Evolutionary approach for large-Scale mine scheduling

Finding the optimum solution for open-pit mine scheduling problems (OPMSPs) is a well-known challenging problem in the mining industry. The existing methodologies for solving large-scale OPMSPs are mostly limited to conventional optimization approaches and heuristics. However, due to the problem’s complexity, represented by high-dimensionality, and hard and soft constraints, the current approaches face challenges in finding good quality solutions with a reasonable computational cost. As an alternative approach, in this paper, we propose an evolutionary approach, based on differential evolution, with three important features: to deal with high-dimensionality, based on the extraction status of each block, we reduce the number of decision variables (blocks) over the planning horizon; to deal with the complex constrained landscape, we develop a repair mechanism that guarantees feasibility and to enhance the algorithm’s performance, we incorporate a local search technique. The experimental results on well-known mine deposits, with up to 112, 687 blocks show that the algorithm has the edge over existing methods to obtain better solutions.