提出了一种粗粒度并行遗传算法，并将其应用于被动控制结构中阻尼器布置位置的优化． 该算法把一个种群划分为多个子种群，各个子种群可以独立完成经典遗传算法操作． 对于给定基因编码的种群个体，通过 Matlab-ABAQUS-Python 的交互使用，利用Matlab编程生成 INP 模型文件，并先调用 ABAQUS 来进行模型分析，后调用 Python 来读取结果数据并传输给 Matlab，求解结构模型的目标函数值． 文中还以层间位移角为控制目标，对 10 层被动控制钢框架结构的阻尼器优化布置进行了实例分析． 结果表明: 粗粒度并行遗传算法与经典遗传算法相比，既提高了种群的多样性，又加快了种群的收敛速度; 对比常规隔层方法，采用该算法可使结构减震率至少提高 19． 3% ，说明该算法能显著提高结构减震率．

A coarse-grained parallel genetic algorithm was proposed to optimize the damper arrangement in passive controlled structures． According to the algorithm，a population is divided into several sub-populations，which can independently complete the operation of genetic algorithm． For a given gene coding individual，the INP model files are firstly generated，numerical analysis is conducted in ABAQUS environment and the result data is read and transmitted in Python environment by using Matlab programming． Through the interactive use of Matlab-ABAQUS- Python，the value of objective function is figured out． A 10-story passive-controlled steel frame structure is selected as the research object，and the optimal design of damper arrangement is conducted by taking the inter-story drift an- gle as the control objective． The results show that the algorithm can not only improve the diversity of population compared with the classical genetic algorithm，but also accelerate the convergence speed of population． Compared with the results of the structure with the dampers placed in every two storeys，the inter-story drift reduction ratio of the structure equipped with the new algorithm can be increased by 19． 3% ，which indicates that the seismic re- sponse reduction ratio of the structure is greatly improved by using the algorithm．