一种改进的抑制早熟收敛的遗传算法  

Genetic Algorithm with Improved Genetic Operation of Suppressing Premature Convergence

作　　者：巩固[1] 郝国生[1] 杨帆[2] 

机构地区：[1]徐州师范大学计算机科学与技术学院,徐州221116 [2]中石化管道储运公司徐州信息中心,徐州221008

出　　处：《计算机与数字工程》2009年第5期6-9,16,共5页Computer & Digital Engineering

基　　金：江苏省高校自然科学基础研究资助项目(编号:08KJD420002);徐州师范大学校级项目(编号:08XBL14)资助

摘　　要：针对遗传算法运算速度低、容易陷入局部最优值、早熟收敛等缺点,提出了遗传算法算子的一些改进策略,对遗传算法的选择、交叉、变异算子以及操作方法进行了改进,采用最佳保留选择策略,改进后的交叉与变异操作,使算法始终保持了种群的多样性,同时也提高了寻优最终结果的精确性。实验表明改进的遗传算法有效的改善了遗传算法的缺点,改进后的算法明显优于传统的遗传算法,该算法具有良好的有效性和可行性。In respect that the genetic algorithm＇s disadvantages of arithmetic speed lowly, running into local optimum easily and so on, the paper describes how to improve the genetic algorithm on the selecting, crossing, mutating operator, and operation method with the multi-operator crossed and mutated as well as adapts mutation. The improved strategies which reserve some elitist genes can reduce useless crossover effectively and thus the convergence speed and the search ca- pability are greatly improved when the elitist reserved genetic algorithm that keeps best strategies compared with the ge- netic algorithm. Experimentation shows that the shortage is meliorated by this improved genetic algorithm, and the improved algorithm is superior to the traditional genetic algorithm and has good validity and feasibility.

关 键 词：遗传操作 改进方法 早熟收敛 全局最优 遗传算子 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]