Demand Forecasting is fundamental to help determine the need for resources for an organization, making it a strategically important task. Moreover, a good prediction contributes to a more efficient supply chain, and is key to decision making. In this context, the performance of a company can be related to the accuracy of demand forecasting. Given such importance, this study aimed to find and propose a quantitative method of demand forecasting of time series that got smaller forecast errors for the consumption of a common bumper sticker in a graphic at Belém. First of all, the element was selected through the ABC classification. After that, quantitative methods to forecast time series existing best known in the literature have been applied to find the most assertive. Moving Mean, Simple Exponential Smoothing Method, Holt and Holt-Winters method (additive and multiplicative) are quantitative forecasting methods applied. The Seasonal Additive Winters’ Model had the lowest MAPE, but not ruled out the application of other methods.
|
