Previsão de demanda de curto prazo para uma indústria metalomecânica

Abstract

Forecasting is a fundamental tool for the organization when it wants to strategically position itself against the goals to be achieved. For planning it is the basis of all decision making in the short, medium and long term. This paper presents a case study in a metalworking industry where the objective is to obtain a demand forecast for the three periods subsequent to the analyzed history. Knowledge of this forecast will guide the organization's planning. The data series provided by the company evaluated using statistical software to perform the calculations through the Forecasting module. The definition of the best model is by evaluating the forecast errors calculated for each model and choosing what best represents the series analyzed. The model that best fit the presented series was the Holt Linear model with deviations calculated at MAE = 15200.9 and MAPE = 12.8589.