Frequency Allocation in Dynamic Environment of Cognitive Radio Networks Based on Stochastic Game

Abstract

We investigate distributed frequency allocation problem in dynamic environment of cognitive radio (CR) networks with a stochastic game (SG) model. Traditional multi-agent reinforcement learning (MARL) algorithms, as the primary online strategy learning algorithms of SG game, are not suitable for the application of wireless communication. Hence, a new MARL algorithm, maximizing the average Q function algorithm (MAQ), is proposed in this article. MAQ not only reduces the intercommunications but also realizes indirect coordination among agents. Simulation results show the learning efficiency of MAQ which is close to that of centric learning method.