Reinforcement Learning-based Trust and Reputation Model for Cluster Head Selection in Cognitive Radio Networks

This paper investigates the effectiveness of trust and reputation model (TRM) in clustering as an approach to achieve higher network performance in cognitive radio (CR) networks. Reinforcement learning (RL) based TRM has been adopted as an appropriate tool to increase the efficacy of TRM. The performance of both the traditional TRM and RL-based TRM schemes was analyzed using the probabilities of packet transmission and dropping in the network. The RL-based TRM scheme demonstrates faster detection of malicious secondary users (SUs). It has significantly shown performance stability in various environment with different malicious SUs' population in the CR networks.