Function finding via genetic expression programming for strength and elastic properties of clay treated with bottom ash

• The GEP predictions of UCS and Es reasonably follow with the measured ones.
• The GEP formulas are performed by significant correlations (p<0.05, R>0.847).
• The GEP predictions have better correlations than the nonlinear regression.
• The bottom ash dosage and the energy absorption capacity are dominant parameters.

In order to understand the treatment of a marginal soil well, the underlying input–output relationship on the strength and elastic responses due to nonlinearity has always been a great importance in the stabilized mixtures for an optimal design. This paper employs a relatively new soft computing approach, genetic expression programming (GEP), to formulations for unconfined compressive strength (UCS) and elasticity modulus (Es) of clay stabilized with bottom ash, using a database obtained from the laboratory tests conducted in the study. The predictor variables included in the formulations are bottom ash dosage, dry unit weight, relative compaction, brittleness index and energy absorption capacity. The results demonstrate that the GEP-based formulas of UCS and Es are significantly able to predict the measured values to high degree of accuracy against the nonlinear behavior of soil (p<0.05, R>0.847). The GEP approach is found to have a better correlation performance as compared with the nonlinear regression as well. The sensitivity analysis for the parameter importance shows that the dominant influence on the predictions of UCS and Es is exerted by the variables of bottom ash dosage and energy absorption capacity. This study reveals that the GEP is a potential tool for establishing the functions and identifying the key variables for predicting the strength and elastic responses of the clay treated with bottom ash. Including a waste material in the proposed formulas can also serve to the environment for the development of recycling and sustainability.