AN OPTIMIZED NEURO-FUZZY GROUP METHOD OF DATA HANDLING SYSTEM BASED ON GRAVITATIONAL SEARCH ALGORITHM FOR EVALUATION OF LATERAL GROUND DISPLACEMENTS

During an earthquake, significant damage can result due to instability of the soil in the area affected by internal seismic waves. A liquefaction-induced lateral ground displacement has been a very damaging type of ground failure during past strong earthquakes. In this study, neuro-fuzzy group method of data handling (NF-GMDH) is utilized for assessment of lateral displacement in both ground slope and free face conditions. The NF-GMDH approach is improved using gravitational search algorithm (GSA). Estimation of the lateral ground displacements requires characterization of the field conditions, principally seismological, topographical and geotechnical parameters. The comprehensive database was used for development of the model obtained from different earthquakes. Contributions of the variables influencing the lateral ground displacement are evaluated through a sensitivity analysis. Performance of the NF-GMDH-GSA models are compared with those obtained from gene-expression programming (GEP) approach, and empirical equations in terms of error indicators parameters and the advantages of the proposed models over the conventional method are discussed. The results showed that the models presented in this research may serve as reliable tools to predict lateral ground displacement. It is clear that a precise correlation is easier to be used in the routine geotechnical projects compared with the field measurement techniques.