A multi-objective optimisation approach for activity excitation of waste glass mortar

Waste glass is promising to be recycled and reused in construction for sustainability. Silicon dioxide the main component of glass, however, its pozzolanic activity latent mainly due stable silica tetrahedron structure. To excite activation waste chemical mechanical grinding powder (WGP) were investigated. As supplementary, hydrothermal combined (mechanical-chemical-hydrothermal) treatments conducted on part WGP samples. The unconfined compression strength (UCS), expansion caused by alkali–silica reaction (ASR), microstructural morphology results showed dosage threshold (around 2%) activators (alkali sodium sulfate) optimal. Besides, a firefly algorithm (FA) based multi-objective optimisation model (MOFA) was applied seek Pareto fronts three objectives: UCS, ASR expansion, Cost mixture proportion. objective functions UCS established through training machine learning (ML) models where FA used tune hyperparameters. cost calculated polynomial function. ultimate values root mean square error (RMSE) correlation coefficient (R) robustness ML models. Moreover, mortars containing 300 ?m 75 WGPs successfully obtained, which contributed practical application mortar production. In addition, sensitivity analysis rank importance input variables. that curing time, activator's content, particle size essential parameters.