Aplicação do método ARIMA como ferramenta de gestão de controle de qualidade de matéria-prima em inventário

Abstract

The quality control of raw materials is a fundamental part for the proper functioning of an industry. Materials that remain stored for a long time end up falling outside the required specification standards and, consequently, losing their added value. In addition, these materials generate costs related to transportation and storage, and carry a capital that could be directed to other investments. In this work, the objective is the management of the raw material production of adhesives in a company that meets the demand from several European countries, being a market leader in this sector. To this end, mathematical modeling was performed through the Box-Jenkins method (ARIMA), using historical data of material consumption, to predict possible materials that will go out of specification. Manual and automatic modeling, in Python language, of this method was done to compare the effectiveness of each method. The methodology of the work consisted of a systematic review of the literature, followed by the study of the ARIMA method. Soon after, the manual and automatic stages of the models for predicting the depreciation values for the 12 months following the last months present in the data were implemented, and the results obtained were submitted to statistical correlation tests with the original raw values. In the comparison, the manual method presented a Root Mean Square Error (RMSE) of 11.18%, while the automatic method obtained 12.81%. It was noticed that both models developed were close to reality and would possibly allow the minimization of the costs associated with the control of the raw material used, reducing the generation of undesirable products (out of specification).