Probabilistic Stability Analysis of Rock Slope Using Soft Computing Techniques ANN, PSO and ANFIS

Abstract

In the present paper, probabilistic reliability analysis is performed for rock slope stability on a rock mass slope section using three different soft computing techniques: Artificial Neural Network (ANN), Particle Swarm Optimisation (PSO) and Adaptive Network-based Fuzzy Inference System (ANFIS). For performing the reliability analysis, the coefficient of variation for cohesion, angle of shear resistance and unit weight is taken as 0.30, 0.20 and 0.05, with mean values of 25 kPa, 37° and 26 kN/m³ respectively. These statistical properties are used to generate 120 datasets, and the factor of safety (FOS) of the rock slope is calculated in an MS-Excel spreadsheet incorporating the planar failure criteria. The generated dataset is divided into 70% for training and 30% for testing of the soft computing models. All models are evaluated based on multiple fitness parameters, Taylor diagram analysis and the Anderson–Darling statistical test to identify the most reliable model. The results indicate that while all models perform satisfactorily, the ANFIS model outperforms the ANN and PSO models, exhibiting the highest accuracy (R² = 0.9921, RMSE = 0.0187) and highest reliability index (β = 3.68) for predicting the FOS of rock slope stability.