A sustainable solution for soil improvement: a decision tree model combined with metaheuristic optimizations for fiber reinforced clays

The rapid urbanization has enhanced the demand for innovative solutions to improve soft soils, prompting the integration of fibers as a sustainable reinforcement strategy to enhance tensile strength. However, effective utilization hinges on the optimal selection of fiber type, soil composition, and environmental conditions. Recognizing the complexity of these interactions, this study employs machine learning to analyze a comprehensive dataset comprising 858 samples derived from laboratory tests, encompassing 19 factors that influence soil-fiber behavior. Utilizing advanced meta-heuristic optimization algorithms—Equilibrium Optimizer, Marine Predators Algorithm, and Manta Ray Foraging Optimization—alongside a decision tree model, the research aims to uncover optimal reinforcement strategies for fiber-reinforced clays. The results reveal that basalt fiber emerges as the most effective reinforcement, while the study highlights the limitations of fiber reinforcement in high-plasticity clays. By addressing a notable gap in the literature regarding the application of optimization techniques to such extensive datasets, this research not only advances understanding in soil-fiber interactions but also promotes environmentally sustainable approaches to soil reinforcement, paving the way for future innovations in geotechnical engineering.