Centralized channel and power allocation for cognitive radio networks: A Q-learning solution

Cognitive radio has been proposed as a novel approach for improving the utilization of the limited radio resources by dynamically changing its operating parameters. This paper deals with the problem of channel and power allocation for cognitive radio networks. In particular, we consider the scenario where the transmission of secondary users is controlled by cognitive base station. We propose an autonomic approach to solve the problem through a form of real-time reinforcement learning known as Q-learning. The secondary users being served and their transmission power on each channel constitute the dynamic environment. Through the “trial-and-error” interaction with its radio environment, the cognitive base station gradually converges to the optimal channel and power allocation policy in a centralized way. Numerical simulation results show that the proposed algorithm can not only realizes the autonomy of channel and power allocation, but also improves system throughput compared to other algorithms.