The optimization of cut-off grades by means of memetic algorithms

Cut-off grades optimization is a fundamental issue for mineral deposits. A cut-off grade is any grade that is used to separate two courses of action; to mine or not to mine, to process or to dump. In order to achieve the maximum discounted cash flow, generally a decreasing order of cut-off grades schedule takes place. Variable mining costs are applied to the extracted material, not to all of the depletion rate as some of the depletion can be left in-situ. Because of access constraints, some of the blocks that have an average grade less than the determined cut-off grade are left in-situ, some of them are excavated and dumped as waste material. The probability density function of an exponential distribution is used to find the portion of the material below the cut-off used that is left in situ. The parts of a mineral deposit that are excavated but will be dumped as waste material and tailings of ore incur some additional cost of rehabilitation. The method of memetic algorithms is a very robust optimization tool. It is a step further from the genetic algorithms. The crossover, mutation and natural selection behavior of the method ensures it escape from a local optimum point, and a further local search improves the optimum further. This paper describes the general problem of cut-off grades optimization, outlines the use of memetic algorithms in cut-off grades optimization and further extension of the method including partial depletion rates and variable rehabilitation cost. This paper is the first application of memetic algorithms to cut-off grades optimization in this context.