A scheduling algorithm for open pit mines

An open pit (opencast) mine can be described by a three-dimensional array of blocks, each of which is assigned a number of values defining its characteristics. Scheduling an open pit consists in finding a sequence in which the blocks should be removed from the mine in order to maximize the total discounted profit from the mine subject to a variety of technical and economic constraints. This paper proposes to model the mine-scheduling problem as one of sequential optimization, and develops an algorithm for its solution. To overcome the difficulty caused by an extremely large number of states in the problem at hand, we consider a technique which is related to dynamic programming but avoids the complete enumeration of the state space. Our algorithm is a combination of this technique with powerful heuristics derived from the specific properties of open pit mining.