Lecture

Multi-Agent Geo-Simulation Research
At the Cognitive Informatics Laboratory

Bernard Moulin, Full Professor, Head of the Cognitive Informatics Laboratory, Laval University, Quebec, Canada

Date: Tuesday, 2nd May 2006,
Time: 11:00 - 13:00, Mediterranean
Mediterranean Studies Seminar Room - FORTH

Host: Grigoris Antoniou

Abstract:

Geo-simulation is an approach which aims at modeling systems at the scale of individuals and entity-level units of the built environment. It provides a new way to simulate how geographic spaces can be used by their future users, particularly in urban environments. Applied to urban design, geo-simulation provides the means to study the characteristics of the urban environment by analyzing the interactions of moving agents simulating the behavior of various actors (such as pedestrians and automobiles) in a urban landscape.
Multi-agent systems are particularly adapted to the simulation of population dynamics in large-scale environments. They are well suited to the exploration of dynamic phenomena in which the interactions of individual entities can be studied at a micro-level and the emergence of behavioral patterns can be observed at a macro-level.
In this talk, we will present our research works on multi-agent geosimulation done during the past 5 years at the Cognitive Informatics Laboratory. We will present the MAGS Platform which we developed to create multi-agent geo-simulations involving several thousands of agents endowed with spatial ‘cognitive’ capabilities (perception, memory, decision making, management of agent’s needs) and evolving in virtual geographic environments (created from geo-referenced data).
We will briefly present several applications in which we used the MAGS platform for the simulation of crowd behavior in a urban environment, the spatial behavior of customers in a shopping mall, the coordination of firemen when atacking a forest fire, the simulation of the propagation of the West Nile Virus. We will also mention our current research works on the creation of large populations of agents and on multi-scale multi-agent geo-simulations.


Short bio:
Bernard Moulin is a full professor at Laval University, teaching in the Computer Science and Software Engineering Department. He is also a member of the Research Center in Geomatics at Laval University and an active researcher of GEOIDE, the Canadian Network of Centers of Excellence in Geomatics. He received his engineering degree from L'Ecole Centrale de Lyon (1976), his Master in economics from the University Lyon II (1976) and his Ph.D. in computer science from the University Lyon I (1979) (all in France).
He leads several research projects in various fields: Multi-agent geo-simulation, Design methods for multiagent systems and software-agent environments; representation of temporal and spatial knowledge in discourse; modeling and simulation of conversations between artificial agents; modeling and design approaches for knowledge-based systems and multiagent systems, as well as several projects at the intersection of geomatics and artificial intelligence. These research projects are funded by the Natural Science and Engineering Council of Canada, by the research council FQRNT from the province of Quebec, the Canadian Network of Centers of Excellence in Geomatics GEOIDE, the Defence Research Establishment Valcartier and several other organizations and private companies.