Optimal decentralized sensing-orders in multi-user cognitive radio networks

This paper investigates the optimal sensing-order problem for the general case of multi-user and multi-channel cognitive radio networks in which channel sensing is performed by the individual cognitive users (CUs) in a decentralized manner. The aim is to determine a sensing-order for each CU so as to maximize the cognitive throughput, which is the expected sum-throughput of all CUs. We consider two broad multi-user scenarios - first, where the network uses a contention resolution strategy whenever two or more CUs try to access the same channel simultaneously, and second, where the CUs are organized into distinct preemptive priority levels. For each scenario, we present a procedure for computing the cognitive throughput in the network. In general, the optimal sensing-orders can then be determined via a computationally expensive brute-force search. We also propose a suboptimal low complexity search algorithm for each scenario and show their proximity to the optimal performance by means of simulations.