O presente artigo objetiva determinar os modelos de previsão de demanda que se adequam com maior precisão a manufatura de ônibus do Brasil. A fim de facilitar a visualização, separou-se o conjunto conforme os quatro principais tipos de carrocerias produzidas nacionalmente, dos quais foram submetidos a cada uma das sete técnicas de previsão, sendo escolhidos, segundo as menores estimativas de erro, o Amortecimento Exponencial Simples (Urbano), o Amortecimento Exponencial Simples com Taxa de Resposta Adaptativa (Rodoviário e Intermunicipal) e a Média Móvel (Micro-Ônibus), que permeia os períodos compreendidos entre Janeiro de 2010 a Março de 2013.Palavras-Chaves: Previsão de demanda, Manufatura de ônibus, Competitividade, Gestão estratégica.Determination of Demand Forecasting Techniques for the Buses Manufacture in BrazilAbstract: Faced with a scenario where the current global strategic management has a crucial role in the conduct of organizations, the demand for forecasting-related techniques are essential to check at the moment the manner in which the company will behave in the future. To this, this article aims to determine the demand forecasting models that adjusts with better precision for the buses manufacture in Brazil. The justificative of this approach is related to the necessity of applications that aid in the establishment of goals and prognostics more consistent to the reality if compared to the traditional methods of verification, seen its intrinsic complexity, considered highly customized, beyond helping in the moment of formation the inventory of feedstock and inputs to the future periods. Therefore, it can be affirmed that the arrangement of the subjects proposed in this research regarding the thematic can be treated as an inedit issue in analogy to the conjuncture addressed. In order to facilitate the visualization of market behavior, the data was splitted in a set of four main types of national products manufactured (Urban, Road, Intermunicipal and Micro buses). These were submitted to the seven forecasting techniques approached in this study, Simple Moving Average, Double Moving Average, Weighted Moving Average, Simple Exponential Smoothing, Adaptive Response Rate Simple Exponential Smoothing, Brown's Double Exponential Smoothing Method, and the Holt's Method, chosen according their ability of modeling the historical series. The selection of the one that better adjusts the reality found was conceived in accordance with the lowest estimative of error found for each measures proposed. By consequence, the Simple Exponenteial Smoothing (Urban), Adaptive Response Rate Simple Exponential Smoothing (Road and Intermunicipal) and the Moving Average (Micro-bus), submitted to a database that permeates the months comprised between January 2010 and March 2013, were indicated as the most propense to generate better estimatives to the context approached. In relation to the objective, it is possible to affirm that the present article has accomplished the proposed, where the models of demand forecast chosen in each segment have adjusted in a reliable way to the reality, what turns possible in the future the achievement of a more accurate projections regarding to the context of sales in each one of the cases. As future expectations, it is waited the application of the models selected in the next demand forecastings for the national bus productions, in order to detail the presented research to each one of the companies addressed, beyond the continuous seek por other techniques able to deliver a better accuracy in the results.Keywords: Forecasting, Bus manufacturing, Competitiveness, Strategic management.
Read full abstract