Previsão De Demanda Intermitente: Uma Investigação Empírica Na Cadeia De Sobressalentes Da Marinha Do Brasil

Abstract

The present work intends to identify intermittent demand forecasting methods best suited to spare parts of the Brazilian Navy, in order to contribute to the modernization project of the Sistema de Informações Gerenciais do Abastecimento (SINGRA), currently underway. Thus, a comparison was made between the methods currently used and the Simple Exponential Smoothing (SES), Croston and Syntetos-Boylan Approximation (SBA) methods. The Mean Absolute Scaled Error was the parameter used in this comparison, which was based on