Enhancing geotechnical design efficiency in compressible soil with artificial intelligence using genetic algorithms

This paper presents an unconventional approach to optimise geotechnical designs in compressible soil using Genetic Algorithms (GA), a subset of Artificial Intelligence (AI) algorithms. GA enhances the efficiency of two geotechnical designs: a parametric study and geometric optimisation in this paper. Geotechnical engineers primarily rely on empirical and analytical approaches over Finite Element Modelling (FEM) due to their computational efficiency and extensive literature. However, FEM holds excellent potential for future geotechnical design due to its ability to capture complex geometry, consider soil-structure interaction, and benefit from tremendous improvements in computational speed. GA integrates with the FEM software PLAXIS2D using the Python programming language and proposes an efficient method for design optimisation in this paper. It explores constrained optimisation approaches tailored to project-specific requirements and facilitates the exploration of various design cases, presenting results in a four-dimensional interpretation. The advantages and drawbacks of the methodology are highlighted through two case studies. This approach aims to advance the integration of AI-driven optimisation methods among geotechnical engineers, paving the way for efficient and sustainable geotechnical solutions.