Simulation paysagère : un modèle multi-agents pour prendre en compte les relations sociales

Abstract

Landscape simulation can be a good mean to understand the observed evolutions and then identify the factors generating the changes. We have worked on the evolution of the land use in a "commune" in the chain of the Puys, by using a model dealing with the ecological growth of vegetation that agricultural practices hinder.The discrete event simulations enable to study long-term evolutions and take space aspects into account, but farm modelling is still simplistic. We have then developed a multi-agent model that enables us to take into account the interactions between the farms subjected to competition and synergy. This work corroborates that a multi-agent modelling is quite adapted to study the influence of social relationship on the territory and then on the landscape.In this presentation, we also show how a multi-agent model can be integrated to discrete event simulations.

Keywords

• simulation
• discrete events
• social interaction
• agent-based model
• landscape modeling/modelling