Unconfined compression strength modelling of expansive soils for sustainable construction: GEP vs MEP

Genetic programming (GP) is presented as a new tool for the unconfined compression strength of expansive soils (‘UCS-ES’) by utilizing an experimental database with 195 datapoints. The five input parameters employed in formulation of gene expression programming (GEP) and multi gene expression programming (MEP) models are clay fraction (CF), liquid limit (LL), plasticity index (PI), specific gravity (G _s), maximum dry density (ρ _dmax), optimum moisture content (OMC), and swell percent (S _p). Simplified mathematical expressions were derived for both the GEP and MEP models to compute the UCS-ES. Various statistics, i.e., mean absolute error (MAE), root mean square error (RMSE), Nash–Sutcliffe efficiency (NSE), and correlation coefficient (R), scatter plots, and the sensitivity and parametric study, were used to evaluate the performance of these models. The results revealed that the UCS-ES are accurately determined by both the approaches, with the GEP model producing relatively superior performance (R ²_TrD = 0.806, R ²_TsD = 0.668). Furthermore, when the formulated models were compared with other AI models, it was found that they performed more efficiently. Hence, both the GEP and MEP models can reliably be deployed for determining the UCS-ES which reduces the time-consuming and laborious testing, thereby attaining sustainability in countering the problematic water-sensitive soils.