Predicting Compressive Strength of High-Performance Concrete Using Metaheuristic-Optimized Least Squares Support Vector Regression

Abstract

This research establishes a novel model for predicting high-performance concrete (HPC) compressive strength, which hybridizes the firefly algorithm (FA) and the least squares support vector regression (LS-SVR). The LS-SVR is utilized to discover the functional relationship between the compressive strength and HPC components. To achieve the most desirable prediction model that features both modeling accuracy and generalization capability, the FA is employed to optimize the LS-SVR. To construct and verify the proposed model, this study has collected a database consisting of 239 HPC strength tests from an infrastructure development project in central Vietnam. Experimental results have demonstrated that the new model is a promising alternative to predict HPC strength.