Development of Multi-similarity index Clustering algorithm in Mathematical Modelling of Mines

Generating a production schedule that will provide optimal operating strategies without all technical and operational constraints is not practical. Creating such models with considering NPV as objective function result to oversize mathematical problems which needs more CPU time. This paper developed a multi-index clustering algorithm to reduce the size of the large-scale mathematical problems by reducing the number of decisions variables. By the way the presented algorithm remove dependency to weight importance coefficients and marks the planning with minimum error and significant reduction in the size of the model and solving time. In application and comparison of the presented clustering technique, 2478 extraction block aggregated in first step in 10 clusters, and in second step in 40 and finally in 109 clusters. By using CPLEX and MATLAB the MILP models of clusters and extraction block created to evaluate the clustering technique. The results show that CPU time has 86% reduction whereas the NPV only show 1.8% difference between clusters and block models.