摘要
借鉴生物免疫原理中抗体多样性产生及保持的机理,建立了一种多目标优化方法.该方法定义了多目标选择熵和浓度 调节选择概率的概念,采用了抗体克隆选择策略和高度变异策略.最后采用四种典型的多目标优化函数,将本方法同几种常用 的多目标遗传算法进行了比较研究,证明了所建立的基于免疫原理的多目标优化方法能有效解决多目标优化问题且具有一定 的优越性.
In this paper, an algorithm is proposed to solve multiobjective optimization problems by introducing the mechanism of producing and preserving the diversity of antibodies in organismal immune system into evolutionary algorithm. The conceptions of multiobjective selection entropy and selection probability based on concentration adjustment are defined, and strategies of antibody clonal selection and hate rofe mutation are introduced. Finally, the approach is compared with several traditional evolutionary approaches. In the comparative study, two metric and four typical testing functions are adopted. Experimental results indicate that the proposed approach can resolve the problem of multiobjective optimization effectively and has better performances.
出处
《小型微型计算机系统》
CSCD
北大核心
2005年第10期1770-1773,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(69772014)资助
关键词
人工免疫系统
多目标优化
多样性
进化算法
artificial immune system
multiobjective optimization
diversity
evolutionary algorithm