Game-theoretic modeling and analysis of relay selection in cooperative wireless networks

In this paper, distributed single relay selection in cooperative wireless networks is modeled as a Chinese restaurant game (CRG). Specifically, the CRG is used to model strategic relay selection decisions of source nodes, taking into account negative network externality due to the potential sharing of relay nodes among source nodes. Two cases are studied as follows: (i) perfect relay transmit power (RTP) knowledge and (ii) imperfect RTP knowledge. Under the first case, a distributed relay selection algorithm is proposed and shown to converge to a Nash equilibrium grouping. Under the second case, a reinforcement learning algorithm is proposed and combined with the distributed relay selection algorithm to allow network nodes to select rate-maximizing relays. Simulation results verify the efficiency of the proposed distributed relay selection algorithm when compared with other relay selection schemes and demonstrate that it yields a network sum-rate that is comparable with that of centralized relay selection. Copyright © 2014 John Wiley & Sons, Ltd.