Résumé en anglais

This paper proposes a participatory and cooperative approach to acquire necessary knowledge to build a MultiAgent-Based Simulation (MABS). This approach is based on role-playing, meeting and Computer Supported Cooperative Work (CSCW) principles. During a meeting firm actors simulate problem solving processes by play- ing their usual roles. The result of several meetings is a corpus which is analyzed to provide a set of scenarios. It allows the building of a multiactor model. In this study, we apply the approach to acquire knowledge to model and then simulate decision-making processes in poultry firms. In particular, we seek to understand the probable impacts of individual behaviors in the decision process of managing raw material.