Selfish users in Aloha: a game-theoretic approach

Abstract

Past studies of Aloha have emphasized system-wide goals such as achieving maximum throughput or minimum delay. We use game theory to analyze Aloha from the perspective of a selfish user. we construct an Aloha game and examine the optimal behavior of individual users. We show that the Aloha game has an equilibrium and that an Aloha system in which the users are selfish will be stable provided the attempt rate is sufficiently low. We then compare the performance of a selfish Aloha system with the performance of a centrally controlled slotted Aloha system. With some system parameters performance is near the optimum performance obtained by a centrally-controlled system. By utilizing a selfish-user assumption, it is possible to build systems which are robust and scalable.