基于蒙特卡洛树搜索方法的空战机动决策

Air Combat Maneuver Decision Based on MCTS Method

针对空战中双方的机动决策问题,考虑角度、距离、速度变化以及高度优势函数,以赢取最大空战优势期望为目标,建立了空战双方机动决策模型,并引入蒙特卡洛树搜索方法进行求解,搜索策略采取极大极小搜索算法。针对双机对抗,通过仿真给出空战态势及空战优势变化曲线,结果表明:我机在整个空战过程中绝大部分是处于优势地位的,验证了使用蒙特卡洛树搜索方法进行空战机动决策的可行性,同时对算法的实时性进行了仿真,计算机动动作所用时间众数位于50~100ms区间内,达到了实时性要求。

Aimed at both sides＇ maneuver decision problems in air combat that changes of angle, distance, velocity and altitude dominant function are considered to make maximum air combat superiority expectation as target, the Monte Carlo tree search method is introduced. The search strategy is used by the mini-max search algorithm. Against the confrontation between the two aircraft, the simulation results of air combat situation and advantage change curves show that our fighter is in the dominant position for the most part in the whole process in air combat by using MCTS method. At the same time, the real-time performance of the algorithm is simulated. The results show that the model time of the algorithm computing maneuver action is located at 50-100 ms range to meet the needs of real-time.