电子学报 ›› 2014, Vol. 42 ›› Issue (8): 1653-1659.DOI: 10.3969/j.issn.0372-2112.2014.08.031

• 科研通信 • 上一篇    下一篇

基于模糊支配的高维多目标进化算法MFEA

毕晓君, 张永建, 陈春雨   

  1. 哈尔滨工程大学信息与通信工程学院, 黑龙江哈尔滨 150001
  • 收稿日期:2013-06-17 修回日期:2013-08-14 出版日期:2014-08-25
    • 通讯作者:
    • 张永建
    • 作者简介:
    • 毕晓君女,1964年11月生于黑龙江省哈尔滨.哈尔滨工程大学信息与通信工程学院教授,博士生导师.主要研究方向为智能信息处理、图像处理.E-mail:bixiaojun@hrbeu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61175126); 中央高校基本科研业务费专项资金 (No.HEUCFZ1209); 高等学校博士学科点专项科研基金 (No.20112304110009)

A Many-Objective Evolutionary Algorithm Based on Fuzzy Dominance:MFEA

BI Xiao-jun, ZHANG Yong-jian, CHEN Chun-yu   

  1. Department of Information and Communication Engineering, Harbin Engineering University, Harbin, Heilongjiang 150001, China
  • Received:2013-06-17 Revised:2013-08-14 Online:2014-08-25 Published:2014-08-25
    • Supported by:
    • National Natural Science Foundation of China (No.61175126); Fundamental Research Funds for the Central Universities (No.HEUCFZ1209); Research Fund for the Doctoral Program of Higher Education of China (No.20112304110009)

摘要:

为提高高维复杂多目标优化算法的收敛性和解集分布性,提出一种基于模糊支配的高维多目标进化算法MFEA.在第二代Pareto支配类高维多目标进化算法模型基础上,利用模糊理论对模型中的环境选择进行改进,提出基于模糊隶属度的支配关系,并结合Harmonic、k邻域法和小生境技术对其中的拥挤密度估计方法进行改进,最后根据高维多目标的特点并结合模糊理论α-截集的思想提出了新的环境选择策略.将该算法与目前性能最好的5种多目标进化算法在标准测试函数集上进行对比试验,结果表明本文算法与其他算法相比具有明显的优势,不仅提高了算法的收敛性能,而且保证了Pareto最优解的均匀分布性.

关键词: 高维多目标优化, 模糊隶属度, 模糊支配, Harmonic平均距离, α-截集

Abstract:

In order to improve the convergence and distribution of Many-Objective Evolutionary Algorithms (MOEAs),this paper proposes a Many-Objective Fuzzy Evolutionary Algorithm (MFEA) which is based on fuzzy dominance.On the model of algorithms based on Pareto-dominance,we improve the environmental selection using fuzzy logic.We present a new dominance strategy based on fuzzy membership.Then,we propose a new estimation method of crowding distance which incorporates Harmonic-distance,k-neighborhood method and niche technique.Finally,according to the characteristics of MOPs and the idea of α-cut set,we design a new environmental selection strategy which is made up of two truncations.The proposed algorithm is compared to 5 state-of-the-art MOEAs on benchmark test problems.Simulation results show that α-MFEA has obvious advantages than other algorithms because MFEA could ensure good convergence while has uniform distribution,especially,applied to solving high-dimensional MOPs.

Key words: many-objective optimization, fuzzy membership, fuzzy dominance, Harmonic average distance

中图分类号: