钢框架优化遗传算法的若干改进  

Improvements on genetic algorithm for steel frame structural optimization

作　　者：孙国富[1] 鹿晓阳[1] 赵晓伟[1] 汲淑丽 

机构地区：[1]山东建筑工程学院工程结构现代分析与设计研究所,山东济南250014 [2]北京中建恒基投资有限公司监理部,北京100834

出　　处：《山东建筑工程学院学报》2005年第2期13-17,共5页Journal of Shandong Institute of Architecture and Engineering

基　　金：山东省教育厅科研计划项目(J01E03)

摘　　要：将遗传算法应用于钢框架结构优化设计时,在实际的应用中,基本的遗传算法存在着收敛速度慢和稳定性差等缺陷。为克服这些问题,采用了改进的适应度函数和约束处理方法、自适应的交叉、变异概率和最优保存策略,提高了遗传算法的效率和可靠性。算例表明改进的遗传算法对离散变量结构优化是有效、可行的。<Abstrcat>In this paper, the genetic algorithm theory was introduced into steel frames design. However, simple genetic algorithm has drawbacks such as slow convergence and less stability. To overcome these problems, such model methods were adopted as improved fitness function and penalty function, adaptive crossover probability and mutation probability in the search of GA. Numerical examples were given which indicated show that the improved genetic algorithm for discrete optimum design of structures was effective and feasible.

关 键 词：钢框架 遗传算法 优化设计 

分 类 号：TU391[建筑科学—结构工程]