遗传算法在单层球壳动力响应优化中的应用  

Application of genetic algorithm in dynamic response optimization of single layer dome

作　　者：叶继红[1] 江季松[1] 

机构地区：[1]东南大学混凝土及预应力混凝土结构教育部重点实验室,南京210096

出　　处：《东南大学学报（自然科学版）》2010年第2期414-419,共6页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金重大研究计划培育项目(90715005);教育部新世纪优秀人才支持计划资助项目(NCET-07-0186);国家教育部博士点基金资助项目(200802860007)

摘　　要：针对标准遗传算法存在局部搜索能力不强、后期收敛速度放慢等缺陷,对遗传算法的杂交算子和变异算子采用自适应策略,同时引入基于预选择机制的小生境技术,提出一种小生境遗传算法(ANGA).以70m跨单层球壳为研究对象,将ANSYS非线性动力响应分析与ANGA算法相结合进行倒塌模式优化.分别以结构在某级水平的地震作用下结点动位移的最大值和结构最大结点动位移的给定区间为优化目标,对球壳进行全局优化和局部优化(径向杆件不变,只优化环向杆件和斜向杆件).结果表明,优化后的球壳可以避免出现无征兆的动力失稳,说明该算法可以有效解决优化变量繁多的大中型网壳结构动力响应优化问题.The standard genetic algorithm（GA） has the limitation of local convergence and slow convergence speed in the anaphase.An adaptive niche genetic algorithm（ANGA） is put forward by introducing an adaptive strategy to crossover operator and mutation operator and using a niche techique with pre-selection mechanism.An optimum analysis of collapse mode of 70 m single-layer latticed dome is carried out by combining nonlinear dynamic response analysis ANSYS and the ANGA.The maximum value of node dynamic displacement and a specified region of the maximum node displacement under a certain level earthquake excitation are considered as two optimization objectives,and global optimization and local optimization for reticulated shells are made（the section areas of members in radial direction keep invariant,while the others are adjusted）.Results show that dynamic instability of latticed dome can be avoided after optimization.Thus,the ANGA can be used to solve dynamic response optimal problem of large and medium reticulated dome with several optimum variables.

关 键 词：遗传算法 自适应策略 小生境技术 单层球壳 动力响应优化 

分 类 号：TU393.3[建筑科学—结构工程]