关于多目标演化算法的策略分析

基于Pareto排序算法的多目标演化算法是多目标演化算法所采用的重要方法,本文叙述了多目标演化算法(MOEAs)的有关概念,在分析已有算法的一些性能和特征的基础上,结合演化算法的有关概念,重点基于Pareto排序算法分析了影响多目标演化算法性能的两大方面:求解过程中解集合的多样性、均匀性分布的保持与维护以及解的收敛性,分析了MOEAs设计中需要注意的策略问题以及今后研究的重点.

Analyzing algorithmic strategy on multi- objective evolutionary algorithms

Multi-objective evolutionary algorithms (MOEAs) based on Pareto sorting are basilic methods , which are often used in a multi-objective evolutionary algorithm.In this paper, some concepts about MOEAs are described ,and the performances and characteristics of some typical MOEAs proposed by some researchers are analyzed,then combining with the concepts about evolutionary algorithms,the paper mainly analyzes algorithmic performance of MOEAs based on Pareto sorting,which is influenced in two ways: while searching Pareto optimal set ,maintaining diversity and spread distribution of solutions,and convergence of solutions to Pareto set .Hence designing a multi-objective algorithm some strategies are proposed and analyzed ,and some emphases in MOEAs to be solved in the future are also suggested.