Abstract Manager

Simulation-based dynamic traffic assignment (DTA), and more generally, mesoscopic simulation as a modelling approach, have grown in recent years in both the diversity of the applications and the geographic scale of the models being developed. Mesoscopic simulation is motivated by a desire to apply traffic simulation to larger scale models that cannot be adequately addressed using conventional traffic (micro) simulation tools, while using mesoscopic simulation in an iterative equilibrium DTA approach allows for greater realism and sensitivities for planning applications compared to traditional travel demand forecasting (static assignment).

In this talk we provide an overview of current trends in simulation based DTA, focusing on recent advances in the underlying algorithms and summarizing a few key applications in North America.  We start with an overview of the key challenges related to scaling traffic simulation models, with a focus on model stability, run times, and calibration effort, and recent advances in algorithms that address these challenges in a systematic way, helping to reduce the cost of model development while simultaneously improving the quality of model outputs and forecasts.

This is followed by an overview of a few recent simulation-based DTA applications in North America, summarizing the motivation and added value of adopting a simulation-based approach, and discussing some of the key challenges encountered in the development of the models. These case studies include the calibration of the simulation based DTA model for the Region of York in the Greater Toronto Area, and the development of several pricing studies for the City of Portland, OR. The talk is concluded with a brief overview of emerging trends in simulation-based DTA, including regional applications and integrated Activity-Based Demand / Dynamic Traffic Assignment (ABM-DTA) models.