Robust routing and channel allocation in multi-hop cognitive radio networks

Jointly consider routing and spectrum selection is essentially and necessary in multi-hop cognitive radio networks. System cost and throughput are commonly used to evaluate performance of routing and spectrum selection schemes. Traditional methods mostly translate these metrics into a single objective function, and corresponding weights are allocated to each metric representing impact on the entire network performance. Optimal solutions of these approaches are sensitive to the weight settings which are usually hard to appropriately chosen. In this work, the task of routing and channel allocation is modeled as a two-objective optimization problem. Two conflicting metric functions system total throughput and total cost are optimized simultaneously, and a novel memetic algorithm which adopts a new neighborhood search procedure is proposed to solve this problem. Incorporated with robustness consideration on routing, a new robustness metric is also presented to work as a decision mechanism to ensure the robustness of the entire network. The aim of this task is to find the best compromise routing and channel allocation scheme on system throughput, cost and robustness among the feasible solution set. Simulation results demonstrate that the optimal solution set obtained by the memetic algorithm can clearly show the conflicting relationship of the system cost and throughput when choosing different routing and channel selection schemes. The best solution made by the additional robustness metric among these solutions can achieve the best performance of the cognitive radio network.