A new development of ANFIS–GMDH optimized by PSO to predict pile bearing capacity based on experimental datasets
A new development of ANFIS–GMDH optimized by PSO to predict pile bearing capacity based on experimental datasets
Hooman Harandizadeh1 Mohammad Mohsen Tawfiq2
1) PhD graduate in Civil Engineering, Department of Civil Engineering, Shahid Bahonar University, Kerman, Iran
2) Professor of Civil Engineering Department, Shahid Bahonar University, Kerman, Iran
IOI Code
:
XEEZ-HBCCF
Publication :
The First Conference on Artificial Intelligence and Smart Computing(aisc2022.semnan.ac.ir)
Abstract :
Predicting the final bearing capacity of piles with the help of field experimental results through artificial intelligence (AI) techniques is one of the most important and complex issues in pile analysis and design. The purpose of this research is to develop new artificial intelligence prediction models to predict pile bearing capacity. The first forecasting model based on the combination of adaptive neural fuzzy inference system (ANFIS) and data management group structure (GMDH) optimized by particle swarm optimization (PSO) algorithm was created, named ANFIS-GMDH-PSO model. The second model was introduced as a group method of fuzzy polynomial neural network type of data management model (FPNN-GMDH). The database contains various pile properties and soil properties, collected from the literature including CPT and pile loading test results, which are applied to the training and testing process of the developed models. Also, a common artificial neural network (ANN) model was used as a reference model for comparison and verification between the hybrid developed models for prediction. The modeling results showed that the improved ANFIS-GMDH model has relatively higher performance compared to ANN and FPNN-GMDH models in terms of accuracy and reliability level based on standard statistical performance indicators such as correlation coefficient (R), mean square error, root mean square error and error obtained standard deviation values.
Keywords : ultimate pile bearing capacity · deep foundation · anfis gmdh pso model · pso algorithm · fpnn gmdh model · gmdh network