An experimental study of Genetic Algorithm for spectrum optimization in Cognitive Radio Networks

The rapid evolution of new wireless technologies and deployment of wireless devices in recent years is resulting in spectrum overcrowding [1]. It is observed that large amount of spectrum is under-utilized causing spectral crisis [1]. Cognitive Radio (CR) has emerged as a cutting edge technology to solve the spectrum management problem and simultaneously meet the QoS requirements of the wireless applications. CRNs allow the secondary user to identify the best available spectrum from the surrounding environment and reconfigure according to the sensed information in order to achieve the optimal performance [2]. To address this problem, we reimplemented Genetic Algorithm (GA) where each spectral band is a chromosome with corresponding Radio frequency parameters as its genes [3]. In this paper, we implemented different selection operators of GA with varied crossover and mutation probabilities to study the behavior of GA on this problem. GA with Roulette wheel selection and exhaustive search is observed to be superior to other variants.