Estimation of strength of ground granulated blast furnace slag concrete using computational algorithm

To make concrete, substitute materials for cement are used, such as ground granulated blast furnace slag ( \(GGBFS\) $\mathit{GGBFS}$ ). The process of making iron in a blast furnace produces \(GGBFS\) $\mathit{GGBFS}$ as a by-product. However, typical statistical approaches were found to be insufficient for describing concrete because of the complex relationships between the mechanical characteristics of the mixtures and their compositions. To address this problem, the study used a dataset of 533 mixes and seven input variables to predict the compressive strength \((CS)\) $(CS)$ of concrete using machine learning \((ML)\) $(ML)$ techniques, notably the multi-layer perceptron ( \(MLP\) $\mathit{MLP}$ ) and radial basis function ( \(RBF\) $\mathit{RBF}$ ). The study additionally employed a meta-heuristic optimizer, the reptile search algorithm ( \(RSA\) $\mathit{RSA}$ ), to enhance the efficiency and accuracy of the \(MLP\) $\mathit{MLP}$ and \(RBF\) $\mathit{RBF}$ models. By leveraging \(ML\) $\mathit{ML}$ algorithms and a meta-heuristic optimizer, the study aims to improve the efficiency of predicting \(CS\) $\mathit{CS}$ in concrete mixtures. It can contribute to resource optimization by reducing the time and resources required for manual testing and experimentation in concrete mixture design. Efficient prediction of concrete properties can lead to cost savings in terms of material usage and production processes. Regarding the results’ descriptions and justifications using error-based metrics, error distributions, correlation findings, and comprehensive evaluator metrics, in terms of prediction, it can be said that the \(MLP\) $\mathit{MLP}$ model in conjunction with the \(RSA\) $\mathit{RSA}$ algorithm is the system that performs better.