摘要
对一种免疫遗传算法的求解性能进行理论分析。首先分析了算法的良好收敛性能;然后,进一步提出了临界浓度的概念,说明该算法与遗传算法的本质不同在与只有低于临界浓度的较优模式才能达到指数级增长,并在此基础上对算法的个体多样性维持能力进行了分析说明。本工作有利于从理论上进一步揭示这类改进遗传算法求解性能得以提高的根本原因。
The performance analysis of the Immune Genetic algorithm was focued on. Firstly, the global convergence of the Immune Genetic algorithm was analyzed. Secondary, after a concept of Critical Density was proposed, the essential difference between Immune Genetic Algorithm and Genetic Algorithm was given that only the better schemas which have lower density than the corresponding Critical Density could exponentially increase. Finally, the ability of maintaining the diversity of individuals was analyzed. This work is useful to theoretically explore and explain why such kind of improved Genetic Algorithm can get better performance.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2006年第4期873-876,共4页
Journal of System Simulation
基金
国家自然科学基金(60404004)
国家博士后科学基金(2003034433)
安徽省教育厅重点项目(2004kj360zd)资助项目
