Computation of equilibria for coalition structures arising from social contexts in resource selection games

In game theory, the centerpiece of analysis is the notion of equilibrium. A branch of game theory (called coalitional game theory) uses equilibrium notions that are defined under the assumption that agents can form coalitions between themselves. In a strategic form game, a strategy profile is a (coalitional) equilibrium if no viable coalition of agents benefits from jointly changing their strategies. Weaker or stronger equilibrium notions can be defined by considering various restrictions on formation of coalitions. In a Nash equilibrium, for instance, the assumption is that viable coalitions are simply singletons. In a strong Nash equilibrium, which lies at the other extreme, every possible coalition (i.e., every non-empty subset of agents) is viable. However, deeming every coalition viable is too demanding for most games that are worth to study; and hence, a strong Nash equilibrium rarely exists in those games. On the other hand, in a partition equilibrium, viable coaliations are restricted to partitions of agents, which is much less demanding then strong Nash equilibrium (since notice that the number of viable coalitions is Θ(n) in the case of a partition equilibrium, and it is 2^n-1 in the case of a strong Nash equilibrium). For example, a partition equilibrium always exists in resource selection games, whereas a strong Nash equilibrium does not exist even in most simple instances of these games. More sophisticated restrictions on coalition formation can be justified by communicational, coordinational or institutional constraints. In this thesis, inspired by social structures in various real-life scenarios, we introduce certain restrictions on coalition formation, and on their basis, we introduce various equilibrium notions. We study our equilibrium notions also in resource selection games, and we present a complete set of existence and nonexistence results for general reseource selection games and their important special cases. We also provide efficient algorithms to compute an equilibrium for the cases where one exists.