自适应改进遗传算法在土木工程优化设计中的应用  

Adaptive and improved genetic algorithm’s application of optimal design of civil engineering

作　　者：范鹤[1] 张延年[2] 范泽[3] 刘斌[1] 

机构地区：[1]东北大学资源与土木工程学院,辽宁沈阳110004 [2]沈阳建筑大学土木工程学院,辽宁沈阳110015 [3]辽宁地质工程职业学院资源系,辽宁丹东118008

出　　处：《辽宁工程技术大学学报（自然科学版）》2007年第6期865-867,共3页Journal of Liaoning Technical University (Natural Science)

基　　金：国家自然科学基金资助项目(50508008);博士后科学基金资助项目(20060400971)

摘　　要：为了改进基本遗传算法的早熟收敛、随机漫游,验证改进后的算法在土木工程中的应用效果,提出4点改进措施,形成自适应遗传算法。利用混沌序列的"遍历性、随机性、规律性"的特点生成初始种群;引入适应度的指数尺度变换;提出自适应的交叉、变异概率公式;幅度系数调节交叉率、变异率。改进后的优化结果与基本遗传算法进行了对比,在进化时间、收敛精度上得到提高。通过土木工程中桁架结构非对称受荷的数值算例,证明自适应改进遗传算法是现实可行的。In order to improve basic genetic algorithm＇s premature convergence and random roam and verify improved algorithm＇s application effect in civil engineering, the article brings forward four improved measures, with which adaptive and improved genetic algorithm came into being. By the use of the chaos serial＇s properties of ＂ergodicity, randomness, regularity＂, original population is generated. Fitness index scaling is adopted. Adaptive crossover and mutation rate formula are presented. Crossover rate and mutation rate are adjusted by extent coefficient. Improved optimal result is compared with basic genetic algorithm. Evolution time and convergence precision are increased. By a numerical example of bar truss structure with unsymmetrical loads in civil engineering, adaptive and improved genetic algorithm is feasible.

关 键 词：优化设计 基本遗传算法 自适应改进遗传算法 数值算例 

分 类 号：TU311.4[建筑科学—结构工程]