Des données géographiques à la simulation à base d’agents : application de la plate-forme GAMA

Abstract

These last years, agent-based simulation has shown its use for geography. However, if nowadays there are many platforms allowing the development of models, these ones are often too complex to be used by non-computer scientist (e.g. Repast Symphony) or too limited to develop rich models integrating numerous agents and data (e.g. Netlogo). The GAMA platform aims at providing modelers a complete platform that is at the same time very simple to use (thanks to a dedicated modeling language) and powerful. In this article, we present the advanced capabilities of the last version of GAMA concerning the management of space. Indeed, GAMA allows to simultaneously manage several environments based on three types of topology (continuous, grid, graph), to easily integrate geographic data (vector and raster) and to represent geometries in a 3D environment. All the capabilities detailed in this article are illustrated by a model on urban segregation.

Keywords

• 3-D GIS
• agent-based model
• geographical data
• modeling/modelling platform