Distributed multichannel random access networks with selfish users

Dynamic spectrum access (DSA) schemes allow the users to share spectrum resources by taking advantage of the variations in spectrum demand over time and space. Carrying out dynamic spectrum allocation centrally, however, can be a complex task. For this reason, distributed schemes in which users can access the available channels independently may be preferable to centralized DSA schemes. Cognitive radio systems, which enable user terminals to sense their environment and form their action accordingly, are particularly well-suited for distributed systems. On the other hand, the freedom in distributed schemes gives the users the option to act selfishly, which has decisive effects on system performance. In this paper we consider a distributed multichannel wireless random access system where users selfishly access the channels in the system. We analyze the behavior of the selfish users by modeling the system as a non-cooperative game and we identify all stable operating points (Nash equilibria) of this game. We then compare the performance of this system with a number of cooperative distributed DSA schemes in terms of user utilities. Our results show that the performance of the selfish multichannel random access system can be comparable to cooperative schemes.