A Centralized Multi-User Anti-Composite Intelligent Interference Algorithm Based on Improved Q-Learning

This paper proposes a central anti-jamming algorithm (CAJA) based on improved Q-learning to further solve the communication challenges faced by multi-user wireless communication networks in terms of external complex malicious interference. This will also reduce the dual factors restricting wireless communication quality, the impact of inter-user interference within the network, and the effect of external malicious interference on the communication system to improve multi-user wireless communication transmission. Firstly, a central base station that coordinates and allocates channels for users within the network is set up using multi-user wireless communication network architecture to constitute a centralized wireless communication network. Secondly, the multi-user system is modeled using the single-user Markov decision process in which the central base station is the main body. Finally, an improved Q-learning algorithm is used to improve overall system transmission income using the central base station, based on the network user number sequential decision action for avoiding external malicious interference. It is designed to avoid the impact of internal network interference on transmission performance during the early stage of communication, achieving overall system transmission income improvement. Simulation results show that in comparison to the existing multi-user independent Q-learning anti-jamming algorithm and the traditional orthogonal frequency-hopping scheme, the proposed algorithm significantly improves overall system transmission performance.