Prediction of Ultimate Bearing Capacity of Pile Foundation Based on Two Optimization Algorithm Models

The determination of the bearing capacity pile foundations is very important for their design. Due to high uncertainty various factors between and soil, many methods predicting ultimate focus on correlation with field tests. In recent years, artificial neural networks (ANN) have been successfully applied types complex issues in geotechnical engineering, among which back-propagation (BP) method a relatively mature widely used algorithm. However, it has inevitable shortcomings, resulting large prediction errors other issues. Based this situation, study was designed accomplish two tasks: firstly, using genetic algorithm (GA) particle swarm optimization (PSO) optimize BP network. On basis, algorithms were improved enhance performance algorithms. Then, an adaptive (AGA) (APSO) network predict foundation. Secondly, test models, predicted results compared analyzed relation traditional model models same type literature based three most common statistical indicators. evaluated evaluation metrics, namely coefficient (R2), value account (VAF), root mean square error (RMSE), metrics set obtained as AGA-BP (0.9772, 97.8348, 0.0436) APSO-BP (0.9854, 98.4732, 0.0332). show that achieved higher accuracy, accuracy reliability, provides new foundations.