基于蒙特卡洛树搜索的机载雷达编队策略优化

Optimal Formation Strategy for Airborne Radar Detection Based on Monte Carlo Tree Search

多机雷达协同探测是飞机协同作战研究的重要内容,编队策略影响着多机雷达协同探测的探测效能。针对机载雷达编队策略的优化,基于部分可观察马尔可夫决策过程(POMDP)模型描述了编队策略优化问题,提出了一种基于蒙特卡洛树搜索(MCTS)的最优策略求解方法。以对敌方目标的探测性能为指标,将该方法与多种其他编队方法的结果进行了比较。试验结果表明,该方法很好地平衡了MCTS中利用与探索的关系,可以在规模较大的搜索空间内有效完成编队策略优化,在编队探测能力提升方面展示出了比其它方法更加优越的性能。

The cooperative detection of multi-aircraft radar is an important part of the cooperative combat research of aircraft,and the formation strategy affects the detection efficiency of multi-aircraft radar.Aiming at the optimization of the formation strategy of airborne radar,the formation strategy optimization problem is described based on the partially observable Markov decision process model,and then an optimal strategy solution based on Monte Carlo tree search is presented.Finally,the target detection performance of the enemy target is used to compare the results with the other methods.The experimental results show that this method can effectively optimize the formation strategy in large search space and show better performance than other methods in improving the detection capability.