一种新的遗传算法停止准则  

A new terminate rule for genetic algorithms

作　　者：李纯莲[1] 王希诚[1] 赵金城[2] 

机构地区：[1]大连理工大学电信学院计算机科学与工程系,辽宁大连116023 [2]大连大学生物信息学与分子设计研究所,辽宁大连116621

出　　处：《辽宁工程技术大学学报（自然科学版）》2004年第1期62-64,共3页Journal of Liaoning Technical University (Natural Science)

基　　金：国家973计划资助项目(G1999032805);国家自然科学基金资助项目(10272030)

摘　　要：遗传算法因其思想简单、易于实现及不需要目标与约束函数的灵敏度信息等优点而越来越受到国内外广大学者的关注。本文以基于空间收缩的多种群遗传算法中搜索空间的收缩精度作为算法停止准则,并与传统的控制算法收敛的方法进行比较分析,实验表明本文算法停止准则能够及时准确地控制算法收敛,较之传统方法更具科学性和有效性。Genetic algorithm is more and more popular in almost every research field. The authors proposed a new GA named An Information-Entropy-based Multi-population Genetic Algorithm, which is paid greet attention to by scholars and abroad for both performances of speed and precision under a set of test problems [1]. In this paper, the terminate rule that used in the paper [1] will be formulated by the comparison with traditional rules. Experimental results show that this terminate rule is both more scientific and efficient, then traditional one.

关 键 词：遗传算法 停止准则 空间收缩 搜索空间 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]