Development and Application of a Regional DTA model for the Twin Cities
Corresponding Author: Jim Henricksen, Minnesota Department of Transportation
Presented By: Steve Ruegg, WSP
The Minnesota Department of Transportation is anticipating several major near-term construction projects in the Twin Cities Metropolitan area and desired of a more detailed assessment of impacts related to construction phasing alternatives and combined impacts of multiple projects involving partial and complete closures of major freeways in the region. In addition, there was a need for better tools to assess operational impacts of metro area roadway operations in general. A dynamic traffic assignment (DTA) model was developed to serve these purposes. Based on the Dynus-T software platform, the entire 13 county region of the greater twin cities area was modeled, matching the modeled area of the current static regional travel demand model. The meso-scale model uses link-based delay and queuing models to more accurately reflect the impacts of capacity constrictions due to construction. Results include detailed traffic flows by minute for a full 24-hour period.
The network was developed from the regional model, and trip tables were also obtained from the regional model with more detailed time of departure distributions applied.
Construction scenarios for I-35W at Lake Street included full closure, and three partial closure scenarios, each with two sub-phases. The model provided comparative performance measures for each scenario, and informed construction strategies. Diversion from the construction corridor was also described.
The model was successfully applied to several other planned construction projects as well.
The development of the DTA model provides a new, valuable tool for metro area planners to evaluate a variety of construction impacts as well as more general impacts of future growth and MnPASS (toll managed lanes) proposals. The model is capable of receiving travel demand data from the regional model and providing updated travel times to the regional model to improve mode choice and distribution forecasts. The model was developed to interface with both the traditional 4-step model as well as the region’s activity-based model.