A multi-stage constraint programming approach to solve clustering problems in open-pit mine planning

The open-pit mine sequencing considering blocks with precedence is an NP-hard problem, which can be subdivided into long-, medium- and short-term plans, and requires different information and constraints in each stage. Through the aggregation of blocks into mining cuts, the size of the mine sequencing problem can be reduced and operational constraints can be added. In this study, a multi-stage constraint programming approach to tackle the mining cut clustering problem through a mixed integer linear programming model is proposed, as well as a geometric propagation heuristic to refine the solution. Unlike previously published studies, this approach optimizes the assignment of blocks to clusters and corrects their boundaries considering the size of the mining equipment. The methodology was validated on a real gold-ore data set. Feasible solutions were obtained in an acceptable computation time, while solutions which allowed more clusters increased their objective function and profit by up to 60%.