单亲遗传算法的选择方式  

Selection Strategies of Partheno-Genetic Algorithm

作　　者：李茂军[1] 童调生[2] 

机构地区：[1]长沙电力学院电力工程系,湖南长沙410077 [2]湖南大学电气与信息工程学院,湖南长沙410082

出　　处：《系统工程与电子技术》2002年第10期87-89,共3页Systems Engineering and Electronics

基　　金：国家教育部博士点基金 (980 5 3 2 0 5 ) ;长沙电力学院科技基金 (9910 10 0 2 0 3 0 3 )资助课题

摘　　要：给出了单亲遗传算法的几种常用选择方式 ,并指出单亲遗传算法的全局收敛性和收敛速度与选择方式有关。锦标赛选择方式和父子竞争选择方式不能保证算法的全局收敛性 ,但有较快的收敛速度 ;按适应度比例选择方式在引入了最优保持操作后能保证算法的全局收敛性 ,但收敛速度较慢。Several main selection strategies of partheno genetic algorithm (PGA) are given. Is PGA globally convergent or not? How fast is convergence velocity of PGA? The answers to these questions are related with selection strategies of PGA. Neither tournament selection nor contention selection between parent and son can ensure that PGA is globally convergent, while both of the selection strategies enable PGA to converge faster than other ones. With the introduction of the optimal maintaining operation, the selection strategy by individual's fitness proportion guarantees a global convergence of PGA, but leads to a slower convergence than other ones.\;

关 键 词：单亲遗传算法 选择方式 收敛性 计算效率 

分 类 号：O242.23[理学—计算数学]