UAV Cooperative Air Combat Maneuvering Decision-Making Using GRU-MAPPO

In this article, a GRU-Multi-agent Proximal Policy Optimization (GRU-MAPPO) algorithm was proposed to address unmanned aerial vehicle (UAV) cooperative air combat decision-making problem. This algorithm adds a layer of GRU to the Actor-Critic network framework, uses update gate to extract the historical temporal information and enhance situational awareness. Finally, experiments in our constructed UAV cooperative air combat environment demonstrate that UAVs using the algorithm proposed in this article can learn effective strategies in air combat environments and achieve high win rates.