A Comparison of Static and Dynamic Traffic Assignment Under Tolls in the Dallas–Fort Worth Region

As the number of drivers in urban areas increases, the search continues for policies to counteract congestion and for models to reliably predict the impacts of these policies. Techniques for predicting the impact of such policies have improved in recent years. Dynamic traffic assignment (DTA) models have attracted attention for their ability to account for time-varying properties of traffic flow. A feature common to all DTA approaches is the ability to model traffic flow changes over time. A variety of formulations exists, with significant differences in how traffic flow is modeled, or in how the mathematical program is described. Simulation is sometimes used to incorporate more realistic flow in traffic models while maintaining tractability. Peeta and Ziliaskopoulos (2001) provide a comprehensive survey of DTA approaches and difficulties. While recognizing the dynamic features of traffic is more realistic, it introduces issues that are irrelevant in static assignment, such as ensuring first-in-first-out queuing disciplines. Also, significantly more input data are required because DTA models require time-dependent travel demand, rather than the aggregate figures that suffice for static assignment. Thus, it is not surprising that DTA formulations lead to complicated solutions that require a substantial amount of computation time when applied to large networks. It is natural to wonder, therefore, what justifies the added computational and data requirements. To this end, this work investigates the differences in results obtained from applying static and dynamic assignment to a large network under a congestion pricing scenario. The Dallas–Fort Worth (DFW) network used here contains 56,574 links and 919 zonal centroids. Comparisons are made of three models: traditional static traffic assignment (STA), the TransCAD approximator (an analytical, link performance–function–based approximation to DTA), and VISTA’s simulation-based DTA approach. An additional contribution is an algorithm that efficiently generates a time-varying demand profile from aggregate demand data (static origin–destination trip tables) by interpolating a piecewise linear curve. This algorithm is described, and is followed by brief descriptions of the TransCAD add-in and the VISTA model, as well as key issues that arise when attempting to compare these models with static assignment. A method to facilitate comparisons of the approximator’s results with those of static assignment is also described, as well as the DFW network results and a summary of modeling contributions and limitations.

• Record URL:
  http://www.trb.org/Publications/Blurbs/160584.aspx
• 
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/isbn/9780309113434
• Authors:
  • Boyles, Stephen
  • Ukkusuri, Satish V
  • Waller, S Travis
  • Kockelman, Kara M
• Conference:
  • Innovations in Travel Demand Modeling Conference
  • Location: Austin Texas, United States
  • Date: 2006-5-21 to 2006-5-23
• Publication Date: 2008

Language

• English

Media Info

• Media Type: Print
• Features: References; Tables;
• Pagination: pp 114-117
• Monograph Title: Innovations in Travel Demand Modeling: Summary of a Conference. Volume 2: Papers
• Serial:
  • Transportation Research Board Conference Proceedings
  • Volume: 2
  • Issue Number: 42
  • Publisher: Transportation Research Board
  • ISSN: 1073-1652

Subject/Index Terms

• TRT Terms: Algorithms; Alternatives analysis; Congestion pricing; Dynamic traffic assignment; Time dependence; Traffic assignment; Traffic flow; Traffic models; Travel demand
• Identifier Terms: TransCAD (Computer program); VISTA (Computer program)
• Geographic Terms: Dallas-Fort Worth Metropolitan Area
• Subject Areas: Highways; Planning and Forecasting; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01121608
• Record Type: Publication
• ISBN: 9780309113434
• Files: TRIS, TRB
• Created Date: Feb 13 2009 4:26PM