MAGéo, une plateforme de modélisation et de simulation multi-agent pour les sciences humaines

Abstract

This paper presents MAGeo (Modeling Agent in Geography), a user-friendly agent-based simulation tool dedicated to spatialized and multilevel researches. Model development with MAGeo does not require any particular programming skills. The conceptual model is realized through graphical objects, such as boxes and links to organize them. Model's ontology is based on the following three entities: Agent - Organization - Behavior that allow the building of multilevel model. The modeler has access to the simulation window in order to develop its model's dynamic. Simulation model is built using cursors and other graphical tools to parametrize the model, as well as maps and graphs to explore the model's behavior. MAGeo is readily accessible to social scientists which aim to develop spatialized and complex models but don’t possess programming skill. Because learning MAGeo and developing model is not time-consuming, MAGeo is also well designed for teaching complex system. Finally, its graphical approach produces perfect systemic and multilevel graphics for communication in multidisciplinary teams.

Keywords

• complex system
• multi-agent system
• agent-based model
• ontology
• simulation platform