Hill, NJ, and DIMACS, Rutgers University. From 2006 to 2007 she was an FNS Assistant Professor in the School of Computer and Communication Sciences, EPFL, Switzerland. She served as an editor for IEEE Communications Letters, and is currently serving as an editor for IEEE Transactions on Communications. She received the Fulbright Fellowship for her graduate studies, the Outstanding Ph.D. Student Award 2000-2001, UCLA, Electrical Engineering Department, and the Zonta award 2008, in Switzerland.

Network coding is an emerging area that re-examines fundamental principles of network information flow. The main idea is that we allow intermediate nodes in a network to not only forward but also to process the incoming information flows. This simple idea promises to have a significant impact in diverse areas that include multicasting, network monitoring, reliable delivery, resource sharing, efficient flow control and security. This talk will present an overview/tutorial on benefits network coding can offer over dynamically changing networks, as well as open research directions. A specific application we will discuss in more detail is broadcasting in wireless networks.

Francisco Martinez is Associate Professor at the University of Chile, Senior researcher in the Millennium Institute of Complex Systems in Engineering, and Advisor to the Rector, University of Chile. He holds the following degrees: PhD with Thesis in the field of Urban Economics (1991), University of Leeds, UK; Master of Arts in Transport Economics (1988), University of Leeds, UK; and Civil Engineer, Universidad de Chile (1984). His research focuses on urban systems (Urban and Transport Modeling, Land use, land prices, location externalities, evaluation, regulations and tax/subsidies, transport and land use interaction), Spatial Choice modeling (constrained behavior, bid-auctions, calibration methods), location models (location demand, optimization and equilibrium location of services (schools and other services)) and Sustainable Urban Systems (member of the Risk Habitat Megacities Project, in partnership with the Helmhold Society, Germany).

The seminar will cover how the urban economic bid-auction theory and the random utility theory are combined to generate the stochastic bid-auction theory that sustains MUSSA (currently CUBE LAND model). The explanations will cover the agents (consumers and suppliers) rational behavior, the auction market equilibrium, how location externalities (e.g. agglomeration economies) and suppliers face economies of scale are modeled, how regulations affect the market, and how the model predicts emerging patterns (agglomeration and segregation, densities), the formation and distribution of rent across space and utility levels attained by consumers.

Francesco Viti was born in Matera (Italy) in 1975. He graduated "Summa Cum Laude" in Civil Engineering, majoring in Transportation Engineering, at the University "Federico II" of Napoli in 1994. In his final M.Sc. thesis he developed a framework for the optimization of urban accessibility using a Park Pricing strategy. In 2001 he joined the Dynamic Traffic Management group of the Delft University of Technology, starting a Ph.D. research which ended in 2006 with the writing of a dissertation entitled "The Dynamics and the Uncertainty of Delays at Signals". During this period he produced a number of high impact articles and presented his work at many international conferences. He was awarded in 2004 with the prize "Best Paper for a Young Researcher" the 10th WCTR Conference in Istanbul. He is currently employed at both Delft University of Technology and at the Katholieke Universiteit Leuven since January 2007, participating actively on research projects on different areas, e.g., ADAS systems, ICT technologies, advanced data collection systems, incident analysis, emission modeling, travel time prediction, real-time dynamic traffic management, traffic flow theory and simulation. During this period he was involved in various projects and teaching activities. He was course leader of "Game Theory: theory and applications in transportation" given at the TRAIL Research School and lecturer of the TU Delft course "Data Collection Methods and Analysis". Moreover he gave his contribution to doctoral and project researches on travel information systems, pricing, travel time prediction, Advanced Driver Assistance Systems etc. He is involved in the evaluation of European Commission projects for both FWP6-INFSO e FWP7-ICT, and reviewer for high impact journals.

Growing concerns over degrading air quality and climate change have led to the implementation of measures to reduce road traffic emissions, among them Intelligent Transport Systems (ITS). Such measures influence traffic flows in many, often subtle ways. The assessment of these measures is not straightforward as the effects are often found on the microscopic level, i.e. the movement of individual vehicles, and are difficult to observe directly. However, as ITS measures become more common, and environmental aspects more important, there is a need for reliable tools to evaluate the effects of ITS measures on emissions. Several traffic and emission models are available; the question is whether they can model accurately enough both the traffic and the produced emissions. Microscopic simulation programs are often used in practice for assessing the effects of ITS strategies, since they enable one to calculate the trajectory and the characteristics of each vehicle and of the network under different scenarios and road conditions, and to model vehicles' and drivers' behavior with enhanced detail and at fractions of a second. However, in these models, measures like speed, acceleration and deceleration are still determined by a limited set of parameters, which are rarely chosen based on field measurements. One main reason is that collecting real vehicle trajectories, able to quantify and explain these parameters, is a very challenging task. At signalized intersections more difficulties arise, since traffic flows are interrupted by the traffic control signal, and driving behavior strongly depends on the current state of the system, i.e. whether drivers have right of way, or they have to stop in queue etc. We propose a method to collect such dataset using image processing techniques, which allows one to obtain a full vision on the traffic process at the signal, and to measure individual vehicle speeds and accelerations at a microscopic level. By using microscopic trajectory data we can gain insight into the behavior of drivers, i.e. the way they approach the signal and interact with the traffic control and the other vehicles. Therefore, analysis of real driving behavior from a microscopic scale enables us to evaluate the speed and acceleration behavior at these systems under different traffic conditions, and to quantify the differences observed in the (longitudinal) driving behavior within these systems, e.g., in terms of interdriver heterogeneity as well as intradriving variability and with different traffic and control states. Using this unique dataset, direct calibration of the main parameters in two widely used commercial microscopic software programs, VISSIM and AIMSUN, could be performed, and validation and comparison could be operated in terms of individual speeds, speed variations and for the estimation of emissions concentrations. As a result of this analysis, we highlight a number of inconsistencies between simulation and reality, due to simplifications in driver's free driving as well as in car-following mode. Lacks in intradriving variability and anticipatory behavior in these programs suggest that they may not provide realistic results for these types of road sections, even when carefully calibrated.

Yingyan Lou is a Ph.D. candidate in the Department of Civil and Coastal Engineering at the University of Florida. She received two bachelor degrees in both Engineering Science and Economics from Peking (Beijing) University, China, in 2005; and earned her master degree in Civil Engineering from the University of Florida in 2007. Yingyan Lou is currently a research assistant at the Transportation Research Center, University of Florida. Her primary research interest is transportation systems modeling and optimization, with applications on system-wide congestion pricing, traffic-responsive tolling for managed-lane operations, dynamic origin-destination demand estimation, robust transportation network design, freeway incident response planning, and infrastructure asset management. She also has a keen interest in traffic flow theory and operations. During her Ph.D. study, Yingyan Lou has co-authored seven papers and made eight presentations at various conferences. She won the best poster award at the 2007 Institute of Transportation Engineers Florida District Annual Meeting and was awarded the graduate scholarship of Women's Transportation Seminar in 2008.

This research investigates congestion pricing strategies in static networks with boundedly rational route-choice behaviors. Under such behavior, users do not necessarily choose a shortest or cheapest route, when doing so does not reduce their travel time by a significant amount. A general path-based definition and a more restrictive link-based representation of boundedly rational user equilibrium (BRUE) are presented. The set of BRUE flow distributions are non-convex and always non-empty. The problems of finding best- and worst-case BRUE flow distributions are formulated and solved as mathematical programs with complementarity constraints. Because alternative tolled BRUE flow distributions exist, our congestion pricing models seek a toll vector or scheme that minimizes the system travel time of the worse-case tolled BRUE flow distribution. As formulated, the models are generalized semi-infinite min-max problems, and we propose a heuristic algorithm based on penalization and a cutting-plane scheme to solve them. Numerical experiments demonstrate that system performance may vary substantially within the BRUE set, and traditional marginal-cost pricing may not evolve the flow distribution to system optimum. The proposed robust pricing models are able to guard against the worst-case scenario effectively and lead to a more stable system performance. This research offers a new alternative for more realistically modeling users' route choices at general networks. It has relaxed the dominance of the assumption of perfect rationality in transportation planning, and led to more realistic models and accurate tools to help and guide in the analysis and evaluation of congestion pricing strategies.

Dr Connors gained his first degree in Mathematics (Oxford) and his PhD in Quantum Chaology (Bristol). This was followed by 3 years working as a MATLAB developer (Cambridge). In 2003 Dr Connors returned to academic research, joining the Institute for Transport Studies (Leeds). His current research concerns the representation of network infrastructure and human behaviour within transport network models across different scales of analysis, and the consistency of these mathematical formulations.

Transport models comprise a representation of the network infrastructure, and of transport users who decide how, where and when to travel. Naturally, the models used to optimise micro-scale/short-term measures (e.g. timing of traffic lights), are different from those used when planning major infrastructure (e.g. increasing motorway capacity), or designing long term policy (e.g. for sustainable future transport systems). However, we should expect such models to somehow 'agree' with each other; transport network models on large spatial and temporal scales should be consistent with those commonly used for smaller scale, nearer term forecasting. In fact there is little understanding of what difference a change in scale makes to model predictions, how to connect current transport models across different scales of analysis and different data resolutions, and what it means for them to be consistent. In transport modelling, aggregation refers to the level of detail included in network and behavioural models, and the methods used to summarise characteristics of detailed models for larger scale analyses. For traffic flow dynamics on a single link, there is some theory on aggregating individual car-following models to give fluid flow PDEs and area speed-flow relationships. For transport networks there is no such theory of aggregation. A key challenge therefore is to establish theory and methods for the aggregation of network models; to connect existing models across different scales, and to provide aggregate representations of transport networks for large-scale, long-term analyses. In this lecture I will outline a new analytic method for network aggregation. With this in mind, I will consider how recent advances in the study of complex networks might be used in the large-scale, long-term analysis of transport networks.

Dr. Nikolas Geroliminis received his Ph.D. in civil engineering from University of California, Berkeley in 2007. Since then, he has been on the faculty of the Department of Civil Engineering at the University of Minnesota. He is a member of the Transportation Research Board's Traffic Flow Theory Committee. His research interests focus primarily on urban transportation systems, traffic flow theory and control, public transportation and logistics. In summer 2009, he will be an academic visitor at the Chair of Sociology, in particular of Modeling and Simulation, ETH. While a student at Berkeley, he received the University of California "Transportation Student of the Year" award in 2007, and the "Outstanding Graduate Student Instructor Award" in 2006.

Various theories have been proposed to describe vehicular traffic movement in cities on an aggregate level. They fall short to create a macroscopic model with variable inputs and outputs that could describe a rush hour dynamically. This work shows that a Macroscopic Fundamental Diagram (MFD) relating production (the product of average flow and network length) and accumulation (the product of average density and network length) exists for neighborhoods of cities in the order of 5-10km2. It also demonstrates that conditional on accumulation large networks behave predictably and independently of their Origin-Destination tables. These results are based on analysis using simulation of large scale city networks and real data from urban metropolitan areas. The real experiment uses a combination of fixed detectors and floating vehicle probes as sensors. The analysis also reveals a fixed relation between the space-mean flows on the whole network and the trip completion rates, which dynamically measure accessibility. The looking-for-parking phenomenon that extends the average trip length is also integrated in the dynamics of the rush hour. An analytical model based on Variational Theory describes the connection between network structure and a network's MFD for urban neighborhoods controlled at least in part by traffic signals. The MFD is applied to develop control strategies based on neighborhood accumulation and speeds and improve accessibility without the uncertainty inherent in forecast-based approaches.

Matteo Salani is an assistant professor at Transport and Mobility Laboratory at EPFL. He received his PhD in computer science in 2006 from University of Milano with a thesis on Branch and Price algorithms for Vehicle Routing Problems. His main research interests are optimization algorithms for complex combinatorial problems arising in transportation. He contributed to the field of Column Generation with publications on real world routing problems, advanced dynamic programming algorithms and resource constrained shortest paths. He is actually interested in robust optimization, intermodal transportation and some involving cases of routing problems.

Transportation systems include complex and structured operations that must be planned in advance in order to exploit the available resources, provide a reliable and competitive service and forecast system's performances. Decisions regarding transport operations are based on data which is frequently due to uncertainty. Moreover, unpredicted events may disrupt the current state of the system and force managers to take reactive decisions to recover to an operational state. Proactive decisions, i.e. decisions which take into account the uncertainty of the data, tend to robust solutions which are able to absorb data deviation and small disruptions. In this talk we address two complex transportation systems, namely air and maritime transportation, we identify some operation planning issues and illustrate a robust planning approach for both cases. For the airline case we focus on a robust recoverable approach for aircraft routing, i.e. a proactive strategy which accounts for the presence of a disruption recovery strategy. In the maritime case we focus on an integrated decision problem arising in container terminals and present a planning approach which accounts for congestion in terminal operations induced by the solution. Both problems are based on real world data.

<p>In 2006, Hoogendoorn was appointed the Antoni van Leeuwenhoek professorship on the topic ‘traffic flow theory and simulation’. Before that, he has been working as an Associate Professor at the Transport & Planning department of Delft University of Technology. He is the author of about 70 journal articles and over 200 conference contributions. </p><p> In the past five years, his research has involved theory, modelling, and simulation of traffic and transportation networks, focussing on innovative approaches to collect detailed, microscopic traffic data and the use of these data to underpin the models and theories using new techniques for model identification. The data collection methods have been applied to a multitude of situations in which driving behaviour could not be studied until now (e.g. adverse weather conditions, narrow lanes, work zones, incident sites). Analyses of these data have resulted in new insights into the large differences in driving behaviour (in terms of car-following, and lane changing) in case of normal and non-recurrent conditions. </p><p> Under his supervision, the pedestrian flow research was also furthered. The NOMAD pedestrian simulation model developed during the project ‘Collective Walking Behaviour in Public Spaces’. The model has been applied in several commercial projects in order to solve design problems. The model functionality (evacuation modelling, revolving doors, train access and egress) was extended based on new walking experiments. </p><p> Hoogendoorn has initiated and supervised the NWO-AMICI (‘Advanced Multi-agent Information and Control of multiclass Integrated traffic networks’) programme, which was aimed at developing new models and control strategies for regional traffic networks. Furthermore, research on travel behaviour in uncertain situations and the impact of traffic information thereon has lead to the development of the Travel Simulator Laboratory (TSL), which is an interactive laboratory to study the travel behaviour dynamics. </p><p> Currently, research on evacuation modelling for buildings and for regional networks is being performed. Hoogendoorn is involved in the upcoming evacuation training (TMO projec) pertaining to the evacuation of the Rijnmond area (area around Rotterdam). Finally, he is organizing the ICEM’09 (first International Conference on Evacuation Modelling and Management; September 2009).</p>

<p>This talk addresses the development of theory and models to explain, predict, and manage dynamic network transportation and traffic operations in case of exceptional events (e.g. incidents, adverse weather conditions, flooding, bushfires). The dynamics involve traveller responses, impacts of the event on infrastructure, the uncertain event dynamics, the spatio-temporal propagation of the event impacts, and the countermeasures (information, guidance, control). These are analysed in a dynamic game-theoretic setting considering three actors: network authorities, network users, and fate (the exceptional event). </p><p> We focus on travel behaviour and traffic operations in relation to the degrading infrastructure (roads, public transit networks, communication, etc.). The transportation system itself is characterised by a high degree of freedom and multitude of choice alternatives of its users and by the fact that the users’ reactions – in terms of driving behaviour, route choice, compliance to an evacuation plan, etc. – are an essential but also uncertain component of the system response. </p><p> During the talk, the newly developed model EVAQ will be introduced. We will explain the main characteristics of the model, and show its application to several case studies. Next to this, we will also show how the model can be used for the optimization of evacuation schemes using dedicated optimization algorithms. In doing so, we show that the efficiency of the evacuation can be improved drastically. </p>

Gunnar Flötteröd graduated in engineering informatics with a focus on systems engineering in 2003 from the Technical University of Ilmenau, Germany. He completed his PhD research on the calibration of microsimulation-based travel demand models in 2008 at the Berlin Institute of Technology. Currently, he is a visiting scholar at Ecole Polytechnique Federale de Lausanne.

The calibration of microsimulation-based demand models is a difficult problem, and there is no source of information that may be disregarded in this process. This work shows how traffic counts can be consistently incorporated in the calibration. The approach is mathematically sound, computationally efficient, and compatible with typical microsimulation techniques.

After studying physics and meteorology in Cologne and Paris, Kai Nagel got his Ph.D. in computer science at the University of Cologne about "fast microscopic traffic simulations". From 1995 to 1999 he was at Los Alamos National Laboratory as part of the "TRANSIMS" team. 1999-2004 he was assistant professor for Computer Science at ETH Zurich at the Institute for Scientific Computing. Since 2004 he is full professor for ``Transport systems planning and transport telematics'' at the Technical University of Berlin. His research interests include: large transportation simulations, modeling and simulation of socio-economic systems, multi-agent simulations.

MATSim stands for "Multi-Agent Transport Simulation". It is a behaviorally-based microscopic simulation of travel behavior and traffic flow. Conceptually, it combines an activity-based demand generation and a dynamic assignment into one coherent modelling and simulation package where synthetic travellers learn good or individually optimal behavior through many iterations. As the most important advantage of MATSim, detailed feedback from the traffic flow simulation is not only used for route choice, but also for other choice dimensions such as activity time scheduling and mode choice. The model is sensitive to policy measures since the synthetic travellers individually adapt to policy measures such as road closures, tolls, public transit price changes, etc. The adaptation is regulated by utility functions of the agents, which are related to utility functions in discrete choice theory, but not the same.

Maya Abou Zeid is a doctoral student at MIT specializing in travel behavior modeling. Prior to joining MIT, she was an Associate of Cambridge Systematics, Inc., with experience in travel demand modeling and market research. She received a Master’s degree in Transportation from MIT, and a Bachelor’s degree in Civil and Environmental Engineering from the American University of Beirut.

Understanding travel well-being is important both for a better representation of travel behavior models and for the design and evaluation of policies. Two data efforts aiming at measuring and modeling travel well-being are described. The first one is a cross-sectional travel and activity well-being survey that was conducted with a sample of commuters in the summer of 2007. Two main findings emerged from the analysis of this survey. First, commute satisfaction was found to be related to commute stress, social comparison, personality, and overall well-being. Second, happiness experienced from performing an activity was found to be related to the propensity of activity participation. Despite the insights gained from the cross-sectional survey, travel behavior is mostly habitual and people don’t usually fully consider their well-being from the travel alternatives available to them. The idea of the second data collection effort, currently being conducted in Switzerland, is to force people to reconsider their travel mode choices and measure their happiness at the time those choices are made. The experiment involves a temporary switch from car to public transport for a sample of commuters with strong car habits. The perceptions, attitudes, happiness, and choices of these commuters will be measured before and after the intervention. Preliminary results from the analysis of this experiment will be described.

Originaire de St-Maurice et Sion (VS), Dominique de Werra est né en1942. Il obtient son diplôme d'ingénieur-physicien de l'EPUL en 1965, puis en 1969, le titre de docteur ès sciences techniques. De 1969 à 1971, il est professeur au département de sciences du management à l'Université de Waterloo; il est professeur invité dans diverses Hautes Ecoles européennes et américaines. Depuis 1971, il est professeur de recherche opérationnelle à l'EPFL. Il a présidé la conférence des chefs de département en 1990, il est nommé vice-président de l'EPFL en 1990 et en plus, directeur de la formation depuis l'automne 93. Ses recherches portent sur les mathématiques discrètes (optimisation combinatoire, théorie des graphes, algorithmique, etc.) et leurs applications aux systèmes industriels et informatiques. Il a participé à et/ou dirigé divers projets interdisciplinaires en productique, distributique, énergétique et ordonnancement. Il dirige des travaux dans les domaines précités et en particulier sur les problèmes d'emploi du temps et plus généralement de gestion de calendriers et de ressources dans l'exécution de grand projets (sports, enseignement, etc.). En 1987-1988, il a présidé l'association EURO qui regroupe les sociétés nationales de recherche opérationnelle en Europe. Il est docteur h.c. de l'Université de Paris et de l'Ecole polytechnique de Poznan et lauréat de la médaille d'or européenne (EURO) de recherche opérationnelle en 1995. En mars 2000, il est nommé doyen des affaires internationales.

En tomographie discrète l’un des problèmes de base est la reconstruction d’une image à partir de ses projections (horzontales et verticales). En d’autres termes on a un tableau m x n où chaque case correspond à un pixel qui peut avoir l’une des couleurs 1,2,…, k. On donne pour chaque rangée (ligne ou colonne) le nombre de pixels de chaque couleur qui s’y trouvent . Il s’agit alors de voir si l’on peut construire une affectation de couleurs aux pixels qui satisfasse ces conditions de projection. Des modèles de graphes ont été proposés pour aborder ces problèmes et on en a étudié diverses variations et généralisations. Ceci a permis de formuler divers problèmes d’ordonnancement en particulier qui n’ont rien de tomographique. Nous discuterons certaines de ces variations. Nous montrerons quels sont les problèmes qui sont facilement solubles et nous étudierons la complexité d’autres problèmes. Nous montrerons comment divers modèles classiques de recherche opérationnelle permettent d’appréhender des cas particuliers du problème de la reconstruction d’image et nous mentionnerons quelques problèmes ouverts.

<p>Part 1: Variable Space Search (joint work with A. Hertz and M. Plumettaz) </p> <p> Three ingredients must be defined when designing a Local Search procedure LS (e.g. a descent method, simulated annealing, tabu search) for a particular problem: a search space S, an objective function f that measures the quality of each solution s in S, and a neighborhood structure N. In 1997, for a given objective function f, a solution space S and a local search procedure LS, Mladenovic and Hansen proposed the Variable Neighborhood Search (VNS) algorithm that uses several neighborhoods N1, …, Nr. VNS was already efficiently applied to many problems. In 2005, for a given objective function f, a neighborhood structure N and a local search procedure LS, Mladenovic, Plastria and Urosevic proposed the Reformulation Descent (RD) algorithm that uses several solutions spaces S1, …, Sr. RD was efficiently applied to circle packing problems. In 2007, we propose to generalize VNS and RD as follows. Consider a set of search spaces {S1, …, Sr} with their respective objective functions {f1, …, fr}. For each search space Si, consider a set Ni = {N(i,1), …, N(i,q)} of neighborhoods which can be used in Si for minimizing fi. Consider finally a set of translators T(i,j) that transform any solution in Si into a solution in Sj. The resulting Variable Space Search (VSS) algorithm successively performs a local search LSi in the different search spaces Si, always using the associated neighborhoods in Ni and the objective function fi. VSS was efficiently applied to the graph coloring problem. </p> <p> Part 2: Ant Local Search (joint work with M. Plumettaz and D. Schindl) </p> <p> We propose a new kind of ant algorithm called Ant Local Search. In most ant algorithms, the role of each ant is to build a solution in a constructive way. In contrast, we propose to consider each ant as a local search, where at each step and in concordance with all ant algorithms, each ant modifies the current solution by the use of the greedy force and the trail systems. We also propose ways to reduce the computational effort associated with each ant decision. Such a new and general ant methodology is then applied to the well-known k-coloring problem, which is NP-hard. Computational experiments give evidence that our algorithm is competitive with the best coloring methods. It is actually the first time that an ant algorithm is competitive with the best coloring methods. </p>

The fixing of variables based on reduced costs is one of the standard acceleration techniques in mixed integer programming (MIP) that has been used in numerous applications: If the reduced cost of an integer variable exceeds the duality gap, the variable can be fixed to zero. An extension of this idea is to consider reduced costs of several decision variables simultaneously: Any solution to a MIP, where the sum of the reduced costs of some integer variables exceeds the duality gap, is necessarily a non-optimal solution. We exploit this idea and show that hybrid exact algorithms with two phases can benefit from this general result. The first phase is the computation of a dual feasible solution to some (strong) relaxation of the problem. For a minimization problem, the dual solution provides a lower bound and reduced costs of the variables. Any upper bound then permits the traditional variable fixing for the individual variables. The second phase is a direct, enumerative decision-tree approach. The additivity of the reduced costs allows bounding the decision-tree algorithm whenever too many decision variables with positive reduced costs have been included in a (partial) solution. In order to empirically test the effectiveness of the new idea, we have developed a hybrid exact approach for the traveling salesman problem with time windows (TSPTW). Its first phase is a cutting-plane algorithm and the second phase a dynamic-programming algorithm. The new hybrid approach clearly and consistently outperforms the currently best methods at hand.

Après des études d'Informatique Théorique à l'Ecole Normale Supérieure de Lyon, Benjamin Lévêque a obtenu son diplôme de doctorat de Mathématiques et Informatique à l'Université de Grenoble en 2007. Depuis janvier 2008, il est en post-doctorat à l'Ecole Polytechnique Fédérale de Lausanne. Ses recherches portent sur les mathématiques discrètes : optimisation combinatoire, théorie des graphes, algorithmique, ...

We consider the problem of minimizing the size of a family of sets G such that every subset of 1,...,n can be written as a disjoint union of at most k members of G, where k and n are given numbers. This problem originates in a real-world application aiming at the diversity of industrial production. At the same time, the minimum of G so that every subset of 1,...,n is the union of two sets in G has been asked by Erdos and studied recently by Furedi and Katona. A simple construction providing a feasible solution is conjectured to be optimal for this problem for all values of n and k and regardless of the disjointness requirement; we prove this conjecture in special cases including all (n,k) for which n <= 3k holds, and some individual values of n and k.

Matthieu de Lapparent is researcher at the French National Institute of Research on Transport and Safety (INRETS), Department of Economics and Sociology of Transport. He holds a PhD in Economics from Université Paris 1 Panthéon-Sorbonne (2004). His areas of special interest are demand modeling and simulation, decision theory, microeconomics, econometrics, and road safety.

There is a plenty of literature on the topic of the value of travel time savings (VTTS) that covers both its theoretical and empirical dimensions. It is a key concept in transport analysis as it is used in assessment of user costs and benefits from a modification of transport infrastructure (traffic capacities, new modes of transport, etc.). Modelling and estimating realistic and reliable willingnesses to pay for saving travel time is therefore of utmost importance to make appropriate planning decisions. The outline of the presentation is as follows. In a first step, the theoretical framework of time allocation subject to budget, time, technological, space, and production constraints is presented. It is adapted from the original analysis frameworks of Becker(1965) and de Serpa(1971). Definition of the VTTS is derived from it. Its properties with respect to travel attributes and characteristics of the traveller are then discussed. The second part of the presentation turns to econometric modelling of the VTTS. It proposes flexible functional forms that take into account stylized facts and theoretical results. Estimation of various discrete choice models is made in a third step with Biogeme by using data from the 2001 Swissmetro SP survey. The results show that non linear utility functions in the presence of both deterministic and unobserved heterogeneity of tastes perform better than conventional approaches. Conclusion of the presentation discusses extensions and future research tracks. <ul class='text'> <li> Becker, G.S. (1965), A theory of the allocation of time, The Economic Journal, Vol. 75 (299), pp. 493-517</li> <li> Ben-Akiva, M., and S. Lerman (1985), Discrete Choice Analysis: Theory and - Application to Travel Demand, the MIT Press</li> <li> Bierlaire, M. (2003), BIOGEME: A free package for the estimation of discrete choice models , Proceedings of the 3rd Swiss Transportation Research Conference, Ascona, Switzerland.</li> <li> Lapparent, M.(de), de Palma, A. and C. Fontan, Nonlinearities in the valuation of travel attributes, Proceedings of the Professional Transportation Research Conference, 2002, Cambridge, UK.</li> <li> Daly, A. and Bierlaire, M. (2006), A General and Operational Representation of Generalised Extreme Value Models,Transportation Research Part B 40(4), 285-305</li> <li>De Serpa, A. C. (1971), A theory of the economics of time, The Economic Journal, 81 (321),. 828-846</li> <li> Hess, S., Bierlaire, M. and Polak, J. (2005), Estimation of value of travel-time savings using Mixed Logit models, Transportation Research Part A, 39(3), 221-236</li> <li> Train, K. (2003), Discrete choice with simulation, Cambridge University Press</li> <li> Truong, T.P and D.A. Hensher (1987), Measurement of Travel Time Values and Opportunity Cost from a Discrete-Choice Model, Economic Journal, vol. 95(378), 438-51 </li> </ul>

Javier Cruz was born in Barcelona, Spain, in May 1982. He received a M.S. degree in Mathematics and a M.S. degree in Telecommunications Engineering, both from the Technical University of Catalonia (UPC), Barcelona, Spain, in 2005 and 2006 respectively. During his studies, he worked on the resolution of the inverse kinematic problem of robotic structures at the Institute of Robotics and Industrial Informatics in Barcelona, he collaborated in the Euro-pean project CHIL (Computers in the Human Interaction Loop) at the UPC and he passed six months in the Signal Processing Laboratory at the Swiss Federal Institute of Technology, Lausanne, Switzerland, in order to do his Master thesis about multi-pose face detection tech-niques. In January 2007 he joined the Transport and Mobility Laboratory, led by Prof. Michel Bierlaire, where he is working towards his PhD degree in image processing.

In this presentation, the problem of generating virtual views of a scene will be presented. A virtual view is a view of a scene from a point where there is not a physical camera. The pres-entation begins with an introduction to the basics of multi-camera systems and how these sys-tems can be used to obtain the 3D information of the scene. With this 3D information a virtual view can be generated, but due basically to non-photorealistic results this is not the approach usually used. For this reason, in the second part of the talk, Image Based Rendering (IBR) techniques will be presented, giving a survey on the state-of-the-art on IBR.

Mr. Newman is currently finishing his Ph.D. in Civil Engineering at Northwestern University. He has been awarded fellowships from the U.S. Departments of Transportation and Homeland Security, as well as the Eno Transportation Foundation and Northwestern's Transportation Center. He also holds a Masters of Public Administration and a BS in Policy Analysis from Cornell University. Between his studies at Cornell and Northwestern, Mr. Newman worked for several years in both the public and private sectors, on projects ranging from public finance to labor relations to economic development. He is a member of the Disaster Action Team for the American Red Cross of Greater Chicago, and has served as a volunteer firefighter. Newman's primary research interests include the development of advanced discrete choice modeling techniques, and transportation policy making. In particular, he is interested in very large scale models, which are of interest in epidemiological research, where it is ideal to model the behavior of hundreds of millions of people simultaneously.

Generalized extreme value (GEV) models provide a convenient way to model choice behavior that is consistent with utility maximization theory, but the development of specific new models within the GEV family has been slow, due to the difficulty of ensuring new formulations comply with all the GEV rules. The network GEV structure provides a tool to quickly generate new models in the GEV family, without the burden of complex analysis of the new model to ensure its properties. However, the plethora of parameters in the model requires that numerous restrictions be imposed to allow for model identification, and different network topographies require different sets of restrictions to maintain an unbiased model in the estimation process. This presentation will examine a few different methodologies for imposing those restrictions, and how those restrictions can be leveraged to create a new type of model with a novel disaggregate correlation structure among alternatives.

Eran Ben-Elia (M.Sc) is a Ph.D. candidate in Transportation Science at the Faculty of Civil & Environmental Engineering at the Technion - Israel Institute of Technology. His interests lie in travel behavior modeling and intelligent transport systems.

Advanced Travel Information Systems (ATIS) are designed to provide real time information enabling drivers to choose efficiently among routes and save travel time. Travel demand modelers have been trying to analyze drivers' response to such systems. Psychological research suggests that route-choice models can be improved by adding realistic behavioral assumptions. However, different generalizations imply deviations in different directions. Specifically, different choices arise when decisions are taken on the basis of information compared to those taken on the basis of personal experience. An experimental study of route choices investigates the combined effects of information and experience on route-choice decisions in a simulated environment whereby the participants can rely on a description of travel time variability and at the same time can rely also on personal experience through feedback. The results show that the effect of information is positive and more evident when participants lack long-term experience on the distributions of travel times. Furthermore, information seems to increase initial risk seeking behavior, reduce initial exploration and contribute to between subject risk-attitudes differences. Based on this data we estimate several advanced discrete choice models using mixed logit specifications with panel data. The model estimation results shows that non-informed participants tend to rely more on recent outcomes and are more sensitive to travel times, less sensitive to travel time variability and show tendency towards risk aversion. In contrast, informed participants are more aware of past sequence of outcomes, are more sensitive to travel time variability and have some inclination towards risk-seeking behavior. These findings have implications for cost-effective ATIS design especially in the conditions characterized by non recurrent congestion. More research is necessary to better understand the behavioral impacts of informed users on the general equilibrium of transport networks especially the effect of driver interaction and joint decision making effects.‏

Ricardo Hurtubia, born Chilean, has an Industrial Engineering degree and a Master of Sciences with a major in Transport Engineering, both from the University of Chile. He has worked mostly in the development of the “Santiago Land Use Model” (MUSSA) in the Transport and Land Use Modeling Laboratory at the University of Chile (LABTUS).

A dynamic equilibrium model for the real estate market is presented. The model is based on the theoretic framework developed for the “Santiago Land Use Model” (MUSSA), but with some relevant differences which will be highlighted through this presentation. Households have stochastic behavior and compete for quasi-unique locations (real estate goods), which are assigned to the best bidder through an auction-type mechanism. Producers are modeled as maximizers of their profits over the long-term through the production of real estate assets, represented by the expected present value of future sales. It is assumed that the producers do not possess complete information about future levels of demand or prices; instead they take the actual and historic prices in each period as the relevant information for their decision-making. A notion of equilibrium is used that adjusts prices given two situations: supply and demand surplus. In the supply surplus case, the prices are diminished and supply in the market is reduced until it equals demand. In the case of demand surplus, the prices rise and demand diminishes (homeless households) until it equals supply. This equilibrium condition yields prices that are jumpy over time, resembling observations of inventories in the real estate market and the manufacture industry.

Jingmin Chen, born in 1985, graduated in 2005 from Sun Yat-sen University, China. His first major in college was mathematics. After two years’ study in math, Mr. Chen transferred to transportation engineering and received a bachelor’s degree of engineering. He is currently a master student in Southeast University. His research interests are transportation network reliability and route choice behavior modeling. He presented a relevant paper in 3th International Symposium on Transportation Network Reliability and has published two papers in national journals.

In this seminar, I would like to present two of my papers, my bachelor’s thesis and master’s thesis respectively, regarding route choice model and algorithm. The title of bachelor’s thesis is “comparative study of Logit-based network loading algorithms”, which concerns analytical and numerical comparisons of traditional STOCH algorithm and Topological-scan based Logit algorithm (T-Logit). Analytical study shows that T-Logit is more accurate than STOCH. Detailed numerical experiments illustrate how Logit parameter influences the loading result. My research topic of master’s thesis is “route choice behavior modeling with travel time reliability requirement under uncertain demand”. Mixed-strategy is employed to model traveler’s route choice behavior, which also considers traveler’s travel time reliability requirement when they face an unreliable traffic network. Demand uncertainty is the source of unreliability. An understandable and resolvable SUE mathematical program is constructed. The model can be easily extended to multi-class user type and numerical examples can prove its validity.

Multi-agent micro-simulation is becoming increasingly important in traffic simulation, traffic analysis, and traffic forecasting. It allows one to model the decision making processes of each single individual explicitly. This is of importance since it is not a vehicle which produces traffic; it is the person who drives the vehicle. Persons do not just produce traffic; instead each of them tries to manage his/her day (week, life) in a satisfying way. Therefore, the decision makers do not only base their decisions on a specific part of their day (i.e. a single trip), each decision is a result of their complete daily needs. So, some of the major challenges in multi-agent transport simulations are - to model the success of a complete individual daily demand, - to model the decisions of each individual based on the success-rate of his/her demand, - to model the diversity of adaption of each individual daily demand to achieve a more successful day, - and---last but not least---to reach a (globally) stable state for all individuals synchronously. We present the open source project MATSim-T (Multi-Agent Transport Simulation Toolkit), a modular approach for large scale scenarios to optimize individual (daily) demand. While the system is able to cope with transport scenarios of millions of individuals, the models are very simple in respect to travel behavior. This presentation is meant to produce a open discussion in the topic of individual decision making processes within MATSim.

Alain Hertz est professeur au Département de Mathématiques et de Génie Industriel de l'École Polytechnique de Montréal. Ses activités de recherche sont centrées principalement sur l'optimisation combinatoire et la théorie des graphes appliquées à la résolution de problème d'horaires et de transport.

We consider a monthly crew scheduling problem with preferential bidding in the airline industry. We propose a new methodology based on a graph coloring model and a tabu search algorithm for determining if the problem contains at least one feasible solution. We then show how to combine the proposed approach with a heuristic sequential scheduling method that uses column generation and branch-and-bound techniques. (joint work with M. Gamache and J. Ouellet)

Fabrice Marchal est charge de recherche CNRS au Laboratoire d'Economie des Transports, Lyon. Ses recherches portent principalement sur la simulation des systemes de transports a grande echelle (modeles METROPOLIS et MATSIM) ainsi que sur les modeles integres transport-occupation du sol.

In this talk we present a methodology to calibrate a class of dynamic traffic simulation models with real data acquired from traffic counts and travel time measurements acquired from GPS devices. Calibration can be described as an optimization process with an objective function that depends on simulation outputs. That function is often noisy because simulations include stochasticity and costly to compute because a single evaluation implies a complete microscopic simulation, a task which can range from a few minutes to up to an hour of computer time. We explore the use of population based optimization algorithms such as genetic algorithms in general and Covariance Matrix Adaptation-Evolutionary Strategy (CMA-ES) in particular. We propose BBOG, a framework designed to swap different algorithms in order to compare their performances for traffic calibration. The framework is tested using CMA-ES to calibrate METROPOLIS simulations running on a cluster of Linux servers. As a proof of concept, we present the results of the calibration procedure with an application to the toy network of Sioux Falls. Two important findings are derived from this exercise. Firstly, state of the art evolutionary algorithms are promising search methods to calibrate traffic simulations, at least when the number of parameters is relatively small. Secondly, a unified framework to compare algorithms provides a richness of analysis that is usually missing from the traditional application of a given algorithm to a particular traffic problem.

Since June 1st, 2007, Dirk Helbing (born January 19, 1965) is Professor of Sociology, in particular of Modeling and Simulation at ETH Zurich. Before, he worked as Managing Director of the Institute for Transport & Economics at Dresden University of Technology, where he was appointed full professor in 2000. Having studied Physics and Mathematics in Göttingen, his master thesis dealt with the nonlinear modelling and multi-agent simulation of observed self-organization phenomena in pedestrian crowds. Two years later, he finished his Ph.D. at Stuttgart University on modelling social interaction processes by means of game-theoretical approaches, stochastic methods and complex systems theory, which was awarded two research prizes. After having completed his habilitation on traffic dynamics and optimization in 1996, he received a Heisenberg scholarship. Both theses were printed by international publishers. Apart from this, Helbing has (co-)organized several international conferences and (co-)edited proceedings or special issues on material flows in networks and cooperative dynamics in socio-economic and traffic systems. He has given 220 talks and published more than 200 papers, including several contributions to journals like Nature, Science, or PNAS, which were discussed by the public media (newspapers, radio, and TV) more than 200 times. He collaborates closely with international scientists. For example, he worked at the Weizmann Institute in Israel, at Xerox PARC in Silicon Valley, at INRETS in Paris and the Collegium Budapest - Institute for Advanced Study in Hungary, where he is now a member of the external faculty.

<p>In the past, physicists have discovered various self-organized phenomena in pedestrian crowds such as the formation of lanes of uniform walking direction, oscillations at bottlenecks in bi-directional flows, the formation of stripes in cross-flows, or "freezing-by-heating" and "faster-is-slower" effects in panicking crowds. These phenomena have all been successfully described by driven many-particle models, as will be demonstrated by animated computer simulations and video recordings of real crowds. </p> <p> Panic stampedes are a serious concern during mass events. However, despite huge numbers of security forces and crowd control measures, hundreds of lives are lost in crowd disasters each year. A high-performance tracking analysis of unique video recordings of the Muslim pilgrimage in Mina/Makkah, Saudi Arabia, has now revealed that high-density flows can even turn "turbulent", which causes people to fall. The occuring eruptions of pressure release bear analogies with earthquakes and are de facto uncontrollable. This talk presents an analysis and interpretation of our recent discoveries and shows that the measurement of the gas-kinetic "pressure" is suitable for an advance warning of critical crowd conditions.</p>

Due to their practical relevance, vehicle routing problems received a particular interest in the operations research during the last decades. In its simplest version, the problem consists in routing a fleet of vehicles in order to deliver goods to a set of customers, minimizing the operating costs. Several important variants have been addressed in the literature, commonly modeling capacities on the vehicles. While in all these variants it is assumed that each customer must be served by exactly one vehicle, relevant savings can be achieved by allowing multiple visits to each customer, each serving only a fraction of its demand. In this talk we describe the so-called Vehicle Routing Problem with Split Deliveries, which is by far less studied and understood than the classical one. We discuss its properties, we outline the main issues arising in the solution process and we present a modeling and algorithmic framework based on column generation. Finally, since the aim of the talk is to sketch our research directions, we discuss related open problems.

Alexandre Alahi was born in Lyon, France, in August 1981. He received the M.S. degree in Communication Systems from the Swiss Federal Institute of Technology, Lausanne, Switzerland, in 2006. During his studies, he earned a one year exchange fellowship to study at Carnegie Mellon University, Pittsburgh, PA, USA. He has working experience in the field of image processing and computer vision in various companies. Among them, we can cite Logitech (Silicon Valley-CA, USA), and Mitsubishi Electric Research Laboratories (Cambridge ­ MA, USA). At Logitech, in 2004, he reduced the cost of their webcam by optimizing the processing of their color processing pipe. From September 2005 to March 2006, at Mitsubishi Electric Research Laboratories, he developed a 3D scene simulator to measure the performance of scheduling algorithms to control pan-tilt-zoom cameras. He is now working towards his Ph.D. degree in scene understanding with multiple moving and fixed cameras.

Scene understanding with multiple modalities has been one of the main focus of several research communities. Low-cost digital cameras and the progress to process large data sets such as digital images have promoted the development of vision-based analysis. In this talk, current trends to detect and track objects of interest in a scene with digital cameras will be discussed. A novel framework will be described to take advantage of multiple fixed and mobile cameras to enhance the performance of the system. Advantages to deal with fixed cameras and challenges to use mobile ones will be discussed.

Gerald Reiner studied Business Administration in Vienna after an education in Industrial Engineering. He received his Master and Doctorate in Business Administration (Quality Management and Production Management) at the Vienna University of Economics and Business Administration (WU). In 2006 he received his Habilitation (venia legendi) in Business Administration at the WU. From 1996-98 he worked as a research assistant at the Department of Industrial Information Processing (WU). From 1999-2006 he was Assistant Professor at the Department of Production Management (WU). In March 2006 he was guest professor at the Aston Business School (UK). In 2007 he was visiting professor at the University of Lausanne (CH). Since 2007 he is full professor at the Enterprise Institute at the University of Neuchâtel (CH). His Research interests lie in the field of Supply Chain Management, Quality Management and Operations Management. Gerald Reiner has published research articles, e.g. in the International Journal of Production Economics, International Journal of Production Research, and OR-Spectrum. Furthermore, he is referee for international journals, e.g., European Journal of Operational Research, Simulation Modelling Practice and Theory.

We develop a procurement model and the framework for performance evaluation for products with fluctuating prices by explicitly considering the transport process. The main decision for each time period under consideration of the selected contract type is the "ideal" procurement quantity. The main constraints are related to speculative inventories and alternative modes of transport. The performance of different procurement strategies is illustrated for a real supply process in the chemical industry by using process simulation.

José Viegas is Full Professor in Transportation at the Department of Civil Engineering and Architecture of the Instituto Superior Técnico, since 1992. Founder and Chairman of the Board of TIS.pt, consultants in Transport, Innovation and Systems, s.a., 2000- Researcher at CESUR (Research Centre for Urban and Regional Systems, of the Technical University of Lisbon), since 1986.

Prof. José Viegas, Director of research group on Transport Infrastructure, Systems and Policy at Instituto Superior Técnico, Lisbon presents a structured view of the ongoing and recently completed projects in his group, as well as the strategy behind those choices. These projects include some on key aspects of problem formulation and some where the main emphasis is the development, appraisal and possibly demonstration of innovative solutions for relevant contemporary problems, from congestion to emissions or infrastructure financing. The central element of this strategy is to avoid riding the waves of fashion on research done in the more developed countries, and concentrate the efforts instead on the identification of the structural elements in the formation of problems that could be tackled from a new angle. Some interesting examples of real or near-to-real application will be shown, as well as suggestions for deeper collaboration between the Lisbon and Lausanne groups.

Norbert Trautmann received the Business Engineering, PhD and Habilitation degrees from the University of Karlsruhe (Germany) in 1997, 2000, and 2004, respectively. His PhD thesis was awarded by the Ger- man Operations Research Society. Since 2005 he is an Assistant Professor in Quantitative Methods at the Department of Business Administration of the University of Bern. Norbert Trautmann's main research areas are Production and Opera- tions Management and Project Scheduling. He has published in Interna- tional Journal of Production Research, International Transactions in Oper- ational Research, European Journal of Operational Research and OR Spectrum.

We consider the problem of planning and scheduling the execu- tion of physical and chemical transformation processes on a multi-product chemical batch production plant. Such a plant consists of several multi- purpose processing units and storage facilities of limited capacity. Given primary requirements for final products, the problem consists in generating an appropriate set of batches for each process and scheduling the process- ing of these batches on the processing units subject to different types of technological constraints. In the literature, the short-term planning problem is generally modeled as a monolithic mixed-integer linear program. Due to the combinatorial nature of the problem, those models generally cannot be used when dealing with problem instances of practical size. In this talk we present a two-level approach which is based on a decom- position of the problem into a batching and a batch-scheduling problem. We formulate the batching problem as a mixed-integer nonlinear program that can be transformed into a mixed-integer linear program and solved using standard software. For solving the batch-scheduling problem, we sketch a schedule-generation scheme and a corresponding multi-pass heuristic. We report on computational results for two sample production processes from the chemical engineering literature.

Giovanni Righini is associate professor of Operations Research at the Department of Information Technologies of the University of Milan, where he founded the Operations Research Laboratory OptLab. He holds a master degree in Electronic Engineering (1988) and a doctorate in Computer and Automatic Engineering (1993) from the Technical University of Milan. His research activities mainly concern the development of mathematical programming algorithms as well as heuristic and metaheuristic algorithms for combinatorial optimization problems, especially in the field of logistics. He is actively involved in applied research projects funded by public institutions and private companies.

The optimization of emergency systems is a challenge for mathematical and statistical methods, both from the viewpoint of model definition and from the viewpoint of algorithms design. In spite of its crucial importance, the optimization of emergency systems has not been so often addressed in a scientific way so far. Therefore it is an open field where Operations Research can provide valuable contributions. In this talk I will present some optimization subproblems arising in the context of a case study concerning the "118 service" in Milan.

Piet H.L. Bovy (1943) graduated as a civil engineer from the Technical University of Aachen (Germany). He completed his PhD at the Delft University of Technology on the subject of stochastic route choice modeling in networks. He gained his professional experience in various occupations such as at a German traffic consultancy firm, a university research institute at DUT, and the Dutch ministry of Transportation. Since 1993 he is full professor for Transportation Planning at the Delft University of Technology, Faculty of Civil Engineering and Geosciences, P.O. Box 5046, NL-2600 GA Delft, The Netherlands. p.h.l.bovy@tudelft.nl Education and research focus on transportation planning, travel behaviour analysis, transport modeling, and networks. Books on transportation to which he contributed are: Route Choice: Wayfinding in Transport Networks (Kluwer Academic Publishers, 1990) and A billion trips a day: Tradition and transition in European travel patterns (Kluwer Academic Publishers, 1993). Recently he edited a series of books on transport research at DUT such as on Motorway traffic flow analysis (Delft University Press, 1998), on Estimators of travel time for road networks (DUP, 2000), and on Modelling for transport systems planning (DUP, 2002). He has written numerous articles in top journals and conference proceedings (ISTTT, IATBR, TRB).

Choice sets of individual travellers play a paramount role in analyzing travel choice behaviour since size and composition of choice sets do matter in cases of choice model estimation and demand prediction. Incorrect choice sets (e.g. because of captivity) can lead to misspecification of choice models and to biases in predicted demand levels. The critical role of choice sets in choice modelling has given rise to profound research into choice set modelling in the transportation field, although largely confined to mode choice. We state that these insights gained on choice set modelling and choice set generation cannot simply be transferred to the route choice realm. For a variety of reasons the specification of route sets for route choice modelling is different and more complex, reason why this topic deserves special attention from researchers and practitioners. The talk will be devoted to a number topics related to the modelling and generation of route choice sets, specifically for application in large networks. In addition the talk investigates recent new approaches to choice set generation for large networks. The one group of promising approaches is the constrained enumeration procedures, often referred to as branch and bound techniques in the research field of road traffic assignment, or as rule-based assignment in the field of schedule-based assignment. Another promising line is combined stochastic (which may include several elements) and deterministic path search. We will summarize in what respect route choice sets differ from other travel choices implying that some proposed choice set modelling approaches cannot be adopted for routes. We will argue that it is necessary and advantageous to distinguish the processes of choice set formation and choice on the part of the traveller. We state that the different purposes of choice set generation, that is supply analysis, model estimation, and demand prediction do matter in choice set modelling. We present a generic conceptual scheme relating the distinct key elements inherent in each choice set generation approach.

Après avoir entamé des études supérieures en géographie à l’université Cheikh Anta Diop de Dakar, Mlle Basse est depuis septembre 2000 étudiante à l’université de Nice Sophia Antipolis. Elle se spécialise dans l’aménagement du territoire notamment dans les effets structurants des grandes infrastructures de transport. Elle est aujourd’hui dans la deuxième année de doctorat et sa recherche concerne les effets structurants des grandes infrastructures de transport sur l’espace géographique.

Les systèmes territoriaux sont composés d’éléments visibles et invisibles qui entrent en interaction et révèlent le dynamisme des territoires. Ces systèmes sont souvent caractérisés par des incertitudes, du fait de leur fonctionnement complexe. Et puisque la modélisation aide à la connaissance et à la compréhension du fonctionnement des différents éléments qui entrent en interaction dans l’espace géographique, il est nécessaire d’utiliser des modèles ayant la capacité de prendre en compte ces incertitudes. La modélisation prédictive, à partir de la théorie des évidences de Dempster-Shafer et de la théorie de Dezert-Smarandache, est un moyen de prendre en compte incertitudes, paradoxes et imprécisions en vue d’une prise de décision. Mots clés : Modélisation prédictive, Incertitude, Territoire transfrontalier, Interactions entre Territoire et Transports, LGV PACA, Théorie des évidences de Dempster-Shafer, Théorie de Dezert-Smarandache

A typical marketing department usually runs a portfolio of mass campaigns such as TV Spots, Radio, Inserts, Promotional offers, Co-Marketing, etc. Running such campaigns is usually a very costly process both in terms of budget and resources. Moreover, there usually is a lack of quantitative measurement of the effectiveness of these campaigns as well as the return on marketing investment in terms of marginal sales and market share. The goal of this study is to build a marketing measurement and optimization system which helps marketing managers to address two key issues in their decision processes: 1) Quantifying the marginal impact of incremental investment in a given type of mass campaign on sales and market share 2) Determine an optimal budget and resource allocation strategy across the portfolio of campaigns which maximizes the total profit and market share. To address these issues, we adopt a modeling process which consists of two steps. We firstly develop a set of machine learning models using historical data. Secondly, we combine those models into a nonlinear objective function which measures the overall Utility of the campaigns portfolio under budget and resource constraints. Optimal campaign diversification and budgeting is then achieved by maximizing such a utility function.

We consider the problem of simulation-based traffic state estimation. Our traffic model is comprised of two major components: a mixed micro/macro traffic flow simulator and a behavioral model of combined route and activity location choice. The physical model moves individual agents based on a macroscopic representation of traffic flow dynamics. The behavioral model is simulated by a combination of a time variant best path algorithm and dynamic programming, yielding a behavioral pattern that minimizes a traveller’s perceived cost. The problem of traffic state estimation is considered in a Bayesian setting. The behavioral model comprises the a priori information, which is combined with anonymous traffic measurements e.g. of flows or velocities. As a solution algorithm, we use an iterative procedure that repeatedly linearizes the available measurements' likelihood with respect to individual turning decisions in the traffic flow model. This results in a convenient solution update scheme which consists in an ordinary simulation run based on additively modified travel cost. An illustrative example is given and potentials for further enhancements are noted.

Les budgets-temps de transport sont réputés stables depuis plusieurs décennies. Etablie sur les travaux fondateurs de Zahavi, cette conjecture suppose que la moyenne par agglomération des temps quotidiens de transport est d’une durée invariable d’une heure. Cette stabilité suggère une gestion paradoxale des gains de temps à deux niveaux. (1) Tout d’abord, au niveau urbain, les gains de vitesse n’ont pas été utilisés pour passer moins de temps dans les transports, mais pour aller plus loin. La mise en cause progressive de cette conjecture nous conduit à redéfinir le sens et à préciser la portée novatrice de la vision des comportements de Zahavi. Dans un second temps, la comparaison internationale des budgets-temps de transport met l’accent sur l’articulation des espaces-temps offerts par la ville. Elle semble indiquer un réinvestissement et un probable surinvestissement des gains de temps en transport supplémentaire. Tous deux attirent l’attention sur les limites des politiques urbaines et des transports en matière de régulation des mobilités et du développe- ment urbain. (2) Ensuite, rapportées au comportement individuel d’allocation des temps, les gains de temps n’ont pas été consacrés à d’autres activités. La substitution attendue entre les temps de transport et d’activités soulève la question des relations entre les durées d’activités et celle de la représentation de la demande dérivée de transport. L’analyse proposée de la dimension temporelle de la mobilité individuelle soumet l’idée selon laquelle le choix du temps de transport relève à la fois du coût d’accès aux opportunités et de l’activité en soi. Un modèle microéconomique de l’allocation des temps aux activités est proposé afin d’intégrer le double rôle du temps de transport. Puis, un modèle de durées est appliqué aux budgets-temps de transport de Lyon. Il révèle l’influence d’attributs individuels, les relations avec les budgets-temps des autres activités et il caractérise la dynamique temporelle du processus de mobilité.