Simulation and optimization techniques for short-term mine production scheduling

Abstract

Short-term mine production schedules aim to meet goals previously defined by the long-term mine production schedule. Unfortunately, the complexity and uncertainty of a mine operation cause deviations from the short-term schedules. A deviation from a schedule corresponds to any difference between this schedule and its execution; for example, deviations in the movement of material, ore sent to the plant, or in the ore grade sent to the plant. One of the challenges that mine planners face is the consideration of operational uncertainty in the generation of short-term mine production schedules that can be reproduced in reality. A concept that quantifies the deviations in a short-term production schedule and its execution is known as adherence, which corresponds to the schedule capability to be reproduced in reality. The objective of the thesis is develop a simulation-optimization framework with the objective of generating short-term mine production schedules with high adherence through an iterative approach. The main contributions of this research are: (i) development of a generic methodology which integrates optimization and simulation techniques in an iterative scheme, that generates short-term mine production schedules with high adherence, validated by a real case study; (ii) formalization of the concept of adherence of a schedule, proposing several adherence indicators to measure the deviation between a schedule and its corresponding simulation; (iii) evaluation of the adherence of a schedule through simulations considering operational uncertainty; (iv) development of an optimization model to perform short-term mine production schedules in open-pit mines, considering multiples objectives, using the weighted sum and the hierarchical method. The model also support the allocation of shovels to mining fronts and stockpiles. The outcomes of the work presented in this thesis demonstrate that the proposed framework improved the mine schedule adherence indicators over iterations and simultaneously maintained the Net Present Value of the mine schedule. The results also show that the weighted sum and the hierarchical method are capable of generating short-term mine schedules by optimizing the various short-term objectives. Finally, the results reveal that schedules with a stockpile obtain higher schedule indicators compared to the ones with no stockpile. They also demonstrate that schedules with a mobile shovel fleet obtain higher schedules' adherence indicators than the ones with a fixed shovel fleet. This thesis reveals the importance and impact of multiple objective optimization methods for the generation of short-term mine production schedules in open-pit mines. It also demonstrates that the simulation provides a better understanding of the impacts of the operational uncertainty in short-term mine production schedules.