改进遗传算法全局收敛性分析  

Global convergence analysis of improving genetic algorithm

作　　者：杨世达[1] 李庆华[1] 阮幼林[1] 

机构地区：[1]华中科技大学计算机学院,湖北武汉430074

出　　处：《计算机工程与设计》2005年第7期1695-1697,共3页Computer Engineering and Design

基　　金：国家自然科学基金项目(60273075)。

摘　　要：传统的遗传算法大多数没有给出收敛性准则。一类新的改进的遗传算法被提出,该算法即考虑了优化问题的全局性要求——每一步构造一个新函数,而这往往却比局部最优理论和方法困难得多;同时通过对选择算子的改进,对遗传算法后期进化缓慢问题得到了有效控制,最后给出了算法的收敛性证明以及收敛性准则。实例证明该算法是有效的。The most traditional genetic algorithms didn＇t give a termination rule. A new kind of genetic algorithm is presented. In this algorithms, an algorithm for finding global minimization was proposed each phase must constructed a new function, which was more difficult than local minimization. Meanwhile a selection operator was presented to make the place of the traditional one. It could prevent the latter slow evolution. The convergence of this algorithm is proved. A termination rule is given. The algorithm is efficiency proved with some instances.

关 键 词：遗传算法 选择算子 收敛准则 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]