Spectrum Allocation in Cognitive Radio Networks Using Evolutionary Algorithms

Abstract

One of the key objectives of evolving communication technologies is to maximize the utilization of the available spectrum by increasing the number of simultaneous users while reducing interferences among users. In cognitive radio networks, this problem is referred to as the spectrum allocation problem, and is shown to be NP-Hard. This chapter studies the use of evolutionary algorithms to solve the spectrum allocation problem in cognitive radio networks. In particular, a Binary Harmony Search Algorithm (BHSA) is proposed and used, for the first time, to solve the spectrum allocation problem. The performance of the proposed BHSA algorithm is evaluated via simulation and is compared with an optimized Genetic Algorithm (GA) under three utilization functions, namely, Mean-Reward (MR), Max-Min-Reward (MMR), and Max-Proportional-Fair (MPF). Extensive simulation results confirm that the BHSA is not only faster, but it also finds better solutions compared to those obtained by the GA. For instance, under the MMR function, the BHSA requires less than 4% of the time needed by the GA in order to find a solution that is 10% better than that obtained by the GA.