A Parallel Composite Genetic Algorithm for Mine Scheduling

L. Fava Lindon, D. Goforth, A. van Wageningen, P. Dunn, C. Cameron, and D. Muldowney (Canada)


genetic algorithm, parallel algorithm, village, scheduling, optimization, diversity.


Mine scheduling is a multi-objective highly constrained optimization problem. Often months are spent by mine planners to achieve one feasible ore production solution. In order to assist in the process and to present alternatives with a higher likelihood of optimality, a parallel genetic algorithm for long-term scheduling of underground mines is developed. For the mine scheduling problems considered, a 2-dimensional map of stopes is given, along with the mineral properties of each. It is required to schedule the extraction sequence of the ore from the stopes to meet the mine’s objectives. The appropriateness of a schedule is determined by applying a fitness function. The fitness function assesses how well the schedule meets objectives and satisfies given constraints. In practice, the scheduling problem is simplified in order to obtain a solution in given time bounds. By modularizing the problem and employing a parallel algorithm with minimal communication requirements, a higher quality mine schedule may be found in given time bounds.

Important Links:

Go Back