Development of an intelligent evolution algorithm for open pit mines’ long-term production scheduling using the concept of block aggregation

Abstract

The method described for production scheduling in this study is a simultaneous use of a clustering algorithm with a genetic algorithm (GA). The aggregating algorithm presented in this study aims to control the concentration of operations and the cluster size, which is evaluated using the Silhouette criterion. The fitness function and the chromosome length in the GA have differences from the usual one. The results showed the number of binary variables in a mixed-integer linear programming model was reduced by 78.5% based on the created clusters. Although the aggregated model's net present value (NPV) is decreased by 7%, the solution time significantly dropped from 3 h to 43.1 s. Also, compared to the non-clustering block model, the aggregated block model's NPV, obtained by GA, was improved.