Evolutionary polynomial regression based modelling of clay compressibility using an enhanced hybrid real-coded genetic algorithm

• An efficient genetic algorithm with a new hybrid strategy combined with a self-adaptive mutation is first developed.
• A new approach for evaluating the compressibility of remoulded clays using the EPR and optimization methods is proposed.
• The EPR-based modelling of compression index gives a more accurate and reliable correlation using physical properties.

A new approach for evaluating the compressibility of remoulded clays using the evolutionary polynomial regression (EPR) and optimization methods is proposed. An efficient hybrid real-coded genetic algorithm (RCGA) with a new hybrid strategy combined with a self-adaptive mutation is first developed. Then, the enhanced RCGA is applied to construct the EPR procedure for compression index. To highlight the performance of the RCGA in the proposed procedure, three other excellent optimization algorithms are selected and compared. All comparisons between predictions and measurements demonstrate that the EPR-based modelling of clay compressibility using the enhanced RCGA gives a more accurate and reliable correlation between the compression index and physical properties of remoulded clays.