基于离散变量GA-APSO算法的钢结构优化分析  

The Optimal Analysis on Steel Structure Based on GA-APSO Algorithm with Discrete Variables

作　　者：郝润霞[1] 袁帅[1] 赵根田[1] HAO Run-xia;YUAN Shuai;ZHAO Gen-tian(College of Architecture and Civil Engineering/Inner Mongolia University of Science and Technology, Baotou 014010, China)

机构地区：[1]内蒙古科技大学建筑与土木工程学院,内蒙古包头014010

出　　处：《山东农业大学学报（自然科学版）》2016年第6期900-905,共6页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：国家自然科学基金项目(51268042)

摘　　要：本文针对遗传算法和粒子群算法收敛早熟、局部搜索能力差等缺点,在改进速度与位置更新算子函数的粒子群算法的基础上,插入了遗传算法的交叉和变异算子,提出了一种新的启发式现代混合算法——遗传-加速粒子群混合算法(GA-APSO)。该算法可以很好的跳出局部最优,扩大搜索域范围,提高收敛速度进而得到更合理的最优解。并基于离散变量将映射函数插入GA-APSO算法中,衍生出一种基于离散变量的GA-APSO算法,以一榀框架为算例通过与基于离散变量的APSO算法进行对比分析,证明了该衍生算法对于检索截面数据库中型钢规格自动选取具有一定的适用性。To overcome the shortcomings of “Premature Convergence" and poor local search ability in Genetic Algo rithm and Particle Swarm Optimization Algorithm, this paper proposed a new heuristic modern hybrid algorithm-Genetic-Accelerating Particle Swarm Optimization Algorithm (GA - APSO) by inserting the cross and mutation operat or of Genetic Algorithm based on Particle Swarm Optimization Algorithm improved speed and position update operator functions. The algorithm could expand the scope of the search field to improve convergence speed so that more reasonable optimal solution was obtained. Furthermore, GA - APSO algorithm with discrete variables was derived by inserting the mapping function. It was feasible to automatically select the medium-sized steels in a database by comparing with the APSO algorithm with discrete variables to take a common framework for an example.

关 键 词：GA-APSO算法 钢结构 优化设计 

分 类 号：TU391[建筑科学—结构工程] O224[理学—运筹学与控制论]