The report evaluates methods for traffic assignment modelling (static macroscopic, dynamic macroscopic and dynamic micro/mesocopic) within strategic transport model systems. Our evaluation of traffic assignment models found that dynamic meso/micro models are most appropriate for all (considered) application purposes in congested urban areas. The biggest advantages are connected to the realistic modelling of congestion and the richness in analysis (allowing to aggregate results in any desirable way). Those models have some practical challenges/disadvantages. They require more detailed input data, are more demanding with respect to implementation, calibration and usage and set high requirements (expert knowledge) on the users. The stochasticity of dynamic meso/micro models imply that distributions of predictions from several model runs should be compared. (Author/publisher)
Abstract