Determining Optimum Inventory of Bauxite 4-8 (80%) Based On Forecasting Demand Using the Long-Short Term Memory Method

Abstract

1. Refratech Mandala Perkasa is a company that operates in the manufacturing of refractory products. One of the materials used to produce the product is Bauxite 4-8 (80%). Based on historical data on inventories and demand from January 2012 to January 2020, it is known that the raw materials have been overstocked. Hence, the inventory analysis is needed by forecasting the demand and determining the optimum inventory of raw materials in order to minimize cost incurred by the company. In this study, the method that used is Long-Short Term Memory. Based on the result of this study, the best model to forcast the demand inventory has a MAPE value of 13.2575%. The forecast results for the bauxite 4-8 (80%) raw material in the next 13 weeks is 67.774,66401 kg. Since the demand is not fixed during the lead time (probabilistic), the method that used to determine the optimal inventory of raw materials is EOQ probabilistic method and an optimal inventory of raw materials is 67.280,2195 kg with total costs for the next 13 weeks is IDR 34.167.024,82.