An alternative approach in determining liquefaction potential based on genetic programming

One of the most important problems in the field of Geotechnical Earthquake Engineering is soil liquefaction. Liquefaction, which can cause loss of bearing capacity and different settlements in loose, saturated and cohesionless soils due to dynamic loads such as earthquakes; it can cause many different damages and financial losses under the structure and its foundations. The magnitude of this damage is not only related to the damage to the structure and building foundations, but also to the repair-improvement process after the damage, loss of workforce and economic losses. Liquefaction analysis based on both field test results and laboratory dynamic test results are the most frequently preferred methods to determine the possible liquefaction risk in the project site. However, developing technology and software enable the use of soft computing methos for liquefaction potential, as well as in different areas of civil engineering. Within the scope of this study, three different Gene Expression Programming (GEP) models, which are expected to provide a flexible and effective solution in determining the liquefaction potential, were created and the performance of the models was evaluated using different metrics. After the evaluations, it was concluded that the GEP models created could be used as a fast and practical method for estimating the liquefaction potential.