Previsão de demanda por gás natural veicular: uma modelagem baseada em dados de preferência declarada e revelada

Abstract

The use of discrete choice models isa effective method of portraying the consumers' behavior in several markets. Its application has been thoroughly reported in the specialized literature and it is largely recognized asa good tool to reveal important market features of products and services in the transportation area. When demand forecast studies are necessary, the most adequate procedure is the mixed use of stated preference (SP) and revealed preference (RP) data. The combination of these data leads to more consistent statistical models, comparing to those estimated with only SP or RP data. In that way, the present research applies a methodology based on discrete choice methods using both SP and RP data. It is named GNVPREV and its objective is to analyze the consumer’s preferences, concerning the choice of fuel and considering the Natural Gas Vehicles – NGV in a competitive context in the vehicular energy market. Such analysis was restricted to users of light vehicle (cars, pickups and vans) that currently use gasoline, alcohol or NGV.The GNVPREV methodology was applied in a part of the central district of the city of Caucaia, situated in the Metropolitan Area of Fortaleza, State of Ceará. The data survey, using SP and RP questionnaires, provided information for utility functions estimation, substitutions patterns, trade-off between alternatives and demand forecast scenarios. The results were satisfactory, even considering the limited availability of primary and secondary data. They confirmed a better performance of the model when combined SP and RP data are used.