Cooperative spectrum allocation algorithm based on reinforcement learning

In order to solve the problem of spectrum allocation in cognitive radio networks, we propose a cooperative reinforcement learning spectrum allocation algorithm based on user experience quality. which simulates the secondary users in the cognitive network as agents in the reinforcement learning, and introduces a cooperation mechanism between the secondary users. New users can absorb and learn from the reinforcement learning experience of other users, and obtain the best spectrum allocation plan at a faster speed. In addition, the price game factor between the primary user and the secondary user is introduced in the spectrum allocation process, allowing the primary user to price the authorized spectrum according to their own situation, and the impact of different spectrum prices on the income of the secondary user is studied, making the algorithm closer to the real scene. In terms of system evaluation, the average opinion score model is used to visually display the service quality of system users. Simulation results show that the algorithm can effectively improve user service quality and system communication performance, and provides an effective solution for understanding the spectrum allocation among users.