Metaheuristic optimisation of underground mining ramp designs for cost-efficient excavation and support

This study evaluates genetic algorithm (GA), adaptive genetic algorithm (AGA), and differential evolution (DE) for optimising underground mining ramps. Using a 177-segment baseline, the algorithms were compared for cost efficiency. DE demonstrated superior performance, achieving an 11.4% cost reduction (580.9s runtime), significantly outperforming AGA (7.15%) and GA (5.5%). The analysis highlights the critical value of location optimisation, where refining ramp paths minimises fault encounters and support requirements. These findings validate that integrating geometric refinements with site-specific geotechnical constraints substantially enhances the financial viability and safety of underground operations.