The summer school is cancelled due to the coronavirus.
Schedule
Monday - 01/06/2020
09:00 - 10:30 Michel Bierlaire
11:00 - 12:30 Claudia Archetti
14:00 - 15:30 Luce Brotcorne
16:00 - 17:30 Claudia Archetti
Tuesday - 02/06/2020
09:00 - 10:30 Claudia Archetti
11:00 - 12:30 Michel Bierlaire
14:00 - 15:30 Luce Brotcorne
16:00 - 17:30 Stein W. Wallace
Wednesday - 03/06/2020
09:00 - 10:30 Stein W. Wallace
11:00 - 12:30 Luce Brotcorne
14:00 - 21:00 Social event
Thursday - 04/06/2020
09:00 - 10:30 Michel Bierlaire
11:00 - 12:30 Stein W. Wallace
14:00 - 15:30 Michel Bierlaire
16:00 - 17:30 Michel Bierlaire
20:00 - 23:00 Conference dinner
Friday - 05/06/2020
09:00 - 10:30 Stein W. Wallace
11:00 - 12:30 Student talks
Lectures
Claudia Archetti (University of Brescia) Solution methods for recent and challenging routing problems
E-commerce is experiencing an exponential growth in the last years. Recent statistics have shown that, while the retail e-commerce sales were around 1 billion dollars in 2014, they are estimated to reach 5 billion dollars in 2021. While this represent a huge business opportunity for companies in almost all fields, it also raises dramatic issues in terms of management of the operations related to satisfying customers’ orders. In particular, focusing on the last step of the supply chain, e-commerce explosion has raised the footlights on last-mile delivery. From a scientific point of view, distribution problems arising in last-mile delivery systems constitutes variants of routing problems. In this talk, some challenging routing problems arising in the context of last-mile delivery will be presented, like routing problems with occasional drivers and problems with release and due dates. We will focus on novel aspects of the context in which those problem arise, which are typically characterized by a high degree of dynamism. In addition, we will discuss the challenges related to developing solution approaches capable of handling the complexity of the problem applications.
In addition to the opportunities and challenges mentioned above, technological developments generated new and fast ways of communication and data exchange. In turn, this created business opportunities related to integrating the management of operations and activities that used to be handled separately. A noticeable example is the inventory routing problem, where a central decision maker defines the distribution plan involving a set of different actors (customers) and manages both transportation and inventories. The inventory routing problem is an extremely challenging problem which has attracted a lot of attention in the research literature. In this talk, we will present exact and heuristic solution approaches for its solution.
Michel Bierlaire (Ecole Polytechnique Fédérale de Lausanne) Behavioral optimization
Modern transportation demand models are based on advanced discrete choice models. They are based on a disaggregate representation of demand (individual-based, and not flow-based), and allow to account for detailed behavioral aspects, including subjectivity. These models are mathematically complex, in the sense that they are not always expressed in closed form, and are non linear and non convex. For these reasons, they have not yet been seriously considered in the operations research literature. In this series of lectures, after an extensive introduction to these models, we propose a methodology to include them in “classical” OR models, such as MILP. This leads to optimization models that explicitly account for the behavior of the actors involved.
Luce Brotcorne (Inria Lille) Bilevel Programming and its applications to Network Pricing and Energy Management
The purpose of this cursus is to introduce the field of bilevel optimization and its applications. The schedule will be divided into two parts: a theoretical study of bilevel programs and a focus on two applications in logistics and in the energy field. Bilevel programming is a fairly recent branch of optimization that deals with programs whose constraints embed an auxiliary mathematical program. More precisely, bilevel programs allow modeling those situations in which a main agent, whom we call the leader, strives to optimize a given quantity but controls only a subset of the decision variables. The remaining variables fall under the control of a second agent, the follower, who solves its own problem by taking into account the decisions taken by the leader.
The optimal solution to such an interactive process constitutes what economists call an equilibrium problem where the demand function results from the solution of an optimization program. Bilevel programs are also closely related to Stackelberg (leader-follower) games, to the principal-agent paradigm in economics, as well as to equilibrium constrained mathematical programs (MPECs), where the lower level problem characterizes the equilibrium state of some physical or economical system, and is frequently modelled as a variational inequality.
Although a wide range of applications fit the bilevel programming framework, real-life implementations are scarce, due mainly to the lack of efficient algorithms for tackling large-scale problems. Indeed, as a general rule, bilevel models are nonconvex and nondifferentiable. Therefore the structure of the problem has to be exploited in the design of efficient solution methods.
In this lecture we first present a brief theoretical study of bilevel problems. Next we focus on two special cases : a price setting problem on a network and a demand side and revenue management problem in the energy field.For each of these two applications, we define the models, study their properties and sketch solution methods based on the structures of the problems.
Stein W. Wallace (NHH - Norvegian School of Economics) Handling randomness in logistics modelling
- What-if analysis, does it work for handling uncertainty in optimization? I explain why it does not, and hence why we need more sophisticated tools.
- Scenarios and stability – how to model uncertainty in optimization models.
- Network design under uncertainty.
- Handling high-dimensional dependent random variables. I demonstrate how we are able to solve vehicle routing problems with over 25,000 correlated random speed variables with an accuracy of about 1%, and then discuss the issue in more generality.