摘要
提出一种基于数字序列编码遗传算法的高层结构黏滞阻尼器优化布置方法,解决了允许各层阻尼器安装数量不同时,二进制编码的标准遗传算法不能完备表达求解空间的问题。数字序列编码用染色体的一个基因位表示一个阻尼器的安装位置,其数值表示该阻尼器的安装层数。在这一编码方案下,优化问题基因型空间中的染色体和表现型空间中的可选布置方案一一对应,编码满足严格的合法性、完备性、Lamarckian性质以及强因果性。与数字序列编码方式相对应,离散重组交叉算子保证了种群的有效进化。此外,基于染色体目标函数值的相对大小构造适应度函数,能够充分体现种群中染色体的适应度差异,加速种群的进化,进而获得优化问题的全局最优解。对多遇地震下20层Benchmark结构的阻尼器布置方案进行了优化,计算结果表明了该方法的有效性。
An optimal locating method of viscous dampers in tall buildings was presented based on digital sequence coding genetic algorithm （DSCGA）.Compared with standard genetic algorithm （SGA）,the presented algorithm could completely represent the solution space of the optimization problem when the number of dampers was not equal in each damper-installation storey.A gene bit of digital sequence code represented an installation position of a damper,its value denoted the damper-installation storey.With the presented digital sequence coding,the size of the genotype space was exactly consistent with that of the phenotype space,while strict legitimacy,completeness,Lamarckian property,and strong causality were satisfied.The discrete-recombination cross operator,corresponding to the digital sequence coding, ensured the effective evolution of the population.Furthermore,a fitness function was constructed based on a relative objective function value,it could not only fully embody the fitness difference in population but also promote effective evolution to obtain the global optimal solution as a result.The method was applied for a typical damper positioning optimization problem,a 20-storey Benchmark structure under a frequent seismic level was considered.The efficiency and applicability of the presented algorithm were assessed.
出处
《振动与冲击》
EI
CSCD
北大核心
2015年第3期101-107,共7页
Journal of Vibration and Shock
基金
国家自然科学基金项目(51378135
11102045)
中国博士后科学基金面上项目(2013M542168)
高等学校博士学科点专项科研基金项目(20134410120005)
广州市属高校科研计划项目(2012A119)
关键词
数字序列编码
遗传算法
相对适应度
阻尼器
高层结构
digital sequence coding
genetic algorithm
relative fitness
damper
tall buildings
作者简介
燕乐纬男，博士，讲师，1978年10月生
通信作者周云男，教授，博士生导师，1965年9月生
