Q-learning-based Power Allocation for Secure VFD NOMA Net-work

This study conducts an in-depth investigation of physical layer security and resource allocation issues in non-orthogonal multiple access (NOMA) communication systems by introducing a virtual full-duplex (VFD) relays scheme. The scheme includes an untrusted amplify-and-forward (AF) half-duplex (HD) relay and trusted decode-and-forward (DF) HD relay to enhance the spectrum efficiency of the communication system. To address potential eavesdropping by an untrusted relay, a simple practical cooperative jamming scheme is designed to confuse the untrusted relay, thereby enhancing the communication security. In terms of resource management, this study employs a Q-learning algorithm for the joint optimization of relay selection and power allocation in the network. This method allows the system to adjust its strategies flexibly in dynamic network environments to maximize communication performance. The simulation results demonstrate the superiority of the proposed Q-learning-based relay selection and power allocation scheme, showing substantial performance gains compared to traditional methods.