An Optimal Resource Allocation Method for Multi-User Multi-Relay DF Cognitive Radio Networks

This letter investigates the resource optimization problem for the multi-user multi-relay decode-forward (DF) cognitive radio networks. The novel optimization metric, namely the bandwidth-power product (BPP), is applied to achieve the optimal spectrum sharing. The formulated problem is found to be a mixed-integer programming problem. Hence, a suboptimal and tractable algorithm with low complexity called the joint optimal relay matching and greedy resource allocation (JRMRA) algorithm is proposed. The optimal relay selection is first regulated by solving the equivalent maximum weighted bipartite matching (MWBM) problem with the Hungarian method in terms of statistic expectation. Then applying the greedy-based clustering method, the resource allocation is performed to minimize the overall BPP. Simulation results show that the JRMRA algorithm achieves the acceptable performance with great reduction of computational complexity and validate the superiority of the proposed scheme compared to conventional spectrum sharing scheme using the water-filling method.