Difficulties Encountered with the Interaction of a Static 4-step Macroscopic Model and a Dynamic Mesoscopic Model

Difficulties Encountered with the Interaction of a Static 4-step Macroscopic Model and a Dynamic Mesoscopic Model


HERVE Aurelie, Cerema, South West Territorial Division, GADRAT Marie, Cerema, South West Territorial Division


Difficulties encountered with the interaction of a static 4-step macroscopic model and a dynamic mesoscopic model: disaggregating origin / destination matrix on peak period, feedback loop...


To estimate the effects of dynamic traffic management measures on GHG emissions and pollutants (Toulouse case study), we used several models successively: macroscopic 4-step model, dynamic mesoscopic model, emissions model. This article focuses on the difficulties encountered in using a macroscopic 4-step model with a dynamic mesoscopic simulation model: scaling problem? compatibility problem? how to make dynamic long-term traffic demand?...

What are the difficulties in disaggregating origin / destination matrices on an evening peak period (4PM-7PM) in dynamic origin / destination matrices by 6-minute interval - which is the aggregation period of input data?
In the Toulouse case study, we first used a 4-step macroscopic model in order to predict traffic flow at each on and off-ramp on the urban motorway resulting in a O/D matrix on the motorway for the evening peak period. Then, we used a FRATAR algorithm to establish several O/D matrixes on 6 minutes intervals. The constraint for this algorithm are the traffic flow data aggregated on a 6 minutes period, assuming that we have the exact same dynamic as we have now, in the present.

Should road capacity gain provided by dynamic traffic management necessarily lead to a feedback loop between dynamic model and static model?
Flows in a 4-step model are theoretical hourly capacity of infrastructure implemented with speeds flow curves and a feedback loop process which takes into account the load on the network (iterative process).
The flow rates of a dynamic model are calculated each time with fondamental diagram and cumulative flow curves.
These two concepts are not easily linkable, especially when observing capacity reserves compared to the theoretical capacity.

First, the paper will present the context of questioning. Then, methods used in an operational study will be discussed. The point being to construct a debate around state-of-practice in multi-scaling traffic modelling.


Association for European Transport