Fleet Management Systems, Operations Research, Stochastic Optimization, Mine Planning, Operations Control, Dispatch Systems, World Mining Congress 1. INTRODUCTION The management of a fleet of trucks and shovels in an open pit mine is complex and often only partially understood. As the backbone of the material transportation, the efficient use of the fleet not only affects the cost of transportation, which can be up to 60% of the total mining costs (Alarie & Gamache, 2002), but also guarantees the continuity of the production plan. As such, breaking down the details of the transportation process to identify its core components is essential to improve the economic and production performance of the dispatch activity. The study of dispatch systems traces its origins back to the start of the Mining Industry 3.0 in the 1970’s when the introduction of computerized automation was first tested in open pit mines to enhance efficiency and reduce streamline costs. Starting with the implementation of White & Olson’s work (1986) as a baseline methodology, the development of multiple approaches and strategies to address fleet-related complexities has advanced significantly over the last 45 years. All the research work is now compiled in the existing Fleet Management Systems (FMS), that integrate in a single specialized software all the routines required for the control of a mine fleet, including functioning algorithms, status reports, data transfer, and many other ancillary services. In the context of the complexities of a mining operation, the degree of uncertainty is considerable across all aspects of the process. This uncertainty in the data and orebody models creates an accuracy risk, which can lead to incorrect production plans that, ultimately, leads to generating multiple scenarios upon which to decide. To simplify matters, mine planners rely on known information, from benchmark parameters to nominal data. Therefore, most production planning calculations are made based on average values, including fleet management estimations. Even with similar truck characteristics, the fleet performance can be chaotic once exposed to a simulation analysis (Ahumada et al., 2020). Contrary to long-term schedules, short-term plans are often more sensitive to these changes and, unless they are properly assessed and time-scaled (Levinson & Dimitrakopoulos, 2023), the impact of variable information may compromise the attainability of the daily plans. This uncertainty has different degrees of criticality, depending on the process under evaluation, from the accuracy of the reserve model (Kumar & Dimitrakopoulos, 2021) to the performance of the machinery (Moradi-Afrapoli & Askari-Nasab, 2019), machinery breakdowns (Paduraru & Dimitrakopoulos, 2019), economic conditions (Pirbalouti & AskariNasab, 2023), primary crusher availability (Bakhtavar & Mahmoudi, 2018), or even weather cycles and environmental difficulties where applicable. The consequences of choosing to include or exclude the uncertainty factor in the operational plans are directly correlated with a risk management strategy. Previous work in stochastic optimization for fleet management in mining activities includes multiple approaches to different aspects of the problem. Nevertheless, most of these approaches are limited in scope, including truck and shovel short-term decisions as part of the scheduling optimization. Villalba & Dimitrakopoulos (2016) presented an assessment on short-term planning and fleet utilization under different degrees of uncertainty to generate well-informed production schedules. Moradi-Afrapoli and Askari-Nasab (2018) reviewed the impact of stochasticity when
RkJQdWJsaXNoZXIy MTM0Mzk2