作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (3): 152-154,157. doi: 10.3969/j.issn.1000-3428.2011.03.054

• 人工智能及识别技术 • 上一篇    下一篇

几种计算超体积算法的比较研究

周秀玲1,郭 平2,陈宝维3,王 静2   

  1. (1. 北京城市学院人工智能研究所,北京 100083; 2. 北京师范大学图像处理与模式识别实验室,北京 100875; 3. 中国辐射防护研究院,太原 030006)
  • 出版日期:2011-02-05 发布日期:2011-01-28
  • 作者简介:周秀玲(1972-),女,副教授、硕士,主研方向:智能 计算;郭 平,教授、博士;陈宝维,副研究员、硕士;王 静, 讲师、硕士
  • 基金资助:
    国家自然科学基金资助项目(10605021, 60675011)

Comparative Research on Algorithms for Computing Hypervolume

ZHOU Xiu-ling 1, GUO Ping 2, CHEN Bao-wei 3, WANG Jing 2   

  1. (1. Artificial Intelligence Institute, Beijing City University, Beijing 100083, China; 2. Image Processing and Pattern Recognition Laboratory, Beijing Normal University, Beijing 100875, China; 3. China Institute for Radiation Protection, Taiyuan 030006, China)
  • Online:2011-02-05 Published:2011-01-28

摘要: 对LebMeasure算法、HSO算法、HSO+MWW算法以及HKMP算法的基本思路、关键问题进行评述,在几种测试数据集上对算法的性能进行比较验证。实验结果表明,对于所有类型的前沿,HSO+MWW的性能好于HSO算法;当处理点的数目超过某一值时,HKMP算法的性能好于HSO算法,与理论分析一致;对于HKMP算法和HSO+MWW算法,在random和discontinuous前沿上,当处理点的数目超过某一值时,HKMP算法的性能好于HSO+MWW算法;但在spherical和degenerate前沿上,HSO+MWW算法的实际性能远好于HKMP算法。

关键词: 进化计算, 多目标进化算法, 超体积

Abstract: The main idea and the key point of LebMeasure algorithm, HSO algorithm, HSO+MWW algorithm and HKMP algorithm are introduced. The performance of the algorithms is compared on the different data set. Experimental results show that the performance of HSO+MWW is better than that of HSO for all of the fronts. For HKMP and HSO, it consists with the theory that the performance of the former is better than that of the latter when the number of the processed point exceeds some value. For the random front and the discontinuous front, the performance of HSO+MWW is worse than that of HKMP when the number of the processed point exceeds some value. But for the spherical front and the degenerate front, the performance of HSO+MWW is better than that of HKMP.

Key words: evolutionary computation, mulitiobjective evolutionary algorithm, hypervolume

中图分类号: