首页 | 本学科首页   官方微博 | 高级检索  
     

遗传算法取代时间的分析
引用本文:郭东伟,周春光,刘大有.遗传算法取代时间的分析[J].计算机研究与发展,2001,38(10):1211-1216.
作者姓名:郭东伟  周春光  刘大有
作者单位:吉林大学计算机科学与技术学院
基金项目:国家自然科学基金 ( 19872 0 2 7),教育部“符号计算和知识工程”重点实验室基金资助
摘    要:在遗传算法中,选择算子具有重要的地位,评价选择算子的标准是取代时间。给出了取代时间的严格定义。对于基于比例的选择算子,详细地分析了算子的取代时间,证明了取代时间的阶不依赖于适应度函数的选取和初始群体的分布,为O(lnN)阶。同时,提出了取代时间系数的概念,可以定量地衡量不同的选择算子、适应度函数变换方式的作用。分析了常见的适应度函数的变换对取代时间系数的影响。

关 键 词:时间分析  优化算法  遗传算法  计算机

ANALYSIS OF TAKEOVER TIME FOR GENETIC ALGORITHMS
GUO Dong Wei,ZHOU Chun Guang,and LIU Da You.ANALYSIS OF TAKEOVER TIME FOR GENETIC ALGORITHMS[J].Journal of Computer Research and Development,2001,38(10):1211-1216.
Authors:GUO Dong Wei  ZHOU Chun Guang  and LIU Da You
Abstract:In genetic algorithm, the selection operator occupies an important position. The character of selection operator depends on takeover time. In this paper, the strict definition of takeover time is shown. The takeover time of selection operator based on proportion is expressed so it is proved that the rank of takeover time is O (ln N ), and it does not depend on specific form of fitness function or initial population distribution. The takeover time coefficient is given, so the influence of different selection operators and transformation of fitness function can be measured quantitatively. The effect of usual transformation on fitness function is also analyzed.
Keywords:selection opertor  takeover time  takeover time coefficient
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号