Diseño de un sistema de gestión de inventarios para una compañía comercializadora de repuestos del sector automotriz

Abstract

El siguiente estudio está enfocado al desarrollo de un diseño de modelo de gestión de inventarios para una compañía comercializadora de repuestos del sector automotriz, la cual busca mejorar la disponibilidad de partes. La investigación se basó en la revisión de literatura de modelos de clasificación, gestión de inventarios y sistemas de ordenamiento que han sido estudiados frente a la misma necesidad. Se aclara, que por el alto número de ítems del inventario (26,000 productos), se tomaron unas muestras para revisar la factibilidad y metodologías a desarrollar en una posible implementación a escala real. Con respecto a la clasificación de inventarios, se plantearon dos tipos: clasificación para control del inventario y clasificación para generar pronósticos. En la clasificación para el control del inventario se propuso un modelo multicriterio basado en el Principio de Pareto y el Proceso de Análisis Jerárquico (AHP). Cómo resultado de la Clasificación Multicriterio ABC, se identificaron los ítems relevantes, asociados al grupo A multicriterio (3,431 productos) en que la compañía debe enfocar sus esfuerzos para aumentar las ventas. Además, para el grupo A multicriterio se encontraron, las mayores líneas y unidades pendientes de despacho por falta de disponibilidad (43.1% en líneas y 66% en unidades), lo que indica la necesidad de enfocar acciones para este grupo principal de inventario. Se estima que, por la falta de disponibilidad de este grupo, la empresa dejó de facturar en el mismo periodo, un 4% mensual de las ventas totales.

This research proposal was focused on the design of an inventory management model for a company of the automotive sector that seeks to improve the availability of spare parts. This study was based on the literature review of classification models, inventory management, supply systems, which have been studied for the same need. It is clarified that due to the high number of ítems in the inventory (26,000 products), samples of ítems were taken to review the feasibility and methodologies to be developed in a possible full-scale implementation. For the inventory classification, two types were generated: classification for inventory control and classification for generating forecasts. In the classification for inventory management, a multi criteria model, based on the Pareto Principle and the Hierarchical Analysis Process (AHP) was proposed. As a result of the ABC Multicriteria Classification, the relevant ítems, in which the company should focus its efforts to increase sales, were identified. The new classification, shows that the multicriteria group A, had the largest lines and units pending dispatch due to the lack of availability of spare parts (43.1% in lines and 66% in units); this indicates the need to focus actions for this main inventory group. It is estimated that, due to the lack of availability of this category, the company stopped billing in the same period, 4% per month of total sales. The second classification apply for forecasting, where the parts of the inventory were classified in one of the 4 demand patterns (Smooth , Intermittent, Lumpy and Erratic). Subsequently, for each group, different time series models were evaluated using R-Studio program and the forecast package: Croston, Simple Moving Average, Holt-Winters, ETS, ARIMA and Neural Networks. With the Average Absolute Error (MAE), it was defined the best the technique for each of the groups. For the group "A" the Neural Networks method was used, for "B", the ETS and for group "C", Simple Moving Average. The forecast results, including the errors, were implemented in the periodic model (R,S) to calculate the units to order in the supply system. By using the periodic model (R,S) it was obtain lower values for the maximum points of inventory, on about 42% average of the ítems analyzed; this would lead to a decrease in the value of the inventory of the company.