Abstract

In this paper a method for optimal location of multi-hubs in a complex network with a large number of nodes is presented. The method is applied to the design of a logistics network composed of many tailings dams and mineral processing plants and combines two data mining techniques, K-Medoids and k-Means, with the multi-criteria decision making model PROMETHEE for the prioritization of nodes to be included into the clusters, based on certain technical and economic decision variables (such as the content of recoverable metals and the costs of transportation). The proposed method contributes to solve a large scale mathematical problem difficult to handle due to the number of variables and criteria. A case study for the recovery of abandoned deposits of mining waste is presented. The case study demonstrates the feasibility and usefulness of the proposed solution and lays the groundwork for further research and other applications of machine learning techniques for big data in support of sustainable production and a circular economy.