| dc.description.abstract | It is customary in transport economics to build models to predict performance of or to optimize urban transport systems. These models, in addition to demand models, require service level models that describe speed or delays of vehicles depending on traffic conditions. As the number of variables that are analyzed has grow, the speed functions used have become more and more ad-hoc, without real foundation on empirical estimation or theoretical relations. Since the traffic data that would be needed for empirical estimation is not available given the number of variables that are relevant and the variability needed, what we do is use a traffic microsimulator to create a very large pseudo-data set for bus corridors. We then use this pseudo-data to estimate bus corridor speed functions considering many relevant right-hand variables, such as frequency, the number of people boarding/alighting, spacing between bus stops, payment technology, distance between stoplights and bus size. We then show how the function can be used in a simple economic transit optimization model á la Jansson, to show its performance, flexibility and impact on transport policies appraisal.
Keywords: Servicel level model, speed-flow relation, microsimulator, transit optimization | |
