| 一种基于聚集距离的多目标进化算法 |
| |
| 引用本文： | 薛娟,郑金华,李旭勇.一种基于聚集距离的多目标进化算法[J].计算机工程与应用,2005,41(36):60-62. |
| |
| 作者姓名： | 薛娟 郑金华 李旭勇 |
| |
| 作者单位： | 湘潭大学信息工程学院,湖南,湘潭,411105;湖南财经高等专科学校信管系,长沙,410205;湘潭大学信息工程学院,湖南,湘潭,411105 |
| |
| 摘 要： | 在多目标进化算法的研究中,解群体的多样性和运行效率是最重要的两个指标。在进化算法中一般采用构造非支配集的方法来保持算法的运行效率和解集的分布性;采用聚类技术来计算和维持解群体的分布性和多样性。文章提出了用庄家法构造非支配集和基于个体距离的聚类方法的多目标进化算法。经试验证明,该算法能够趋近到Pareto最优解,并且能保证较好的分布度。
|
| 关 键 词： | 多目标进化算法 聚类算法 非支配集 |
| 文章编号： | 1002-8331-（2005）36-0060-03 |
| 收稿时间： | 2005-03 |
| 修稿时间： | 2005-03 |
Multi Objective Evolutionary Algorithm Based on Clustering Distance |
| |
| Xue Juan,Zheng Jinhua,Li Xuyong.Multi Objective Evolutionary Algorithm Based on Clustering Distance[J].Computer Engineering and Applications,2005,41(36):60-62. |
| |
| Authors: | Xue Juan Zheng Jinhua Li Xuyong |
| |
| Affiliation: | 1.Zhe Institute of Information Technology, Xiangtan University,Xiangtan, Hunan 411105 ;2.Department of Information Management,Hunan College of Finance and Econonics,Changsha 410205 |
| |
| Abstract: | In the research of MOEA(Multi Objective Evolutionary Algorithm),diversity of the solution population and the efficiency of the algorithm are two important measures.Most MOEAs employ clustering to keep the diversity and variety of solutions and construct the non-dominated set to make algorithms efficient.This paper puts forward a new clustering based on clustering distance and employs the Dealer's Principle to construct the non-dominated set.It is proved by experimental results that the algorithm can approach the Pareto front and attain good diversity. |
| |
| Keywords: | Multi Objective Evolutionary Algorithm clustering non-dominated set |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
