Research on Multi-Agent D2D Communication Resource Allocation Algorithm Based on A2C

Device to device (D2D) communication technology is the main component of future communication, which greatly improves utilization spectrum resources. However, in D2D subscriber multiplex network, interference between links serious and system performance degraded. Traditional resource allocation schemes need a lot channel information when dealing with problems system, have weak dynamic capability low throughput. Aiming at this challenge, paper proposes multi-agent algorithm based on Advantage Actor Critic (A2C). First, multi-D2D cellular model A2C established, then parameters actor network critic are updated, finally scheme users dynamically adaptively output. The simulation results show that compared DQN (deep Q-network) MAAC (multi-agent actor–critic), average throughput improved by 26% 12.5%, respectively.