Minority game for cognitive radios: Cooperating without cooperation

Abstract

In this paper we report on our preliminary results and recent work on building cognitive wireless networks that achieve resource sharing in a local environment. We emphasize two major issues. First, the cross-layer optimization at the network level requires exchange of information between OSI-layers in the terminal and often among the nodes that form the network. Second, the cooperative behavior among the radios is often believed to require a rich exchange of information. We show in this paper that cooperation can be an emergent phenomenon without any complex signalling. We apply Minority Games to cognitive wireless networks to show that resource sharing can be achieved without detailed information exchange or coordination between strategies. We further argue that Minority Games are not only a useful analysis tool, but a potentially efficient method to develop actual resource sharing algorithms. We conclude the paper by pointing out that also other swarm intelligence type of solutions could be applied to cognitive radio communications.