Probabilistic Stability Analysis of Earthen Slope Using ANN, PSO-ANN, GPR and GA-ANFIS Soft Computing Techniques

In the present paper, reliability analysis is performed for soil slope stability on c–ϕ soil slope section using four different soft computing techniques ANN, PSO-ANN, GPR and GA-ANFIS. For performing reliability analysis, the coefficient of variation for cohesion, angle of shear resistance and unit weight is taken as 0.3, 0.2 and 0.03, mean value as 10 kPa, 30° and 20 kN/m³, respectively, was used to generate 100 datasets and factor of safety (FOS) of soil slope was calculated by Morgenstern-price method using the GeoStudio 2016 softwareAfter the generation of the actual dataset for the factor of safety of soil slope stability, the dataset is divided into 70% and 30% of the training and testing, respectively, of the soft computing models (ANN, PSO-ANN, GPR and GA-ANFIS). Soft computing models were used to evaluate factor of safety while in training and testing. All of the models are analysed based on various fitness parameters, Taylor diagram and statistical Anderson Darling test to find the most reliable model for the slope stability analysis of c–ϕ soil slope section under study. From the results, it was found that all models performed well, but GA-ANFIS outperformed on comparison among the models i.e. GA-ANFIS model is having higher accuracy and lowest error for the prediction FOS of the soil slope stability.