基于MDP的战机对抗导弹措施优化方法  

MDP method for optimization of fighter aircraft's countermeasures against missile

作　　者：宋海方[1,2] 肖明清[1] 陈游[1,2] 胡阳光[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安710038 [2]电子信息控制国防重点实验室,成都610036

出　　处：《北京航空航天大学学报》2017年第5期942-950,共9页Journal of Beijing University of Aeronautics and Astronautics

摘　　要：对战机对抗导弹的措施优化问题进行了研究,将该问题归结为不确定条件下的序贯决策问题,提出了一种基于Markov决策过程(MDP)的导弹对抗措施优化方法。首先,该方法根据作战过程将作战阶段分为不同的状态,将电子对抗措施(ECM)和战术机动作为可供战机选择的行动;然后,通过不同战术行动下状态之间的转移概率来反映行动的不确定性,以雷达搜索状态和导弹命中状态的期望值来表征不同策略下的飞机生存力;最后,通过建立Markov决策模型得到飞机生存力最大时各个状态对应的最佳行动。仿真结果表明:飞机生存力随着时间推移而逐渐降低;不同策略下飞机的生存力不同,基于MDP的策略可以有效提高飞机的生存力;单步优化措施不能提高战机的生存力,必须考虑长期状态转移的影响。Missile countermeasure optimization problem for fighter aircraft was studied. The missile countermeasure optimization was modeled as a sequential decision-making problem under uncertainty, and a Markov decision process （MDP） based approach was proposed. First, the engagement process was divided into several phases, and these phrases were marked by states. Electronic counter measures （ECM） and strategic maneuvers were treated as actions. Then, the state transition probability was used to reflect the uncertainty of each action, and average occupancy of ＂hit＂ state and ＂search＂ state was used to evaluate the aircraft survivability under different policies. Finally, the policy iteration algorithm was used to get the optimal policy, which maps the optimal action to be taken in each state. Simulation indicates that the aircraft survivability decreases as time goes on; the proposed MDP-based approach can effectively improve the aircraft survivability; one-step optimization is useless for the aircraft survivability and the influence of the state transition should be considered from the view of long-time horizon.

关 键 词：飞机生存力 电子对抗措施(ECM) 战术机动 Markov决策过程(MDP) 导弹对抗 序贯决策 

分 类 号：TN972[电子电信—信号与信息处理] TJ765[电子电信—信息与通信工程]