基于聚类的多目标演化算法在航迹规划中的应用研究  

Application study of a clustering-based multiobjective evolutionary algorithm in path planning

作　　者：王立晶[1] 李桂英[1] 李欣 WANG Li-Jing;LI Gui-Ying;LI Xin(School of Mechanical & Electrical Engineering, Heilongjiang University, Harbin 150080, China;Centre for Control Theory andGuidance Technology, Harbin Institute of Technology, Harbin 150001, China)

机构地区：[1]黑龙江大学机电工程学院,哈尔滨15080 [2]哈尔滨工业大学控制理论与制导技术研究中心,哈尔滨150001

出　　处：《黑龙江大学工程学报》2019年第2期77-83,共7页Journal of Engineering of Heilongjiang University

基　　金：哈尔滨市科技创新人才研究专项资金项目(2017RAQXJ133,2017RAQXJ137)

摘　　要：巡航导弹的航迹规划问题是一类复杂的多目标优化问题,利用多目标演化算法可以很好地求解此类问题。考虑到聚类算法的数据挖掘功能可以提高多目标演化算法的局部搜索能力,并能恰当地平衡搜索过程中的开采与勘探以获得均匀分布的逼近前沿,提出了一种基于聚类的多目标演化算法(HCEA)。HCEA算法利用层次聚类算法挖掘种群分布信息,然后利用配对控制概率平衡全局搜索与局部搜索,并在局部搜索与全局搜索中分别采用不同的差分系数,使算法的搜索能力进一步加强。实验结果表明HCEA能够有效地求解巡航导弹的航迹规划问题。The path planning problem of the cruise missile is a kind of complicated multiobjective optimization problem while the multiobjective evolutionary algorithms can be used to well solve this kind of problem. Considering that the data mining ability of the clustering algorithm can improve the local search ability of the multiobjective evolutionary algorithm, and appropriately balancing the exploitation and exploration during the search process helps to obtain evenly distributed approximation front, A clustering-based multiobjective evolutionary algorithm (HCEA) is proposed. HCEA utilizes the hierarchical clustering algorithm to excavate the solution distribution information and then uses the mating restriction probability to balance the global search and the local search. Among them, the values of difference coefficient are different in the local search and the global search to further improve the search capability. The experimental results show that HCEA can effectively solve the path planning problem of the cruise missile.

关 键 词：多目标优化 演化算法 聚类 导弹航迹规划 

分 类 号：TH131[机械工程—机械制造及自动化]