PREDICTION OF SOIL-WATER CHARACTERISTIC CURVE USING GENE EXPRESSION PROGRAMMING

Abstract

Soil–Water Characteristic Curve (SWCC) is one of the most important parts of anymodel that describes unsaturated soil behavior as it explains the variation of soil suction withchanges in water content. In this research, Gene Expression Programming (GEP) is employed asan artificial intelligence method for modelling of this curve. The principal advantage of the GEPapproach is its ability to generate powerful predictive equations without any prior assumption onthe possible form of the functional relationship. GEP can operate on large quantities of data inorder to capture nonlinear and complex relationships between variables of the system. The selectedinputs for modelling are the initial void ratio, initial gravimetric water content, logarithm ofsuction normalized with respect to atmospheric air pressure, clay content, and silt content. Themodel output is the gravimetric water content corresponding to the assigned input suction.Sensitivity and parametric analyses are conducted to verify the results. It is also shown that claycontent is the most influential parameter in the soil–water characteristic curve. The resultsillustrate that the advantages of the proposed approach are highlighted.