摘要
多目标遗传算法(MOGA)大体上可以分为两个步骤:构造非支配集和保持解分布度。而ε支配能将两者有机地结合起来,具有良好的时间效率及分布度。但是采用ε支配时,其ε参数难以设定,为此文章提出了一种基于自适应ε支配的快速多目标遗传算法(AEMOGA)。通过与其它的2个多目标遗传算法NSGA2和SPEA2比较,实验结果表明该文提出的算法具有良好的时间效率分布性、收敛性及时间效率。
Multi-objective Genetic Algorithm has two main steps:creating non-dominated set and keeping diversity.ε- domination can combine the two steps to one with low time complexity and good diversity.It is difficult to set ε- parameter when ε-domination is used,for this,a fast Multi-objective Genetic Algorithm(AEMOGA) based on self-adaptation ε-domination is suggested in this paper.Comparing AEMOGA with the two other MOGAs,NSGA2 and SPEA2,the result of the experiment shows that the algorithm suggested in the paper gets good diversity,convergence and low time complexity.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第15期42-44,共3页
Computer Engineering and Applications
基金
国家自然科学基金资助项目(编号:90104021)
湖南省自然科学基金资助项目(编号:01JJY2060)
关键词
自适应
ε支配
多目标遗传算法
self-adaptation,ε-domination,Multi-objective Genetic Algorithm(MOGA)