文档介绍：第 22 卷第 6 期山东理工大学学报(自然科学版) Vol. 22 No. 6
2008 年 11 月 Journal of Sha ndong U nive rsity of Technology (Natural Science Edition) Nov. 2008
文章编号:1672 - 6197 (2008) 06 - 0010 - 03
基于改进自适应遗传算法的仿真研究
姜静1 , 谭博学1 , 姜琳2
(1. 山东理工大学电气与电子工程学院, 山东淄博 255049 ;
2. 河南安阳钢铁公司第二炼钢厂, 河南安阳 455004)
摘要: 交叉概率 Pc 和变异概率 P m是遗传算法中重要的参数,自适应遗传算法中 P c和 Pm 能根据
个体适应度差异自适应地调节其大小,在快速收敛和全局最优之间获得了较好的平衡,但自适应遗
传算法对于进化初期不利. 改进的自适应遗传算法避免了进化初期较优个体处于停滞不前的状态.
分别用 3 种算法对典型的测试函数进行训练,仿真结果表明:改进的自适应遗传算法在收敛速度和
寻最优解方面是最优的.
关键词: 自适应; 遗传算法; 参数选择
中图分类号: TP18 文献标识码: A
Simulation analysis based on improved self2a da ptive ic algor ithm
J IAN G J ing1 , TA N Bo2xue1 , J IA N G Lin2
( 1. Sc hool of Electrical and Elect ronic Engineering , Shandong Univer sity of Technology , Zibo 255049 , China ;
2. The Second Steel Plant of Anyang Steel and Iron Corpora tio n , Anya ng 455004 , China)
Abstract : Crossover and mutation probabilities are important parameters in ic algorithm ,
self2adaptive ic algorithm can reach good balance between convergence speed and global opti2
mization , but it is not suitable to use the algorithm at the beginning of the evolution. The im2
proved algorithm can avoid t his ing. Training the typical test f unction by ic algo2
rithm , self2adaptive ic algorithm a nd the improved algorithm , the improved algorithm out2
pe rfor ms the other two o nes.
Key wor ds : self2a