Cooperation in multi-access networks via coalitional game theory

Abstract

We study the performance of cooperation strategies in multiple-access networks in the framework of coalitional game theory to determine how stable coalitions arise and evolve in response to the potential throughput gains enabled by cooperation. Two strategies are examined, both inspired by practical scenarios. The first strategy is representative of a cooperative random access system. Each user randomly alternates between two states, indicating its desire to transmit. Users cooperate to avoid interfering transmissions by sharing knowledge of their state with members of the same coalition. A scheduler determines the active user within each coalition that can access the channel. Collisions occur when users belonging to different coalitions transmit simultaneously. In this case, the grand coalition formed by all users is both sum-rate optimal and stable, in the sense that users do not have any incentive to leave the coalition. The second strategy is representative of a cooperative token based system. Users are statically scheduled in a round robin fashion. They cooperate to avoid wasteful idle cycles by sharing their right to access the channel with members of the same coalition. A scheduler selects the user that transmits among the ones within the coalition. In this case, the grand coalition is sum rate optimal but cannot always be stabilized because some group of users may always have an incentive to deviate.