Computational Approaches in Slope Stability Analysis: A Comprehensive Review

Slope stability analysis is crucial in preventing landslide disasters, and artificial intelligence (AI) techniques have become an integral part of this research area. This review examines a wide range of studies that use AI methods to analyze slope stability. The analysis covers three key aspects: initial slip surface description, Factor of Safety (FOS) computation, and critical slip surface selection. Most studies focus on two-dimensional non-circular slip surfaces, using the typical method and its variations to generate feasible slip surfaces. The limit equilibrium method, known for its analytical simplicity and compatibility with optimization algorithms, is the most commonly used method for FOS computations. Studies are grouped into four categories based on AI methods: quasi-physical intelligence, simulated evolutionary methods, swarm intelligence, and hybrid intelligence. Swarm intelligence methods are the most prevalent, followed by simulated evolutionary and quasi-physical intelligence methods, with hybrid intelligence being less commonly used. The review highlights the benefits of these studies, including the global search capability for critical slip surfaces, applicability to diverse slope types, flexibility in FOS computation methods, introduction of novel AI techniques, and the proposal of typical methods for describing non-circular slip surfaces. This comprehensive review provides valuable insights into the evolving field of slope stability analysis, paving the way for future advancements.