Optimization of surface mining dig limits with a practical heuristic algorithm

An algorithm is presented for optimizing the classification of surface mine material subject to excavating constraints. High-resolution, expected-profit models are optimized to classification maps subject to site-specific excavating constraints. This optimization problem defies traditional closed-form analytical solutions. A practical heuristic algorithm quickly determines the optimum final destination for material subject to realistic selectivity constraints. The expected profit could be calculated from an estimated model if the value calculations are all linear; otherwise, simulation could be used. In both cases, a single expected profit for all destinations on a high-resolution grid is the starting point for dig limits optimization. Maximizing expected profit in a risk-neutral manner is correct given the repeated nature of grade-control decisions over relatively short timeframes.