Open-pit design and scheduling by use of genetic algorithms

Paper presented at a meeting on: Artificial intelligence in the minerals sector, held in Nottingham, UK, 20 April 1993. The design and scheduling of open pit operations are consistent problems encountered by the surface mining industry. Numerous techniques and algorithms have been developed to assist the mining engineer in the creation of optimum development strategies. Virtually all the practical techniques that have been produced have concentrated on one of the aspects (limit definition or scheduling) at the expense of the other. It has long been recognised, however, that the two activities are intrinsically linked and thus solving either problem independently of the other can only ever produce sub-optimal results as a whole. This has been termed the "circular reasoning problem". It has also been recognised that, for all practical-sized situations, solving both problems simultaneously is virtually impossible using conventional techniques due to the massive size of the search space involved. A technique is presented which is based on a self-learning genetic algorithm, which provides a combined pit limit and extraction schedule simultaneously in order to maximise the NPV (net present value) of the extraction. The theory is developed in two-dimensions and a number of examples of its application are presented. It is proposed that this approach is a highly efficient, heuristic technique offering the potential to transcend the circular reasoning difficulties of separate pit limit and extraction scheduling systems. Extension of the technique into three-dimensions and its application to realistic problem sizes are discussed