Dynamic congestion modeling

This project aimed to model recurrent congestion observed on the highway sector A50/A501, a strategic entry point into Marseille, which is particularly busy at peak times.
The model was designed for analyze the functioning of the convergent between these two branches and test the relevance of a dynamic control system with the help of lights, in addition to a future reserved lane for public transport (VRTC).

The work was based on the exploitation of SIREDO data (counts every 6 minutes) And of Floating Car Data, making it possible to characterize the speeds and flows over the entire linear path (13 km).
Two typical days have been reconstructed — Median day and Strong day — in order to assess the behavior of the network according to different load levels.

The model selected, of type Dynamic LWR, has been calibrated to accurately reproduce the observed flow rates and travel times, with a discrepancy of less than 10% on the main sections.
The tool simulates the interactions between flows from the A50 and the A501 and the propagation of congestion waves to Marseille.

The study then tested several regulatory scenarios:

• differentiated shoulder light settings (alternating green/red times);
• elasticity to travel time;
• redistribution of flows between branches;
• combination with the VRTC.

The results show that regulation allows a rebalancing travel times between the two branches, a significant fluidity between the convergent and La Penne, and a measurable reduction in hourly variability without major impact on other users.

• Understand the mechanisms for the formation and dissipation of congestion on a structuring axis of the Marseille metropolis.
• Test the relevance of a fire control system, at low development costs, to support the commissioning of a Bus/carpooling reserved lane.
• Provide a dynamic simulation tool capable of evaluating in near real time the effects of regulation on flows, speeds and the variability of travel times.
• Identify the most effective control parameters (cycles, durations, activation) for modal shift to public transport.

This modeling made it possible to:

• Objectifying congestion on the A50/A501 corridor;
• quantifying time savings for public transport using the VRTC;
• Demonstrate the technical feasibility of dynamic control at low cost, without major geometric changes;
• and provide DIR Med with a freely reusable model prototype, available as open source on GitHub.

The approach confirmed that a well-parameterized control could reduce up to 30% the variability of travel times and improve the regularity of interurban lines on this corridor.

This project was a successful proof of concept of application of the macroscopic dynamic modeling to operational traffic management.By integrating regulation, reserved lanes and elastic demand, Explain demonstrated the ability of a lightweight scientific model to inform planning decisions and to support the transition to more efficient mobility on metropolitan highways.