离散变量结构优化设计的混沌遗传算法  

Chaotic genetic algorithm for structural optimization with discrete variables

作　　者：郭鹏飞[1] 韩英仕[1] 

机构地区：[1]辽宁工学院土木建筑系,辽宁锦州121001

出　　处：《辽宁工程技术大学学报（自然科学版）》2007年第1期68-70,共3页Journal of Liaoning Technical University (Natural Science)

基　　金：辽宁省自然科学基金资助项目(20032136)

摘　　要：利用混沌搜索的遍历性、随机性、规律性等特点,提出了一种求解离散变量结构优化设计的混沌搜索方法;将混沌搜索技术嵌入遗传算法,与基本遗传算子共同构成了一种离散变量结构优化设计的混合遗传算法—混沌遗传算法;通过自适应的退火因子和罚函数来处理约束条件,使算法逐渐收敛于全局可行最优解。计算结果表明,该方法有效地克服了基本遗传算法中的“早熟”现象,并具有更快的收敛速度。By use of the properties of periodicity, stochastic property and regularity of chaos, a chaotic approach method is presented for structural optimumal design with discrete variables. Through defining a chaotic operator in the genetic algorithm, a hybrid genetic algorithm for structural optimization with discrete variables, combined the advances of both genetic algorithm and the chaotic design method is presented in this paper, which adopted adaptive annealing penalty factors and penalty function. The numerical results show that the hybrid genetic algorithm has a rather high convergence speed and can locate the global optimization with a rather large probability for solving structural optimal design with discrete variables.

关 键 词：混沌搜索 离散变量 结构优化 遗传算法 退火罚函数 

分 类 号：TU311.4[建筑科学—结构工程]