Agent Group Workshop

Date: September the 27th, 2013

Planning

  • 09:00-09:30: Keynote talk of Cedric Boittin
  • 09:30-12:00: Discussions
  • 12:00-14:00: Lunch Break
  • 14:00-14:30: Keynote talk of David Meignan
  • 14:30-16:00: Discussions
  • 16:00-16:30: Keynote talk of Sebastian Rodriguez
  • 16:30-18:00: Discussions



Keynote talk of Cedric Boittin

Calibration/Validation of LUTI models and associated optimization problems

The management of transportation facilities and the evolution of the urban land-use has been at all time a major stake in the cities’ development. In the last century, many studies tackled these two aspects, and emphasis has been set on the relationships between them. It is now of common knowledge that land-use evolution induces a change in the transportation needs : the separation of human activities creates the need for travel and goods transport. On the other hand, history is full of examples of the reverse impact of major transportation changes on the land-use, when important trade routes made and unmade the wealth of whole regions. This simple explanation model quickly spread and has been massively used by urban planners ever since, and is still in use as of today. It has been enriched with knowledge from various domains such as economics or social theories, but the core idea remains. Several distinct Land-Use Transport Integrated (LUTI) models have emerged from these additions. The aim of such models is to obtain the best possible representation of the influence of a modification in the transport infrastructure on the land-use, and reciprocally. Therefore, one of the objectives of the LUTI simulation is to run a proper prediction with a 20-30 year horizon on the evolution of the urban system based on assumptions made in the considered scenarios or policies under study or comparison. With the ever-growing globalization comes the need for urban planners to more precisely control the evolution of cities, and thus of more accurate and reliable forecasts, with a broader horizon. The simulation of LUTI models as a forecasting tool poses a twofold problem : how to efficiently calibrate these models, and how to measure and ensure the correctness of the results. The state of the art in simulation calibration indicates a strong correlation with optimization problems.


Keynote talk of David Meignan

Interactive optimization

Interactive optimization is the combination of combinatorial optimization methods with expertize of a human operator or user. Over the years, Operation Research has produced numbers of successful computational approaches and software tools for solving hard combinatorial optimization problems of practical significance. However, in several application domains there is still a large gap between research and application of advanced optimization methods, in particular for decision support tools. Interactive optimization is an active and promising paradigm to fill this gap between research and application. In this presentation I will give an overview of the project “Interactive metaheuristics for optimization-based decision support systems” recently funded by the DFG (Deutsche Forschungsgemeinschaft). I will first give a general introduction to interactive optimization with a quick overview of existing interactive optimization methods (interactive multiobjective optimization, interactive evolutionary computation and human-guided search). In a second part, I will present the proposed methods and results of a preliminary study, and will discuss the recent trends and perspectives on interactive optimization.


Speaker

Dr. David Meignan, Universität Osnabrück (Osnabrück, Germany) David Meignan received his PhD in Computer Science from the University of Belfort-Montbéliard (Belfort, France). From 2009 to 2011 he held a postdoctoral position at École Polytechnique de Montréal (Montréal, Canada). Since 2011 he work as a researcher at the University of Osnabrück (Osnabrück, Germany) where he obtained in April 2012 a research grant from Google for a project on interactive optimization. His current project “Interactive metaheuristics for optimization-based decision support systems” has been recently funded by the DFG (Deutsche Forschungsgemeinschaft) for two years. His research interests include heuristic and metaheuristics methods for solving hard combinatorial optimization problems, and distributed problem solving.



Keynote talk of Sebastian Rodriguez

Speaker

Sebastian Rodriguez is a Full Professor of the Department of Computer Science, National Technology University (NTU), Argentina. He is also the founder and Head of the Advanced Technology Research Center of Tucumán, Argentina, and an associate researcher of the Systems and Transportation Laboratory at the University of Technology of Belfort-Montbéliard (UTBM), France. He received a Computer Engineer degree for the National University of Tucumán, Argentina, a M.S. degree in computer science from the University of Franche-Comté and a Ph.D. degree in computer science of the UTBM.

Retrieved from "http://multiagent.fr/TW0913"
This page was last modified on 23 September 2013, at 14:38. This page has been accessed 10,524 times.
Copyright 2010-2018 © Laboratoire Électronique Informatique Image - Université Bourgogne Franche-Comté - Privacy policy