Plenary speakers

 


Plenary talk: Robustness in Scheduling

Marjan van den Akker (Utrecht University, The Netherlands)

 fotoMarjan7.png

Traditionally, scheduling is based on the assumption that all input parameters are deterministic. Since in real-life situations uncertainty is unavoidable, recently more attention has been given to scheduling with stochastic processing times. Our goal is to obtain robust solutions for stochastic scheduling problems. To the best of our knowledge, no universal formal definition of robustness exists. An intuitive definition is: a schedule which does not significantly degrade in the face of disruption is called robust. In this presentation we discuss different models and algorithms for robustness in scheduling with an emphasis on parallel machine scheduling with precedence constraints.

After her PhD in scheduling algorithms supervised by Jan Karel Lenstra, Marjan van den Akker has worked CORE (Louvain-la-Neuve) as postdoc and at the Netherlands Aerospace Centre NLR as expert on modelling, optimization, and simulation in Air Traffic Management and Electronic Road Pricing. Since 2000, she is at the Department of Information and Computing Sciences at Utrecht University.

Her research area is advanced algorithmsrobustness and simulation. In her research characteristics from practice are combined with state of the art theoretical models. When solving optimization problems, important side-constraints are included as much as possible, even though this increases the complexity: the goal is to push the boundary of what is computable as far as possible. To achieve this, her research includes the application of LP-based decomposition techniques such as column generation, as well as approaches using local search. The last few years, she has focused on planning and scheduling under uncertainty. Traditionally, scheduling and planning problems were modeled as problems with fixed (deterministic) data. Since in real-life situations disturbances occur frequently, robustness is receiving an increasing amount of attention.  The purpose is to develop algorithms for finding solutions that, either remain valid in case of a disturbance, or can easily be adjusted to a feasible solution without having to solve the problem all over again. The question is “How can we capture uncertainty as well and efficient as possible in a deterministic model?”  Her work includes applications in energy networks and public transportation, where she is supervising different PhD students including projects with companies.

Her work is published in well-known international journals and conferences in Computer Science and Operations Research. She is a board member of the Dutch Network on the Mathematics of Operations Research (LNMB).

 


Plenary talk: Modeling and solving complex job-shop scheduling problems

Stéphane Dauzère-Pérès (Ecole des Mines de Saint-Etienne, France)

 Photo_Dauzere_Peres.jpg

This talk focuses on the flexible job-shop scheduling problem, first extensively studied in the 1990’s, which was later extended to include additional constraints and criteria to become complex job-shop scheduling problems. The complexity related to the problems is first discussed, in particular by differentiating with the classical job-shop scheduling problem. The main characteristic of the flexible job-shop scheduling problem is that operations can be performed on several resources, i.e. that operations must be both assigned to and sequenced on resources. Modelling choices and solution methods will be surveyed, including some recent contributions related to the consideration of batching constraints, sequence-dependent setup times and multiple criteria. Some of the results are based on a long-term collaboration of more than 15 years with two manufacturing sites of the French-Italian semiconductor company STMicroelectronics.

Stéphane Dauzère-Pérès is Professor at Mines Saint-Etienne in its site of Gardanne, France, and Adjunct Professor at BI Norwegian Business School, Norway. He received the Ph.D. degree from Paul Sabatier University in Toulouse, France, in 1992 and the H.D.R. from Pierre and Marie Curie University, Paris, France, in 1998. He was a Postdoctoral Fellow at the Massachusetts Institute of Technology, U.S.A., in 1992 and 1993, and Research Scientist at Erasmus University Rotterdam, The Netherlands, in 1994. He has been Associate Professor and Professor from 1994 to 2004 at the Ecole des Mines de Nantes, France, where he headed the team "Production and Logistic Systems" between 1999 and 2004. He was invited Professor at the Norwegian School of Economics and Business Administration (NHH), Norway, in 1999. Since March 2004, he is Professor at Mines Saint-Etienne, where he headed the research department "Manufacturing Sciences and Logistics" from 2004 to 2013.

His research interests broadly include modeling and optimization of operations at various decision levels (from real-time to strategic) in manufacturing and logistics, with a special emphasis on production planning (lot sizing) and scheduling and on semiconductor manufacturing.

He has published nearly 80 papers in international journals and has contributed to more than 200 communications in national and international conferences. Stéphane Dauzère-Pérès has coordinated numerous academic and industrial research projects, including 4 European projects and 24 industrial (CIFRE) PhD theses, and also six conferences. In particular, he co-organized in 2010 the first edition of the International Workshop on Lot Sizing which was held in Gardanne, France. In 2014, he created with Bernardo Almada-Lobo (University of Porto, Portugal) the EURO Working Group on Lot-Sizing (LOT), that he coordinated until 2018. He was runner-up in 2006 of the Franz Edelman Award Competition, and won the Best Applied Paper of the Winter Simulation Conference in 2013. His h-index is 36.

 


Plenary talk: Data driven Project Management

Mario Vanhoucke (Ghent University, Belgium)

MarioPas2010.jpg

This presentation will give an overview of the past endeavours and the recent trends in integrated project management and control, with a focus on linking scheduling to risk and control management. The presentation will show the relevance of using artificial and real project data for both research and practice. It will be shown that not many organisations have as much data as they often claim, and researchers therefore have to fall back on the use of artificial project data. Consequently, an overview of the most important artificial project datasets (each with advantages and disadvantages) will be given, and also a new set of empirical projects (freely available to researchers) will be presented.
Some recent research trends will be highlighted, illustrating that the integrated use of empirical data and advanced techniques (machine learning) might lead to promising results, and should therefore define the path for future research. References to a literature overview of project control will be given to outline the future research on integrated project management and control.
During my talk, I will also present my nice team of young researchers to you!

Prof Dr Mario Vanhoucke is a Full Professor at Ghent University (Belgium), Vlerick Business School (Belgium) and UCL School of Management (University College London, UK). He teaches "Project Management", "Applied Operations Research" and "Decision Making for Business". He obtained a Master’s Degree in Business Engineering (1996) and a PhD in Operations Management (2001), and he was director of EVM Europe (www.evm-europe.eu) and partner at the company OR-AS (www.or-as.be) until 2018.

Mario is responsible for various research projects in the field of Integrated Project Management and Control, which has led to more than 60 papers in international journals, five Project Management books published by Springer, three free online books (www.or-as.be/books), three computerised business games and an online learning platform known as PM Knowledge Center (www.pmknowledgecenter.com).

His research has received multiple awards, including awards from PMI Belgium (Belgium, 2007), International Project Management Association (Italy, 2008) and the American Accounting Association (USA, 2010) and multiple awards from Belgian companies. He currently has a team of +10 enthusiastic PhD students who are jointly working on improving decision making in project management, with a strong focus on (1) developing methods to improve project scheduling, (2) analysis risk in projects and (3) validating current and newly developed statistical methods for project control.

 


Industrial plenary talk: Industrial project and machine scheduling with Constraint Programming

Philippe Laborie (IBM France Lab, Gentilly, France) 

Philippe_Laborie.jpg

More often than not, project and machine scheduling problems are addressed either by generic mathematical programming techniques (like MILP) or by problem-specific exact or heuristic approaches. MILP formulations are commonly used to describe the problem in mathematical terms and to provide optimal solutions or bounds to small problem instances. As they usually do not scale well, one usually resorts to using heuristics for handling large and complex industrial problems.
Though constraint programming (CP) techniques represent the state of the art in several classical project and machine scheduling benchmarks and have been used for almost 30 years for solving industrial problems, they are still seldom considered as an alternative approach in the scheduling community. A possible reason is that, for years, in the absence of efficient and robust automatic search algorithms, CP techniques have been difficult to use for non-CP experts.
We will explain why we think this time is over and illustrate our arguments with CP Optimizer, a generic system, largely based on CP, for modeling and solving real-world scheduling problems. CP Optimizer extends linear programming with an algebraic language using simple mathematical concepts (such as intervals, sequences and functions) to capture the temporal dimension of scheduling problems in a combinatorial optimization framework. CP Optimizer implements a model-and-run paradigm that vastly reduces the burden on the user to understand CP or scheduling algorithms: modeling is by far the most important.  The automatic search combines a wide variety of techniques from Artificial Intelligence (constraint programming, temporal reasoning, learning etc.) and Operations Research (mathematical programming, graph algorithms, local search, etc.) in an exact algorithm that provides good performance out of the box and which is continuously improving.

Philippe Laborie is a Principal Scientist at IBM. He is one of the main designers of the mathematical modeling language for scheduling problems offered in CPLEX Optimization Studio and a significant contributor to the underlying automatic search algorithm. He graduated from Telecom ParisTech in 1992, and received a PhD in Artificial Intelligence from LAAS/CNRS (Toulouse) on the integration of Artificial Intelligence Planning and Scheduling in 1995.

Before joining IBM/ILOG in 1998, he worked at Electricité de France (Paris) and INRIA/IRISA (Rennes) on the Supervision and Diagnosis of complex systems (telecommunication and power distribution networks). His main scientific interests include planning, scheduling, supervision and diagnosis of complex systems and more generally, all decision problems dealing with time.

He received the 2011 ICAPS Influential paper award. Philippe is member of the editorial board of the Journal of Artificial Intelligence Research and serves in the Program Committee of many conferences in AI (IJCAI, AAAI, ECAI, ICAPS, CP, CPAIOR, ...).


Online user: 1 RSS Feed