Modeling Reliability Benefits

Abstract

This article compares the performance of several models forecasting travel time variability for road traffic using before and after data from the introduction of the Stockholm congestion charges. Models are estimated on before-data, and the models’ forecasts for the after-situation are compared to actual aftermeasurements. The accuracy of the models varies substantially, but several models can be used to forecast the benefits from reduced travel time variability with sufficient accuracy to make them useful for decision-making.