引用本文:吴亚丽,付玉龙,李国婷,张亚崇.高维多目标头脑风暴优化算法[J].控制理论与应用,2020,37(1):193~204.[点击复制]
WU yali,FU Yu-long,LI Guo-ting,ZHANG Ya-chong.Many-objective brain storm optimization algorithm[J].Control Theory and Technology,2020,37(1):193~204.[点击复制]
高维多目标头脑风暴优化算法
Many-objective brain storm optimization algorithm
摘要点击 2687  全文点击 961  投稿时间：2018-11-19  修订日期：2019-04-23
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2019.80898
  2020,37(1):193-204
中文关键词  头脑风暴优化算法  聚类  决策变量聚类  分解策略  参考点
英文关键词  brain storm optimization (BSO)  cluster  decision variable clustering  decomposition strategy  reference points
基金项目  国家青年基金
作者单位E-mail
吴亚丽* 西安理工大学 yliwu@xaut.edu.cn 
付玉龙 西安理工大学  
李国婷 西安理工大学  
张亚崇 西安飞行自动控制研究所  
中文摘要
      多目标优化的两个核心指标是收敛性和多样性，而对二者加以优化和权衡是多目标进化算法的关键。头脑风暴优化算法作为一种新型的群体智能优化算法，一经提出便引起了众多研究者的关注。本文在对现有的多目标头脑风暴优化算法研究的基础上，通过对决策变量进行分析，围绕收敛性和多样性分别进行优化，在对收敛性优化时通过分解策略增加选择压力，而在对多样性优化时以参考点更新种群增加多样性，最终扩展并提出了高维多目标头脑风暴优化算法。此外，本文提出一种以角点为聚类中心的自适应聚类方式，明确个体的导向，提高种群的扩展性。与现有的几种效果较好的多目标进化算法进行比较，大量的仿真结果表明了本文的算法具有优秀的性能。
英文摘要
      Convergence and diversity are two core indicators of multi-objective optimization. The optimization and the balance of them are the keys of Multi-Objective Evolutionary Algorithms(MOEA). Brain Storm Optimization (BSO) as a new kind of swarm intelligence optimization algorithm, which has paid more attention of more researchers in different fields. Based on the research of the existing MOBSO, this paper optimizes the convergence and diversity by analyzing the decision variables of problems. Decomposition strategy is carried out to increase the select pressure while the convergence optimization is performed, and the strategy of reference points is adopted to update the population to increase the diversity while the diversity optimization is performed. Finally, we extend and propose the Many-Objective Brain Storm Optimization Algorithm(MaOBSO). In addition, this paper proposes a new adaptive clustering strategy based on the corner points, which could clarify the orientation of individuals and improve the expansibility of population. Compared with several existing multi-objective evolutionary algorithms with better performance, a large number of simulation results show that the algorithm of this paper has excellent performance.