Comparison between three different traffic micro-simulations and reality in Dallas

It is certainly desirable that transportation forecasting models are correct in the sense that the traffic patterns they predict correspond to what would happen in reality under the circumstances assumed in the forecasting model. Unfortunately, it is notoriously difficult to transform the above common sense statement into a technical specification. Since one cannot run controlled experiments in socio-economic systems, it is usually impossible to check the forecasts. The authors describe three traffic microsimulations which operate at different levels of fidelity. They are used to iteratively generate a self-consistent route-set based upon microsimulation feedback. They compare the simulation results of all three simulations to aggregated turn count data of actual field measurements.