一种改进的遗传算法及其性能研究  

An improved genetic algorithm and its performance study

作　　者：吴星[1] 齐虹[1] 陈冲[1] 

机构地区：[1]福州大学电气工程与自动化学院,福建福州350002

出　　处：《福建工程学院学报》2006年第4期407-410,共4页Journal of Fujian University of Technology

基　　金：福建省教育厅科研基金资助项目(K04007)

摘　　要：为克服遗传算法收敛速度慢、稳定性差的缺点,从模式定理出发,讨论遗传算法的本质,以基本遗传算法为基础,对选择算子进行改进,在引入最优个体保护策略以确保收敛性的基础上设定动态阈值,既提高交配池的平均适应值又兼顾群体多样性。仿真研究表明,改进的遗传算法有效提高群体的平均适应值,加快收敛速度,提高稳定性。To overcome the drawbacks of slow convergence and poor stability of genetic algorithm, an improved genetic algorithm is proposed based on the standard genetic algorithm by studying the essence of genetic algorithm according to the Schema Theorem. The improved genetic algorithm improves the selec- tion operator by introducing the preservation of the optimal individual algorithm that assures the conver- gence,setting a dynamic threshold that allows the individuals of higher adaptability to enter the crossing pool to enhance the average adaptability of the population and to assure the diversity of the population. Through the simulation experiments, the proposed algorithm shows its faster convergence and better stability.

关 键 词：遗传算法 模式定理 快速收敛 稳定性 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]