Mineral Supply Chain Optimisation under Uncertainty Using Approximate Dynamic Programming

The optimisation of mine complexes and related value chains is a challenging problem due to the simultaneous presence of a highly-dimensional decision space, integer decision variables, non-linear constraints and recovery functions, as well as geological and market uncertainty. In addition, if the uncertainty is revealed progressively, then decisions need to continuously adapt to new information, which poses additional challenges. Stochastic dynamic programming is a well-known class of methods for addressing problems where a decision-maker reacts to new information. While simplistic applications of dynamic programming using exact representations would be intractable for any realistic mine planning problem, new dynamic programming methods using approximate value function representations have been successfully implemented in the past to address problems of similar complexity. This work formulates the problem of mine complex optimisation as an approximate dynamic program and presents an application of the resulting method to a gold-copper deposit.CITATION:Paduraru, C and Dimitrakopoulos, R, 2014. Mineral supply chain optimisation under uncertainty using approximate dynamic programming, in Proceedings Orebody Modelling and Strategic Mine Planning Symposium 2014 , pp 415–422 (The Australasian Institute of Mining and Metallurgy: Melbourne).