Incorporating realistic congestion effects into large scale strategic transport (demand) models
Incorporating realistic congestion effects into a large scale strategic transport (demand) model. Research focuses on (1) development of a semi dynamic assignment model (called STAQ) that incorporates flow metering and spillback effects resulting from congestion, and that is capable of describing traffic behaviour on large scale transport networks containing both highways as well as urban areas; and (2) incorporating STAQ into large scale transport demand models by developing a compatible matrix estimation method and computationally efficient methods to calculate the converge to the user equilibrium.
- Develop computation efficient methods to incorporate a realistic traffic assignment model into both the demand (matrix estimation) and supply (traffic assignment) side of strategic transport models.
- Find the optimum between model realism and model usability as an instrument for strategic transport planning.
- Exploit Big data (observed travel times, network states, destination choices, route choices) that has become more widely available recently.
Strategic transport models are used to assess effects of proposed policy measures to enable policy makers to choose between these measures before implementing them. In the vast majority of studies in which these models are being used the effects that road congestion have on the level of service, travel times and environmental emissions on the network play an important role. However, almost all strategic transport model systems to date still use static traffic assignment models that are not capable of describing the effects of congestion adequately and are known to provide large underestimation and sometimes counter-intuitive indirect effects in study areas with structural congestion levels.
Dynamic assignment models are capable to adequately describe the congestion effects, but suffer from a poor scalability, mathematical tractability and a high computational burden. The (less known) class of semi-dynamic assignment models (combining features of both static and dynamic) can potentially overcome all of the drawbacks of dynamic models, thereby enabling policy makers to safely develop policy in areas with structural congestion based on these assignment models.