Application of artificial neural networks to prediction of deformation in deep foundation pit

The application and research of artificial neural network in foundation deformation prediction, over the past decades and at home and abroad, are reviewed. Based on the two main aspects of foundation pit deformation, which are horizontal displacement of supporting structure and surface settlement around foundation, the evolution and corresponding research results of neural network prediction are reviewed and summarized when the input vector of neural network changes from historical deformation to multi-source data then to combination of them. On the basis of a thorough analysis and discussion, this paper points out the further research direction and development prospects.