A simulated annealing based approach for open pit mine production scheduling with stockpiling option

Abstract

Production scheduling plays a pivotal role in successfully executing any open-pit mining operation. It defines the most profitable extraction sequence of mineralized material from the ground subject to various physical and operational constraints. Different mathematical formulations have been proposed to achieve this goal; however, solving these models for real-sized deposits with multiple constraints is a challenging and computationally expensive task. Moreover, the inclusion of stockpiling option further complicates this task. The stockpile option adds flexibility to the operation by allowing excess low-grade ore storage for processing at a later stage when processing capacity is available. Accurate integration of stockpiling option in the production scheduling process through mathematical approaches leads to nonlinear constraints. This could further complicate the already challenging task since linear approximation of these nonlinear constraints could lead to sub-optimal results. Metaheuristic techniques could play an essential role in handling such situations. Though several attempts have been made to solve this problem through these techniques, little effort has been made to incorporate stockpiling option in the optimization process. This article presents a Simulated Annealing based approach for production scheduling of open-pit mines with stockpiling option. The proposed approach uses a stockpile and a greedy heuristic with a Simulated Annealing algorithm to achieve this goal. The greedy heuristic improves the Simulated Annealing algorithm's computational efficiency by managing its randomness. The proposed approach's performance and efficiency are demonstrated through three case studies (A, B, and C) under different algorithmic settings. Results of these case studies reveals that compared with the CPLEX solver, the proposed approach produced near optimal solution, within reasonable amount of time, proving the applicability of the proposed approach.