Indian Journal of Science and Technology

Abstract

Cognitive Radio is a forthcoming technique to improvise the utilization of radio frequency spectrum in wireless network. However, Cognitive Radio Network has several challenges such as channel selection, efficient spectrum sharing, network throughput, etc. This paper presents a resource (channel) selection strategy by the Secondary Users (SUs) in a dynamic environment via game theoretic approach namely potential games. A distributed, Stochastic Learning based Resource Selection (SLRS) and Q-Learning based Resource Selection (QLRS) are the two different algorithms proposed here. The strategy followed by the SU is based on its own action-reward history, even without knowing the actions in other SUs. The simulation results prove that the QLRS algorithm achieves higher throughput and fairness performances than the SLRS algorithm.
Keywords: Cognitive Radio, Channel Selection, Q-Learning, Stochastic Learning, Spectrum Sharing