文献类型：期刊文献

中文题名：一种用于多目标约束优化的改进进化算法

英文题名：Advanced evolutionary algorithm used in multi-objective constrained optimization problem

作者：俞国燕[1];李鹏[1];何真[1];孙延明[2]

机构：[1]广东海洋大学工程学院,广东湛江524025;[2]华南理工大学工商管理学院,广东广州510640

年份：2009

卷号：15

期号：6

起止页码：1172

中文期刊名：计算机集成制造系统

外文期刊名：Computer Integrated Manufacturing Systems

收录：CSTPCD、、EI(收录号：20092912195714)、Scopus(收录号：2-s2.0-67650045736)、CSCD2011_2012、北大核心2008、北大核心、CSCD

基金：国家自然科学基金资助项目(50675069);广东省海洋渔业局资助项目(A200899G02)~~

语种：中文

中文关键词：差分进化;约束优化;多目标优化;仿真

外文关键词：differential evolution; constrained optimization; multi-objective optimizationl ; simulation

中文摘要：当前求解多目标优化的进化算法主要考虑如何处理相互冲突的多个目标间的优化,很少考虑对约束条件处理的问题。对此,给出了一种基于双群体搜索机制的改进差分进化算法,以求解多目标约束优化问题。采用两个不同种群,分别保存可行个体与不可行个体的双群体约束处理策略,利用基于Pareto的分类排序多目标优化技术,完成对进化个体解的评价。并通过群体混沌初始化、自适应交叉和变异操作来提高基本差分进化算法的性能。对三个经典测试函数的仿真结果表明,文中算法在均匀性、逼近性及收敛速度三方面均优于非支配排序遗传算法,而收敛速度也优于另两种改进进化算法。

外文摘要：Evolutionary algorithm for constrained multi-objective optimization problems mainly focused on how to deal with conflicts among multi-objectives, while little consideration was given on how to deal with constraint condition. To deal with this problem, based on double populations searching scheme, an improved differential evolution algorithm was proposed for multi-objective constraint optimization problem. Two different populations were adopted to preserve constraints in optimization process, one was feasible solutions, and the other was infeasible solutions. To evaluate evolutionary individual, Pareto-based sorted ranking multi-objective technology was adopted. In addition, population chaotic initialization, adaptive crossover and mutation were adopted at the same time in order to improve the algorithm performance. Through experiments on three benchmark functions with constraints and multi-objectives, it showed that the proposed algorithm was superior to nondominated sorting genetic algorithm Ⅱ in the measure of uniformity, approximability and convergence speed, and was better than other two advanced evolutionary algorithms in the measure of convergence speed.