基于EGA与Alopex算法的非可微函数混合全局优化算法  

A Hybrid Global Optimization Algorithm for Non-differential Function Based on Elitist Maintained Genetic and Alopex Algorithms

作　　者：翟海峰[1] 苏宏业[1] 褚健[1] 赵明旺[2] 

机构地区：[1]浙江大学工业控制技术国家重点实验室先进控制研究所,浙江杭州310027 [2]武汉科技大学自动化系,湖北武汉430081

出　　处：《电路与系统学报》2002年第1期1-4,共4页Journal of Circuits and Systems

基　　金：国家杰出青年科学基金(NSFC: 60025308);高等学校优秀青年教师教学和科研奖励基金资助项目

摘　　要：遗传算法是一种全局优化算法,能以较大概率搜索到全局最优解。本文将Alopex算子嵌入到保留最优个体遗传算法(EGA)中,对非可微或求导困难函数从而得到既能以较大概率搜索全局极值,又能进行局部细致搜索的混合全局优化算法;并对其全局收敛性和计算效率作了证明与分析。数值计算结果表明该算法优于求解函数优化的EGA和Alopex算法。Genetic algorithm is a global optimization algorithm. In this paper, a hybrid global optimization algorithm for non-differential function is proposed. Within this algorithm, Alopex algorithm is used to Elitist maintained Genetic Algorithm (EGA). This new algorithm possesses not only the capability of global optimization, but also strong performance of locally searching, which is the feature most conventional Genetic algorithms do not possess. Convergence and efficiency of proposed algorithm are also analyzed. Numerical simulation results demonstrate that the new algorithm is superior to EGA and Alopex algorithm.

关 键 词：EGA ALOPEX算法 全局优化算法 非线性函数 遗传算法 全局优化 

分 类 号：O232[理学—运筹学与控制论]