Soil liquefaction modeling by Genetic Expression Programming and Neuro-Fuzzy

Liquefaction of soils induced by the earthquake is one of the major complex problems for the geotechnical engineering. It is generally determined from in situ tests and laboratory test of which application is very difficult, expensive and time consuming. They also require extreme cautions and labor. Hence the development of new models for the prediction of liquefaction potential of soils provides fairly significant facility to design the constructions. This study presents the potential of Genetic Expression Programming (GEP) and Adaptive Neuro-Fuzzy (ANFIS) computing paradigm to forecast the safety factor for liquefaction of soils. To develop the models, a total of 569 data set collected from the literature were used. Five parameters such as standard penetration test ((N1)60), percentage of finest content (FC), effective overburden stresses (σ′), cyclic stress ratios (CSR) and angle of shearing resistance (ϕ′) were used as input parameters. The performance of models was comprehensively judged using several statistical verification tools. The results revealed that GEP and ANFIS models are fairly promising approach for the prediction of the soil liquefaction potential and capable of representing the complex relationship between seismic properties of soils and their liquefaction potential.