Unmanned aerial vehicle confrontation game training control method based on reinforcement learning

The invention discloses an unmanned aerial vehicle confrontation game training control method based on reinforcement learning. The method comprises the following steps that: a main agent is made to fight with all opponents in an opponent pool, and the winning rate of the main agent is calculated; whether the winning rate of the main agent meets a preset requirement or not is judged; if the winning rate of the main agent meets the preset requirement, an confronting opponent is selected according to the winning rate of the main agent; and the main agent and the confronting opponent are trained until the strategy of the main agent converges. The unmanned aerial vehicle confrontation game training control method based on the reinforcement learning is effective and can make the main agent have higher learning ability. 本申请公开了一种基于强化学习的无人机对抗博弈训练控制方法，包括如下步骤：使主智能体与对手池中所有对手对战并统计所述主智能体的胜率；判断所述主智能体的胜率是否满足预设要求；如果所述主智能体的胜率满足预设要求，则根据所述主智能体的胜率选择对抗对手；使主智能体与所述对抗对手训练直至所述主智能体策略收敛。本申请的有益之处在于。本申请的有益之处在于提供了一种行之有效的基于强化学习的无人机对抗博弈训练控制方