Sistema de suporte à decisão para priorização de investimentos em rodovias utilizando inteligência geográfica

Abstract

The lack of funds affects mostly all government sectors which, in order to fulfill its institutional functions, needs to make investments. This is a quite typical scenario faced by transportation agencies responsible for highway construction, maintenance and operation. Regarding the role of the road transportation, where a road allows to flow thousands of vehicles flowing, it is imperative to affirm that even minor improvements in the infrastructure cause major economic impacts by reducing the transportation cost, what indeed explains the importance for prioritizing investments in roads. However, there is a lack of planning for the road infrastructure in Brazil, once some of the road that have the worst pavement condition are located in the most relevant geographic areas, like the Triangulo Mineiro region or the areas with mining activities. This investigation addresses a methods for identifying, classifying roads segments candidates for receiving funds for maintenance and recovering programs for paved roads under jurisdiction of the State of Minas Gerais - Brazil based on geographic intelligent modeling. In order to assist the investigation, robust GIS tools were used compute road sinuosity, rise&fall, and the horizontal and vertical road curvatures, required in the estimation of the flow speedy. The origin-destination matrix was estimated, as well as the traffic flow using the software Aimsun. It was also computed geographic-contextual variables from spatial interpolation, Euclidian distances and density operator (Kernel). Based on these geographic-continuous variables four different scenarios were proposed: socioeconomic perspective, logistic perspective, traffic flow perspective and the perspective from technical criteria for pavements. Experts were consulted for weightening by using the Delphi method. The road segments with volume of traffic were overlaid to the maps produced. This allowed the identification and ranking of the critical roads through a simple geographic intersection. The model demonstrated to be robust regarding the solid algorithms used, as well as demonstrated to be flexible as per the scenarios, rules and weighting strategies. Findings proved the hypothesis of a geographic-based multicriterial analysis framework allows to identify and classify road segments which infrastructure requires priori investments for maintenance, what is a key strategic tool for the decision making process.