Ultimate Shear Strength Evaluation of the Stirrup-Reinforced Concrete Beams by Ensembled Models

The main goal of engineering training is to create secure, reliable structures that are simple to design and evaluate using straightforward analytical methods. Consequently, a standard design method assumes that one is familiar with all fundamental structural typologies and then chooses the one that best meets the presumptive requirements. Nevertheless, Advances in computer modeling enable progress, yet shear load distribution in reinforced concrete beams with stirrups remains unclear. This work used artificial intelligence (AI) models to avoid these cognitive constraints. The present study introduces a data-driven machine learning methodology utilizing Random Forests Regression (RFR) in conjunction with the Pufferfish optimization algorithm (named RF_Pu ) and Red-tailed Hawk optimization models (named RF_Re ), for forecasting the shear strength (P_u) capacity of medium-to ultra-high-strength concrete beams featuring longitudinal reinforcement and vertical stirrups. A database that comprises 315 experimental data points from studies is employed for the reason of calculate the nine input elements for the P_u. This is done to accomplish the objective of computing the Pu. Given the available information, it is likely that both RFPu and RFRe will compute P_u correctly. Throughout the training and evaluation, the RF_Pu model’s good functional dependability was proved by the R^2 (0.97309 and 0.96698). With learning and assessment values of 20.9673 and 27.9776, respectively, the RFPu produced the smallest results on the RMSE measure.