Capacity Estimation in Support of Mesoscopic Simulation as Part of Dynamic Traffic Assignment Models

Abstract

Dynamic traffic assignment (DTA) is increasingly being considered to model advanced strategies. Capacity is a crucial parameter in the calibration of traffic flow models utilized as part of DTA modeling. The Highway Capacity Manual has been used as the authoritative source for defining and estimating capacity in the United States. With increased traffic detector data availability in recent years, it is now feasible to locally measure capacity at bottleneck spots with different methods. This study investigates the benefits and necessity of implementing DTA in the analysis of advanced strategies, such as managed lanes, and the importance of the calibration of the associated traffic model parameters. In this regard, the importance of coding capacity based on detector measurements in DTA tools is illustrated, particularly when there is evidence that the modeled corridor capacity is lower than estimates based on the Highway Capacity Manual. The shortcomings of utilizing the traffic flow model of static assignment tools for assessing managed lanes, even when the measured capacity values are coded, are also demonstrated; this drawback illustrates the need to utilize simulation-based DTA modeling for such assessments.

Deployment and Calibration Considerations for Large-Scale Regional Dynamic Traffic Assignment: Case Study for Sydney, Australia

“…The bilevel strategy increased the share of traffic on local streets, lowering the root mean square error (a common calibration metric) of the model. Shabanian and Hadi used previous data from loop detectors to estimate the capacity at bottleneck locations (16). Their study found the evaluated capacities at some locations to be significantly different from those of the Highway Capacity Manual.…”