Using the integrated neural network of radial basis function (RBF) via optimization algorithms to estimate pile settlement range

Abstract

The Pile motion seems to be one of the most critical in pile failure that requires appraisal before installing piles. The variables to estimate the Pile Settlement parameter, there are several methods. Among existing theoretical ways to investigate the pile movement mathematically, most studies have tried to model the piles’ settlement overloading period using artificial intelligence. Thus, this research has used the Artificial Neural Network to have the actual status of pile motion vertically over the loading periods dynamically and statically. Therefore, the present research has utilized the Radial Basis Function Neural Network joint with Equilibrium Optimizer Algorithm and Grasshopper Optimization Algorithm to figure out the optimum number of neurons within the hidden layer. Kuala Lumpur’s Klang Valley Mass Rapid Transit transportation network, Malaysia, opted to model the piles’ settlement and earth properties via the proposed hybrid RBF-GOA and RBF-EOA frameworks. By modeling both frameworks, the error index of RMSE for RBF-GOA and HRBF-EOA were gained to 0.6312 and 0.5947, respectively. However, the VAF indicator showed identical results of the rates 96.98 and 97.33, respectively. Overly, the RBF-EOA represented better than RBF-GOA by little efficiency.