Swarm Game Decision Making Method Based on Deep Reinforcement Learning

Aiming at the decision making problem of unmanned swarm system in non-cooperative environment, this paper proposes an intelligent decision making algorithm combining deep reinforcement learning and improved reward allocation method, which can quickly respond to changes of complex game environment and opposite agents, and generate dynamic intelligent strategies for unmanned swarms. The simulation results show that the intelligent decision-making algorithm can not only successfully complete the task of the unmanned swarm in the face of the opposite intelligent swarm non-cooperative behavior, but also alleviate the problem of sparse reward and difficult convergence of the deep reinforcement learning decision-making algorithm, the intelligent decision-making algorithm has great performance.