A local branching heuristic for the open pit mine production scheduling problem

Abstract

This paper considers the well-known open pit mine production scheduling problem (OPMPSP). Given a discretisation of an orebody as a block model, this problem seeks a block extraction sequence that maximises the net present value (NPV) over a horizon of several periods. In practical applications, the number of blocks can be large, and therefore, the problem can be difficult to solve. It is even more challenging when it incorporates minimum resources requirements that are represented as lower bounds on resource constraints. In this study, we propose to tackle OPMPSP by using a novel metaheuristic technique known as local branching. To accelerate the search process, we combine local branching with a new adaptive branching scheme, and also develop a heuristic to quickly generate a starting feasible solution. Despite consideration of minimum requirements being seldom taken into account in the literature, this method yields near-optimal solutions for a series of data sets we have conceptually generated. To judge the performance of our methodology, the results are compared to those of two techniques from the literature, as well as to those obtained by a mixed integer linear programming (MILP) solver.