基于改进型多目标粒子群优化算法的武器-目标分配

Weapon-target Assignment with an Improved Multi-objective Particle Swarm Optimization Algorithm

在作战中武器-目标分配(WTA)问题包含众多的变量,是典型的非确定性多项式完全问题。针对毁伤效能最大和用弹量最少两个目标函数,建立了基于改进型多目标粒子群优化(MOPSO-Ⅱ)算法的WTA模型。由于粒子群优化算法存在"维数灾难"瓶颈,应用了变量随机分解策略和合作协同进化框架,按照带精英策略的非支配排序遗传(NSGA-Ⅱ)算法中的排序方法对粒子群编码数据进行非支配排序。通过实例仿真分析,结果表明MOPSO-Ⅱ算法比NSGA-Ⅱ算法具有更好的求解精度与运行效率,能够获得满意的分配结果,且计算快速有效,比较适合较大规模的WTA问题实时求解。

Weapon-target assignment（ WTA） with numerous variables in modern campaign is a typical non-deterministic polynomial（ NP） complete problem. An optimization model based on improved multiobjective swarm optimization algorithm（ MOPSO-II） is established to solve the objective functions of maximum damage probability and minimum ammunition consumption. Since ＂curse of dimensionality＂occurs in the objective swarm optimization algorithm（ PSO）,the random variable decomposition strategy and cooperative co-evolution evolutionary frame are used for variable decomposition,and also all swarms are composited by using the non-dominated set algorithm in NSGA-II. The simulated results show that MOPSO-II is quicker and more effective than NSGA-II,and can give good WTA quickly,especially when the scale of WTA problem is large.