Apports de la simulation multi-agents à l’étude des processus de diffusion

This paper presents the main contributions of an agent-based approach for modelling and simulating the processes of diffusion in geographical space. This approach is compared with the macro modelling traditionally employed to describe such phenomenons. The logistic model is thus presented as representative of this category of model. The agent-based models represent a possible alternative to a macroscopic approach of the processes of diffusion. Related to the theories of self-organization, it is assumed that the phenomena observed on a certain level are the result of the many interactions which occur on one or more lower levels. As a simulation tool predisposed to the analysis of such a dynamic, the multi-agent systems are briefly presented. This paper ends with a brief example of application, the modelling and the simulation of the diffusion of an agricultural innovation model (Hägerstrand's model revisited)