Research on Deformation Prediction of Deep Foundation Pits Based on Gene Expression Programming

This article first analyzes the current problems in the excavation and operation of deep foundation pits, especially the damage to the country's society, economy, and casualties after a major accident. Then it emphasizes the importance of deformation prediction, and briefly introduces the gene expression programming technology and its application advantages in deformation. Combined with a deep foundation pit project in Guangzhou, Guangdong Province, the first 33 phases of the monitoring data are used as training samples and the next 7 phases are used as test samples. The gray GM (1,1), BP neural network and gene expression are used. Three models are programmed to perform deformation prediction in the next seven periods, and their absolute, relative, and average relative errors are calculated. The average relative errors of these three models are calculated to be 2.7771%, 0.8881%, and 0.5606%, which proves that the gene expression programming model is feasible in the prediction of deep foundation pit deformation.