基于ACOR-LSSVR的基坑最大水平位移预测模型  

Prediction of Maximum Horizontal Displacement of Foundation Pit based on ACOR-LSSVM

作　　者：王磊 王军 崔志鹏 Wang Lei;Wang Jun;Cui Zhipeng(Jiangsu Nanjing Geo-Engineering Survey Institute,Nanjing 210041,China)

机构地区：[1]江苏南京地质工程勘察院,江苏南京210041

出　　处：《市政技术》2022年第4期205-208,共4页Journal of Municipal Technology

摘　　要：准确合理地预测基坑支护的变形,特别是最大水平位移,对于提高设计人员的工作效率和现场作业的经济效益具有重要的意义。因此,采用连续域蚁群智能优化算法(ACOR)结合最小二乘支持向量机(LS-SVR),构建出ACOR-LSSVR预测模型,并依托已有的基坑支护结构最大位移实测数据,分别运用ACOR-LSSVR、LS-SVR、BP模型进行了预测。研究结果表明:较之其他2种预测模型,ACOR-LSSVR预测精度最高;佐证了该模型在基坑支护结构最大位移预测方面,具有一定的优势。It is important to accurately predict the deformation of foundation pit support, especially the maximum horizontal displacement, for improving the working efficiency of designers and economic benefit. In this paper, combined the ant colony optimization for continuous domains(ACOR) with least squares support vector machine(LSSVR), prediction model of ACOR-LSSVR was established. Based on the existing maximum displacement measured data of the foundation pit, the prediction was conducted respectively by ACOR-LSSVM, LS-SVM and BP model. The research results show that ACOR-LSSVM model has the highest prediction accuracy compared with the other two models. The model is proved to have advantages in predicting the maximum horizontal displacement of foundation pit support.

关 键 词：基坑支护结构 最大位移预测 最小二乘向量机 连续域蚁群算法 

分 类 号：TU433[建筑科学—岩土工程]