基于改进遗传算法的大型体育场馆传感器优化布置研究  

Optimal sensor placement for large-scale stadium based on improved genetic algorithm

作　　者：邢国华[1] 陈思锦 苗鹏勇 张雯 武艳如[1] 柳明亮 XING Guohua;CHEN Sijin;MIAO Pengyong;ZHANG Wen;WU Yanru;LIU Mingliang(School of Civil Engineering,Chang'an University,Xi'an 710061,Shaanxi,China;Shaanxi Architecture Science Research Institute Co.,Ltd,Xi'an 710082,Shaanxi,China)

机构地区：[1]长安大学建筑工程学院,陕西西安710061 [2]陕西省建筑科学研究院有限公司,陕西西安710082

出　　处：《建筑科学与工程学报》2024年第6期19-30,共12页Journal of Architecture and Civil Engineering

基　　金：陕西省重点研发计划项目(2022LL-JB-13);秦创原引用高层次创新创业人才项目(QCYRCXM-2023-059);陕西省自然科学基础研究计划项目(2022JC-LHJJ-06)。

摘　　要：为高效、准确地对大型体育场馆结构进行健康监测,根据其结构特点,提出了一种基于多种群非支配重组遗传算法(MP-RGA)的传感器优化布置方法。引入二重编码解决离散优化问题,通过多种群策略提高算法的搜索能力,通过非支配重组解决以往遗传算法优化后期的多样性丧失问题,最后采用自适应化交叉、变异策略改善算法前期的收敛速度和后期的寻优能力。以西安国际足球中心屋盖为例,通过数值算例验证不同传感器布置数量下利用该算法确定传感器布置方案时的有效性。结果表明:与量子遗传算法(QGA)和粒子群算法(PSO)相比,多种群非支配重组遗传算法收敛速度更快,以模态置信度准则为目标的优化效果分别提升了41.88%和91.27%;依据所提算法优化得到的传感器布置方案实际应用效果良好,表明该算法适用于大型体育场馆的传感器优化布置问题。In order to efficiently and accurately monitor the health of large-scale stadium structures,a sensor optimization layout method based on multi population non-dominated recombination genetic algorithm(MP-RGA)was proposed according to the structural characteristics of large-scale stadium.Introducing dual encoding to solve discrete optimization problems,improving the search ability of the algorithm through multiple population strategies,and solving the diversity loss problem in the later stage of genetic algorithm optimization through non-dominated recombination.Finally,adaptive crossover and mutation strategies were adopted to improve the convergence speed of the algorithm in the early stage and the optimization ability in the later stage.Taking the roof of Xi'an international football center as an example,the effectiveness of the algorithm in determining the sensor layout scheme under different sensor layout numbers was verified through numerical examples.The results show that compared with quantum genetic algorithm(QGA)and particle swarm optimization algorithm(PSO),multi population non-dominated recombination genetic algorithm has a faster convergence speed,and the optimization effect with modal confidence criterion as the objective can be improved by 41.88%and 91.27%,respectively.The actual application effect of the sensor layout scheme optimized based on the proposed algorithm is good,indicating that the algorithm is suitable for the sensor optimization layout problem of large-scale stadium.

关 键 词：大型体育场馆 健康监测 传感器优化布置 改进遗传算法 网壳结构 

分 类 号：TU311.3[建筑科学—结构工程]