电子学报

• 学术论文 • 上一篇    下一篇

高维多目标优化问题融入决策者偏好的集合进化优化方法

巩敦卫, 王更星, 孙晓燕   

  1. 中国矿业大学信息与电气工程学院, 江苏徐州 221116
  • 收稿日期:2013-04-22 修回日期:2013-07-24 出版日期:2014-05-25
    • 作者简介:
    • 巩敦卫 男,1970年3月出生于江苏铜山,博士,教授,博士生导师,主要研究方向:智能优化与控制,基于搜索的软件工程. E-mail:dwgong@vip.163.com王更星 女,1989年8月出生于河南开封,硕士研究生.主要研究方向:多目标进化优化理论与应用. E-mail:geng8659@163.com
    • 基金资助:
    • 国家自然科学基金 (No.61105063); 中央高校基本科研业务费专项资金 (No.2013XK09); 江苏省普通高校研究生科研创新计划 (No.CXZZ13-0932)

Set-Based Evolutionary Optimization Algorithms Integrating Decision-Maker’s Preferences for Many-Objective Optimization Problems

GONG Dun-wei, WANG Geng-xing, SUN Xiao-yan   

  1. School of Information and Electrical Engineering, China University of Mining and Technology, Xuzhou, Jiangsu 221116, China
  • Received:2013-04-22 Revised:2013-07-24 Online:2014-05-25 Published:2014-05-25

摘要: 高维多目标优化问题普遍存在且非常重要,但是,已有的解决方法却很少.本文提出一种有效解决该问题的融入决策者偏好的集合进化优化方法,该方法首先基于决策者给出的每个目标的偏好区域,将原优化问题的目标函数转化为期望函数;然后,以原优化问题的多个解形成的集合为新的决策变量,以超体积和决策者期望满足度为新的目标函数,将优化问题转化为2目标优化问题;最后,采用多目标集合进化优化方法求解,得到满足决策者偏好且收敛性和分布性均衡的Pareto优化解集.将所提方法应用于4个基准高维多目标优化问题,并与其他2种方法比较,实验结果验证了所提方法的优越性.

关键词: 进化优化, 高维多目标优化, 决策者偏好, 期望函数, 降维

Abstract: Many-objective optimization problems are common and important in real-world applications,previous theories and methods suitable for them,however,are few so far.We presented a set-based many-objective evolutionary optimization algorithm with integrating a decision-maker's preferences to effectively solve the problems above in this study.In the proposed method,each objective function of the original optimization problem was first transformed into a desirability function based on preference areas given by the decision-maker over it;thereafter,the optimization problem was further transformed into a bi-objective optimization one by taking such indicators as hyper-volume and the decision-maker's satisfaction as two new objectives in which a set formed by multiple solutions of the original optimization problem is as the new decision variable;finally,the transformed bi-objective optimization problem was solved by using a set-based evolutionary optimization algorithm to obtain a Pareto optimal set which meets the decision-maker's preferences and balances the convergence and the distribution.The proposed method was applied to four benchmark many-objective optimization problems and compared with the other methods.The experimental results showed its advantages.

Key words: evolutionary optimization, many-objective optimization, decision-maker's preferences, desirability function, dimensionality reduction

中图分类号: