Abstract:Aiming at the problem of multi-metal and multi-objective ore blending in open pit, an optimization method of ore blending based on multi-objective evolutionary algorithm was proposed. According to the actual situation of the mine, the optimization model of open-pit ore blendingwas established with the minimum production cost, ore grade deviation and ore lithology ratio deviation as the optimization objectives. Based on the decomposition based multi-objective evolutionary algorithm (MOEA/D), the updating process of the algorithm was improved, and an angle based updating strategy was proposed by using the spatial position relationship between the population and the weight vector, so that the algorithm can better balance the diversity and convergence of the population when solving multi-objective problems. Due to insufficient consideration of mineral processing factors, the comprehensive recovery rate of ore can not be effectively improved. In this paper, a random forest prediction model of comprehensive recovery rate integrating oxidation rate and harmful material parameters was established. Through the prediction model, multiple groups of ore blending results obtained by the algorithm were screened to obtain a group of ore blending plan more suitable for the actual situation of the mine. Finally, a large molybdenum-tungsten copper mine in China was taken as an example, and the experimental results showed that the ore blending plans can effectively improve the comprehensive utilization rate of ore and the economic benefits of enterprises when solving the problem of multi-metal and multi-objective ore blending.