地对空武器—目标分配的多目标决策问题研究  

Research on Multi-target Decision-making of Ground-to-air Weapon-target Assignment

作　　者：许可[1] 吉兰萍 孙文娟[1] 雷鸣[1] 高宏宇 叶彩霞 XU Ke;JI Lanping;SUN Wenjuan;LEI Ming;GAO Hongyu;YE Caixia(Shenyang Ligong University,Shenyang 110159,China)

机构地区：[1]沈阳理工大学理学院,沈阳110159 [2]沈阳理工大学自动化与电气工程学院,沈阳110159

出　　处：《沈阳理工大学学报》2022年第5期13-20,共8页Journal of Shenyang Ligong University

基　　金：辽宁省教育厅高等学校基本科研项目(LJKQZ2021057,LJKZ0260);辽宁省“百千万人才工程”资助项目(2021921089)。

摘　　要：地对空信息化作战中,合理高效的武器-目标分配方案对于降低我方损失、提高防御效能具有重要意义。本文以敌方所有来袭目标的剩余生存概率最小、我方弹药消耗量的价值最小、敌方对我方资产的损害最小为优化目标建立地对空武器-目标分配模型;设计一种改进第二代快速非支配排序遗传算法(NSGA-Ⅱ)对模型进行求解,针对两个种群进行不同的遗传操作,利用强化学习方法动态调整两个种群中个体迁徙参数,以保证种群的多样性;对获得的武器-目标分配模型的非劣解集(Pareto解),使用模糊决策进行评价,得出最优的分配决策方案。In ground-to-air information warfare,a reasonable and efficient weapon-target allocation scheme is of great significance to reduce our loss and improve defense efficiency.In this paper,a ground-to-air weapon-target allocation model is established to minimize the residual survival probability of all enemy attack targets,our ammunition consumption and the enemy′s damage to our assets.An improved second generation fast Non-dominated sequencing genetic algorithm(NSGA-Ⅱ)is designed to solve the model.Different genetic operations are carried out for the two populations,and reinforcement learning method is used to dynamically adjust individual migration parameters in the two populations to ensure the diversity of the populations.The non-inferior solution set(Pareto solution)of weapon-target assignment model is evaluated by fuzzy decision,and the optimal assignment decision scheme is obtained.

关 键 词：武器-目标分配 快速非支配排序 强化学习 模糊决策 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]