Analyzing the efficacy of waste marble and glass powder for the compressive strength of self-compacting concrete using machine learning strategies

Self-compacting concrete (SCC) is well-known for its capacity to flow under its own weight, which eliminates the need for mechanical vibration and provides benefits such as less labor and faster construction time. Nevertheless, the increased cement content of SCC results in an increase in both costs and carbon emissions. These challenges are resolved in this research by utilizing waste marble and glass powder as cement substitutes. The main objective of this study is to create machine learning models that can predict the compressive strength (CS) of SCC using gene expression programming (GEP) and multi-expression programming (MEP) that produce mathematical equations to capture the correlations between variables. The models’ performance is assessed using statistical metrics, and hyperparameter optimization is conducted on an experimental dataset consisting of eight independent variables. The results indicate that the MEP model outperforms the GEP model, with an R 2 value of 0.94 compared to 0.90. Moreover, the sensitivity and SHapley Additive exPlanations analysis revealed that the most significant factor influencing CS is curing time, followed by slump flow and cement quantity. A sustainable approach to SCC design is presented in this study, which improves efficacy and minimizes the need for testing.