Analyzing Selfish Topology Control in Multi-Radio Multi-Channel Multi-Hop Wireless Networks

Typically, topology control is perceived as a per-node transmit power control process that achieves certain network-level objectives. We take an alternative approach of controlling the topology of a network purely by assigning channels to multiple radio interfaces on nodes. Specifically, we exploit the synergy between topology control and channel allocation to reduce the overall interference in multi-radio multi-channel wireless ad hoc networks. We formulate channel assignment as a non-cooperative game, with nodes selecting low interference channels while maintaining some degree of network connectivity. This game is shown to be a potential game, which ensures the existence of, and convergence to, a Nash Equilibrium (NE). Next, we evaluate the performance of NE topologies with respect to interference and connectivity objectives. By quantifying the impact of channel availability on interference performance, we illuminate the tradeoff between interference reduction that can be achieved by distributing interference over multiple channels and the cost of having additional channels. Finally, we study the spectral occupancy of steady state topologies, and show that despite the non-cooperative behavior, the NE topologies achieve load balancing.