Traditional short-term production planning is carried out in two separate sequential optimizations, typically based on mixed integer programming formulations, the physical mine production schedule, and followed by the assignment of the available fleet. However, (a) the fleet availability, hauling time and mining considerations do not influence the production scheduling sequence; and (b) the fleet optimization algorithms do not consider uncertainty in their parameters or account for the variability of the local mineralization of the various mine sectors being considered. The above have adverse effects on short-term production scheduling performance and very large operating and opportunity costs, stemming from erroneous materials blending and processing destination decisions, sub-optimal mining equipment utilization and so on.
A new short-term mine production scheduling formulation is developed herein based on stochastic integer programming. The formulation simultaneously optimizes fleet and mining considerations, production constraints, uncertainty in both orebody metal quantity and quality, as well as fleet parameters and equipment availability, all leading to a well-informed sequence of mining that is expected to have realistic, as well as high performance during a mine's operation. To assess the latter performance, the proposed formulation is applied at a multi-element iron mine and the resulting monthly schedules are assessed and compared to the conventional mine scheduling approach showing: lower cost, minable patterns, efficient fleet allocation ensuring higher and lesser variable utilization of the fleet.
Published August 2014 , 20 pages