Résumé en anglais

This article presents a coalition formation method for multi-agent systems that finds a Pareto optimal solution without aggregating the preferences of the agents. This protocol is adapted to problems requiring coordination by coalition formation, where it is undesirable, or not possible, to aggregate the preferences of the agents. This method has been extended to enable dynamic restructuring of coalitions when changes occur in the system. Then it presents the behaviours of the agents teaching schedule at university using the utility function of the professors and the students. To evaluate these methods, we have implemented this application and presented the obtained results.