基于改进ABC算法的有-无人协同空战行为建模  

Behavioral Modeling of Manned-unmanned Cooperative Air Combat Based on Improved ABC Algorithm

作　　者：王鹏[1] 刘昊雨 李妮[1] 于泽熙 贾尚杰 Wang Peng;Liu Haoyu;Li Ni;Yu Zexi;Jia Shangjie(School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China;Aerospace Industry Shenyang Aircraft Design and Research Institute,Shenyang 110035,China)

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100191 [2]航空工业沈阳飞机设计研究所,辽宁沈阳110035

出　　处：《系统仿真学报》2024年第12期2871-2883,共13页Journal of System Simulation

摘　　要：为解决典型有-无人机空战场景下建立协同行为模型困难,模型对抗能力较弱的问题,提出了基于混合决策方式的有-无人机行为建模框架。通过协同规则集、规则子集、战术动作集等工具在该框架下构建了支持推磨战术、单边迂回战术等5种协同战术的分层决策协同行为模型;提出了一种基于改进人工蜂群算法(artificial bee colony,ABC)的行为模型参数优化方法。利用梅森旋转法初始化种群,得到更优的初始蜜源,并改善侦察蜂阶段的探索策略。结果表明:该协同行为模型具备较强的对抗能力,利用改进ABC算法对战术动作集的参数进行优化,可以提升行为模型构建效率和对抗效果。To solve the problem of difficulty in establishing collaborative behavior models and weak adversarial capabilities in typical MAV/UAV air combat scenarios,a mixed decision based MAV/UAV behavior modeling framework is proposed.Using collaborative rule sets,rule subsets,tactical action sets,and other tools,a hierarchical decision collaborative behavior model supporting five types of collaborative tactics,including grinding tactics and unilateral flanking tactics,is constructed in this framework.a behavior model parameter optimization method based on an improved artificial bee colony(ABC) algorithm is proposed.By using the Mason rotation method to initialize the population,a better initial honey source was obtained,and the exploration strategy during the reconnaissance bee stage was improved.The results indicate that the collaborative behavior model has strong adversarial capabilities when simulated and validated in typical aerial combat scenarios.By optimizing the parameters of the tactical action set using the improved ABC algorithm,the efficiency of behavior model construction and adversarial effects can be improved.

关 键 词：行为建模 有-无人机协同 分层决策 参数优化 战术动作集 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]