A hybrid technique adopted for study on the properties of alkali-activated binder concrete by using recycled concrete aggregates

This manuscript presents a novel hybrid technique for assessing the geopolymer concrete' properties incorporating recycled concrete aggregates. Named AO-GBDT, this hybrid approach combines the Aquila Optimizer (AO) and the Gradient Boosting Decision Tree algorithm. The primary objective is to reduce errors and predict the optimal strength of RCA geopolymers. The method investigates the interaction of key parameters and prediction values, like sodium silicate and sodium hydroxide to improve the binding materials. AO is specifically utilised for optimising the mixture proportions of RCA geopolymers, while GBDT is used to forecast the compressive strength of the optimised concrete mixtures. Additionally, the effects of fly ash (FA) and ground granulated blast furnace slag (GGBS) on the concrete's hardened and fresh characteristics are investigated. Performance evaluation conducted in MATLAB demonstrates the effectiveness of the proposed system, which is contrasted to existing strategy. The proposed strategy exhibits a root-mean-square error of 1.8, a correlation coefficient (R) of 0.95 and a mean absolute error of 0.9, indicating its superior predictive accuracy and efficacy compared to existing approaches.