基于双层自适应MOEA/D的多阶段武器目标分配方法  

Multi-stage weapon target assignment method based on two-layer adaptive MOEA/D

作　　者：王洲强 陈万春[1] 陈中原 Wang Zhouqiang;Chen Wanchun;Chen Zhongyuan(School of Astronautics,Beihang University,Beijing 100191,China)

机构地区：[1]北京航空航天大学宇航学院,北京100191

出　　处：《战术导弹技术》2022年第4期113-123,共11页Tactical Missile Technology

基　　金：中国博士后科学基金资助项目(2021M700321)。

摘　　要：针对多阶段武器目标分配问题,建立了多阶段多目标约束组合优化问题模型。在考虑资源约束、可行性约束的前提下,寻求对敌方目标造成最大伤害的同时,资源损失最少。在MOEA/D算法(multi-objective evolutionary algorithm based on decomposition)基础上,引入了差分进化算子,并提出了启发式初始化种群机制,以避免算法陷入局部收敛并加快收敛速度。采用随机修复策略对产生的不可行解进行修复,引入双层自适应机制,以种群进化状态来提供基准值差分进化因子及邻域数量,并以每个个体进化状态提供修正项,以进一步提高算法搜索能力并加快收敛。实验仿真表明,所提优化方法具有较高的收敛性且解集分布更均匀,能够有效地解决多阶段武器目标分配优化问题。Aiming at the multi-stage weapon target assignment,a combinatorial optimization model of multi-stage and multi-objective constraints is established.On the premise of resource constraints and feasibility constraints,it seeks to cause maximum damage to enemy targets while minimizing resource loss.Based on the MOEA/D algorithm,a differential evolution operator is introduced,and a heuristic initialization population mechanism is proposed to avoid local convergence.The random repair strategy is used to repair the infeasible solutions,and a two-level adaptive mechanism is introduced to provide the base value differential evolution factor and the number of neighborhoods based on the population evolution state,and provide the correction term based on each individual evolution state,so as to further improve the algorithm′s searching ability and speed up the convergence.Experimental simulation shows that the proposed algorithm has higher convergence and more uniform solution set distribution,which can effectively solve the multi-stage weapon target allocation optimization problem.

关 键 词：多阶段武器目标分配 多目标 MOEA/D 双层自适应 动态邻域 

分 类 号：TJ765.3[兵器科学与技术—武器系统与运用工程]