A Aplicação de modelos estatísticos baseados na previsão da demanda: o problema da linha Cidade Universitária

Abstract

The question regarding mobility in urban areas is an important subject in discussions that aim to create efficient ways to take people from a place to another, considering different transportation means and mobility strategies. This way, the objective of t his study is to make use of time series to forecast the number of commuters riding the Cidade Universitária bus service route in the city of Dourados. Therefore, a case study was developed along with the public transportation operator in the city that aimed to understand better the costumers’ behavior and also provide a forecast model for commuters demand for the three following years. The statistical analysis showed seasonality in the series of data, but it did not show trend. Through the ratio -tomoving-average method, it was obtained the seasonal indices for each four period and, in order to obtain the forecast for the following three years, the Holt-Winters method was used. It could be verified that for the demand curve behavior of the studied bus route, the Holt-Winters method showed to be the best option as a demand forecast technique and also a good decision making tool for bus fleet management plan.