Highway case study investigation and sensitivity testing using the Project Evaluation Toolkit

• Main Author: Fagnant, Daniel James
• Format: Others
• Language: English
• Published: 2011
• Subjects: Transportation; Planning; Project evaluation; Traffic forecasting; Sensitivity testing; Monte Carlo simulation
• Online Access: http://hdl.handle.net/2152/ETD-UT-2011-08-3923

As transportation funding becomes increasingly constrained, it is imperative that decision makers invest precious resources wisely and effectively. Transportation planners need effective tools for anticipating outcomes (or ranges of outcomes) in order to select preferred project alternatives and evaluate funding options for competing projects. To this end, this thesis work describes multiple applications of a new Project Evaluation Toolkit (PET) for highway project assessment. The PET itself was developed over a two-year period by the thesis author, in conjunction with Dr. Kara Kockelman, Dr. Chi Xie, and some support by others, as described in Kockelman et al. (2010) and the PET Users Guidebook (Fagnant et al. 2011). Using just link-level traffic counts (and other parameter values, if users wish to change defaults), PET quickly estimates how transportation network changes impact traveler welfare (consisting of travel times and operating costs), travel time reliability, crashes, and emissions. Summary measures (such as net present values and benefit-cost ratios) are developed over multi-year/long-term horizons to quantify the relative merit of project scenarios. This thesis emphasizes three key topics: a background and description of PET, case study evaluations using PET, and sensitivity analysis (under uncertain inputs) using PET. The first section includes a discussion of PET’s purpose, operation and theoretical behavior, much of which is taken from Fagnant et al. (2010). The second section offers case studies on capacity expansion, road pricing, demand management, shoulder lane use, speed harmonization, incident management and work zone timing along key links in the Austin, Texas network. The final section conducts extensive sensitivity testing of results for two competing upgrade scenarios (one tolled, the other not); the work examines how input variations impact PET outputs over hundreds of model applications. Taken together, these investigations highlight PET’s capabilities while identifying potential shortcomings. Such findings allow transportation planners to better appreciate the impacts that various projects can have on the traveling public, how project evaluation may best be tackled, and how they may use PET to anticipate impacts of projects they may be considering, before embarking on more detailed analyses and finalizing investment decisions. === text