基于改进自适应多种群遗传算法的结构-控制系统一体化优化  

Integrated optimization of structure and control systems based on a modified adaptive multi-population genetic algorithm

作　　者：梅真 龚嘉诚 高毅超[1] 魏琳 李海锋[1] MEI Zhen;GONG Jiacheng;GAO Yichao;WEI Lin;LI Haifeng(College of Civil Engineering,Huaqiao University,Xiamen 361021,China;Construction&Development Co.Ltd.,China Construction Fourth Bureau,Xiamen 361006,China)

机构地区：[1]华侨大学土木工程学院,福建厦门361021 [2]中建四局建设发展有限公司,福建厦门361006

出　　处：《中南大学学报（自然科学版）》2024年第2期799-809,共11页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(51778248);福建省自然科学基金资助项目(2022J01291,2020J01057)。

摘　　要：提出一种改进的自适应多种群遗传算法,以更好地解决建筑结构-主动控制系统一体化优化问题,即同时对被控结构参数、控制算法参数、主动作动器布置位置进行优化。该遗传算法对编码方法、初始种群生成、选择策略、交叉概率和变异概率的自适应调整、多种群协同进化中移民策略等进行改进。研究结果表明:改进的自适应多种群遗传算法和改进的基本遗传算法优化结果总体一致,表明前者分析结果是正确的,并且具有较高的精度;改进的自适应多种群遗传算法和改进的基本遗传算法首次得到优化分析最优解的平均进化代数分别为320与730,表明前者比后者收敛速度更快;改进的自适应多种群遗传算法每次能达到或接近最优解,可有效克服基本遗传算法优化结果随机性较强的缺点;经改进的自适应多种群遗传算法优化的主动控制系统取得明显减振效果,E1 Centro波输入时,主动控制结构层间位移角峰值和绝对加速度峰值较无控时分别平均减小54.5%与46.7%。算例结果表明了改进的自适应多种群遗传算法的有效性,实现了对建筑结构-主动控制系统的一体化优化。A modified adaptive multi-population genetic algorithm(MAMPGA)was proposed to better solve the integrated optimization problem of building structure and active control systems.Specifically,structure parameters,control algorithm parameters and the placement of actuators were optimized simultaneously.The proposed MAMPGA was improved in coding method,initial population generation,selection strategy,adaptive regulation of crossover probability and mutation probability,migration strategy in multi-population co-evolution,etc.The results show that the optimization result of the MAMPGA is generally consistent with that of an improved simple genetic algorithm(ISGA),indicating the correctness and accuracy of the former algorithm.When the optimal solution is obtained for the first time,the average evolutionary generations of the MAMPGA and the ISGA are 320 and 730 respectively,which means that the former algorithm converges faster than the latter.The MAMPGA can reach or approach the optimal solution each time,so it can overcome the shortcoming of strong randomness in optimization results of the ISGA.The active control system optimized by the MAMPGA achieves a good vibration reduction effect.The peak values of inter-story drift angles and floor absolute accelerations of the actively controlled structure under E1 Centro wave are decreased by 54.5%and 46.7%respectively on average compared to those without control.The numerical example demonstrates the effectiveness of the MAMPGA,and the integrated optimization of building structure and active control systems is realized.

关 键 词：主动控制 结构-控制系统 一体化优化 自适应遗传算法 多种群 

分 类 号：TU352.1[建筑科学—结构工程]