Estimation of pile-bearing capacity of rocks via reliable hybridization techniques

Precisely calculating the pile-bearing ability is essential to the safe and effective design of foundations in civil engineering. This article's principal objective is to create cutting-edge artificial intelligence prediction models for pile-bearing capacity (P_u) assessment. This study uses a basic prediction technique based on the K-nearest neighbor (KNN) architecture. A unique hybrid method that combines the Ebola Optimization Search (EOS) and the Aquila Optimizer (AO) has been applied to provide predictions that are both optimum and accurate. A dataset of 200 sample histories from static load tests conducted on driven piles was used to build and validate the model. These datasets were employed for training, validating, and testing models at every stage of the development process. The method used in the investigation produced accurate results, demonstrating the efficacy of the predicted models. The performance of the traditional KNN model has been significantly enhanced by the incorporation of a hybridization approach, which has produced trustworthy predictions for P_u. The KNN model optimized with AO optimizers produces reliable results, as indicated by the model's R² and RMSE values of 0.990 and 34.878, respectively.