Behavioral coalition structure generation

Coalition structures are partitions of agents, i.e., collections of pair-wise disjoint coalitions, called blocks, whose union yields the entire population. Given a coalitional game, assigning a worth to each coalition, the worth of coalition structures obtains as the sum of blocks’ worth. Optimal coalition structures have maximal worth. For generic coalitional game, searching optimal coalition structures is NP-hard. Here, a behavioral coalition structure generation algorithm is presented. It specifies a selfish behavior leading agents to move across blocks depending on how coalitions allocate their worth across members. This latter issue is addressed through cooperative game theory. Given that optimality is sought through agents’ selfish behavior (rather than searched directly), the system is adaptive: if the underlying coalitional game varies over time, then whenever a change occurs the algorithm redirects agents toward new optimal coalition structures. 1. This work is partially supported by the EU within the 6th Framework Programme under contract 001907 (DELIS). 2. Department of Computer Science, University of Bologna, Mura Anteo Zamboni 7, 40126 Bologna, Italy.