A fuzzy evolutionary approach for collaborative clustering in multi-agent systems with application to emergent virtual organizations

An approach to dynamic collaborative clustering in multi-agent systems is proposed. The approach determines the optimal configuration of an emergent virtual organization by clustering the best partners for each activity contributing towards the organizations' goal. The approach consists of two parts: 1) in a dynamic, ever expanding agent domain it performs an open evolutionary search for the best possible partners to be added to the organization; and 2) it uses fuzzy entropy minimization to continuously maintain the optimal configuration of the partnering agents in terms of their interactions to best fulfill the global goal of the organization.