Evaluation of parameters and techniques for genetic algorithm based channel allocation in Cognitive Radio Networks

Abstract

Genetic Algorithm is a promising optimization technique for solving the problem of Channel Allocation in Cognitive Radio Networks(CRNs). This work involves exploration of various parameters and techniques used in Genetic Algorithm(GA). The selection of parameters and techniques influence the run-time and ability of genetic algorithm to arrive at a globally optimal solution. Therefore, this paper validates various crossover and mutation techniques to be fit/unfit for use and their effect on convergence of genetic algorithm for optimum channel allocation strategy. Extended version of current genetic algorithm for channel allocation using partial mapped crossover(PMX) is proposed. The simulation results show that PMX crossover is a robust method for dealing with the interference problem encountered in the extended version of the genetic algorithm.