一种改进的小生境遗传算法  

Improved niching genetic algorithm

作　　者：郏宣耀[1] 王芳 

机构地区：[1]浙江大学宁波理工学院信息科学与工程分院,浙江宁波315100 [2]School of Business Management University of East Anglia,Norwich,England,NR4 7TJ

出　　处：《重庆邮电学院学报（自然科学版）》2005年第6期721-723,744,共4页Journal of Chongqing University of Posts and Telecommunications(Natural Sciences Edition)

基　　金：浙江大学宁波理工学院青年创新基金资助项目(2004-11)

摘　　要：简单遗传算法(SGA)存在早熟收敛和后期收敛速度慢的弱点,基于小生境(niche)技术的改进遗传算法因其较好地保持了种群多样性,显示出更优的性能,但它存在操作复杂、比简单遗传算法更费时的缺陷,因此提出了一种基于自适应的小生境遗传算法。该算法在多模函数的优化中能够保持种群多样度的稳定性,获取合适的子种群规模,从而以更快的收敛速度获得更优的解。仿真结果表明该算法高效、可靠,易于实现。Simple Genetic Algorithm （SGA） has the weaknesses of premature convergence and low speed of convergence in later stage then, the improved Genetic Algorithm based on Niche technique shows a better performance because it keeps the population diversity well, but it is more complex than SGA in operation and is more time consuming. This paper presents a new method based on self-adaptive, it can keep the population diversity stable and determine a suitable size of sub population in optimization of multimodal functions, so it can obtain more optimal solution at a much higher speed. The simulation experiment indicates that this new algorithm is efficient, reliable and easy to program.

关 键 词：简单遗传算法 小生境 多模函数优化 早熟收敛 自适应 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]