Optimización Multi-Echelon del inventario basado en pronósticos de la demanda por el método Random Forest.

Abstract

The need to optimize resources and develop strategies that allow better synergy within the entire supply chain has led both academia and companies to investigate and implement new methodologies that respond to these realities. This research project allows, based on the study of x the data of the company Ambacar focused on the most important brand of the company and the 50 most important spare parts, to implement the Machine Learning Random Forest methodology for the forecast of demand, thus achieving a better understanding of the market behavior in the spare parts area at each of the network's dealers, taking into account the most relevant customer factors and external effects such as those caused by the COVID pandemic. Based on the results obtained from the demand forecast, the inventory optimization philosophy for each dealer is changed to a philosophy that encompasses the entire system, thus achieving that the relationship and interaction of each of the workshops with the rest is considered. of the network. All of this is achieved without negatively affecting service levels and maintaining high-quality standards.