A matheuristic algorithm based on Iterated Local Search for the ore blending planning and green shovel allocation problem

This study addresses the ore blending planning and green shovel allocation problem in an open-pit copper mine, to minimize deviations in production, quality, and fuel consumption. To deal with it, an algorithm called Math-ILS was proposed, which combines an Iterated Local Search with a Matheuristic, using a mathematical formulation for the shovel allocation sub-problem. To generate a good initial solution, a constructive algorithm was proposed based on the ranking of the stockpiles and mining fronts. Five neighborhood structures were proposed to explore the search space. To avoid getting stuck in local optima, a perturbation algorithm was proposed that randomly removes ore from one stockpile or mining front and inserts it into another. To evaluate it, 10 real instances were used and the results showed that Math-ILS found solutions close to the optima in all instances.