Florida Department of Transportation (FDOT) Express Toll Lane Modeling Workshop Peer Review Report
4.0 Florida DOT's Charge
The Blue Ribbon Expert Panel was charged with providing advice and recommending technical guidance for a new managed lane model design, including specifications for a consistent modeling practice for express lane implementation. The need for this new managed lane model design was a direct result of the recent shift to requiring managed lanes as an alternative for analysis in the study of major capacity improvements on existing limited access facilities in the State of Florida.
FDOT developed a workshop agenda that first provided an overview of Florida's current modeling practices and procedures for analyzing managed and express toll lanes. Following a number of presentations on the specific experience to date in Florida, a handful of invited speakers expanded the discussion to identify current practices and emerging trends being employed nationally across the country.
After an intermission for lunch two technical breakout workshop sessions were convened: "Planning and Demand Forecasting" and "Operational Analysis." Roughly ninety participants in total attended the two-day workshop. Participants elected which of the breakout sessions to attend based on their own interests, background, and expertise. As a result, there was approximately a 70/30 split of total participants in the "Planning and Demand Forecasting" and "Operational Analysis" breakout sessions, respectively.
4.1 Planning & Demand Forecasting Breakout Session
This session was moderated by the following four individuals:
- Jennifer Fortunas (FDOT Central Office);
- Steve Ruegg (PB);
- Jack Klodzinski (URS/Florida Turnpike); and
- Amy Perez (FDOT District 1).
The expert panel members invited to serve during this session included:
- Scott Ramming (DRCOG);
- Matthew Kitchen (PSRC);
- Kara Kockelman (University of Texas);
- Hugh Miller (CDM Smith);
- Bill Olsen (CDM Smith);
- Eric Pihl (FHWA); and
- Jim Ely (HNTB).
The moderators led a series of discussions focusing on questions related to three different topic areas that included: Express Lanes: Planning, Policy, and Process; Model Capabilities and FSUTMS Applications; and Screening and Planning Phase Considerations.
The specific questions posed by the moderators to the expert panel and breakout session participants are listed below. There were five questions regarding Topic Area #1, five questions regarding Topic Area #2, and twelve questions regarding Topic Area #3.
Topic Area #1 = Express Lanes: Planning, Policy and Process
- Which criteria should be applied in determining when a capacity improvement may be needed within the planning horizon (i.e., the next 20 to 40 years)?
- How can the feasibility of express lanes as a strategy for capacity improvement be determined?
- Which key features and capabilities are necessary in a travel forecast model to provide a case for the feasibility of express lanes?
- If express lanes are not feasible, which other non-priced managed lane strategies may be evaluated as suitable alternatives?
- If express lanes are feasible:
- How can lane requirements be established?
- Which dynamic tolling scheme provides maximum throughput?
- Which dynamic tolling scheme provides the optimal financial solution?
- What important considerations must be evaluated to determine an Internal Rate of Return over 50 years?
Topic Area #2 = Model Capabilities and FSUTMS Applications
- Which model sensitivities and capabilities must be focused on during each stage of project development and LRTP development?
- Demand Sensitivity to Price
- Price and Demand Equilibrium
- Time- of-Day Travel Sensitivities
- Distributed Value-of-Time
- Income Market Segmentation
- Purpose Market Segmentation
- Transit Demand Sensitivities
- Multi-class Assignment
- What guidelines should be established for model application?
- Should these guidelines vary by phase and, if so, how would they vary?
- Fixed Person Demand across All Alternatives
- Assignment Closure Criteria
- Forecast Years
- Quality Assurance Procedures and Technical Review
- Network/zone Level of Detail
- Level of Detail of Reported Results
- Subarea Isolation or "Windowing" of Regional Model (size of detailed network area)
- Identify dataneeds at each of the phases below to support the development and maintenance of demand forecasting tools for express lanes:
- Development Phases
- Initial Screening
- How should thedevelopment and application of modeling tools be coordinated from initial screening through T&R forecasts to ensure consistency?
Topic Area #3 = Screening and Planning Phase Considerations
- What modeling efforts should be undertaken for "Sketch Level" studies?
- Should the data available define the model tools?
- Define potential modeling approaches/tasks (e.g., utilize and run an existing travel demand model).
- Which applications should be utilized in conjunction with travel demand models (e.g., ability to model managed express lanes)? Consider both model run times and structural complexity.
- Should additional modeling be conducted (e.g., time-of-day)?
- Is it possible for a demand model to be supplemented with an additional modeling tool (e.g.,time-of-day) to meet the defined goal(s) of the project scope?
- How much consideration should be given to modeling capabilities for "Sketch Level" studies in the initial stages of project definition?
- How are project limits and access locations identified?
- What considerations should be taken when defining a potential project?
- Should there be two stages of model capability review: one for project definition and one for feasibility analysis?
- What input data should be required for a "Sketch Level" study?
- Traffic counts, capacity (e.g., general use and express lanes), speed (e.g., posted, free flow), etc.?
- What output data should be required for "Sketch Level" project feasibility determination?
- Should multiple projects in a defined network of limited access facilities be considered?
- Which considerations should be taken in defining output? Examples include: the defined goal(s) of the managed lanes project(s), administrative directives (i.e., all new capacity should consider tolling), congestion management, filling funding shortfalls, covering operations and maintenance (O&M), schedules, etc.?
- Which output aside from toll rates and traffic volumes are necessary?
- What should function as the determining factor for whether a project should graduate to the next phase of analysis?
- Which considerations should be taken in determining the phase of analysis? Examples include: the defined goal(s) of the managed lanes project(s), administrative directives (all new capacity should consider tolling), congestion management, filling funding shortfalls, covering O&M, schedules, etc.?
- How can project feasibility be determined from a "Sketch Level" evaluation?
- What modeling efforts should be undertaken for "Planning Level" analysis?
- Define potential modeling approaches/tasks (e.g., project level validation of demand model, refinement of access points).
- Should available model tools define data requirements? (e.g., if more than one model is available and one particular model requires more data than the others due to complexity, which should be selected for application?)
- Which applications should be included with demand modeling (e.g., toll mode choice), considering model run times and complexity?
- What additional modeling tools (e.g., time-of-day) could be utilized to meet the defined goal(s) of the project scope?
- What model input data should be required for a "Planning Level" study?
- What can be considered in addition to traffic counts by time-of-day, travel surveys, and O/D surveys?
- What model output data should be required for a "Planning Level" study?
- What can be considered in addition to traffic volumes by time-of-day, toll rates by time-of-day, and revenue estimates based on time-of-day data?
- What model results make a successful "Planning Level" project?
- What can be considered in addition to the implications of alternatives and ability to select a preferred alternative?
- When and to what degree of detail should statistical probability/risk analysis be introduced in the "Planning Level" process?
- Should Dynamic Traffic Assignment (DTA) be introduced as part of the modeling process and, if so, when?
- Identify any additional factors/concerns:
- Should full model redevelopment be required for producing bonding capacity results for "Investment Grade" studies?
- Should intensive microsimulation be implemented for "Investment Grade" studies?
- The panel's discussion of these questions is presented in Section 5.0 of this report.
4.2 Operational Analysis Breakout Session
The Operational Analysis Breakout Session was moderated by the following three individuals:
- James Sturrock (FHWA);
- Andrew Velasquez (URS/Florida Turnpike); and
- David Stroud (PB).
The expert panel members invited to participate in this session included:
- James Sturrock (FHWA);
- Yi-Chang Chiu (University of Arizona); and
- Peter Vovsha (PB).
The moderators led a series of discussions focusing on questions related to three different topic areas that included: Process and Policy, Dynamic Traffic Assignment, and Microsimulation.
The specific questions posed by the moderators to the expert panel and the breakout session participants are listed below. There were eleven questions regarding Topic Area #1, eight questions regarding Topic Area #2, and six questions regarding Topic Area #3.
Topic Area #1 = Process and Policy
- At what point should operations be involved in the determination of feasible access points?
- What are the required analysis procedures for an Interstate System Access Request? When is a PD&E or Design required?
- In what project development phase is O/D survey data and/or SPS data needed?
- What should function as the source for express lane demand for operational analysis? Potential sources include, but are not limited to: regional demand models, corridor time-of-day models, DTA models, or microsimulation choice models.
- What is included in a concept of operations, and at what point in a study should a concept of operations (ConOps) be executed?
- Which pricing policy should be modeled at the PD&E and Design levels: revenue maximization or traffic maximization?
- When should the operational models be integrated with the T&R output for PD&E and Design evaluation?
- How does FDOT address pricing policy changes from an operations perspective?
- Would you recommend that FDOT establish a life-cycle travel demand model/ microsimulation model philosophy? For example: should models created for tolled-managed lanes in the project development planning step be passed forward for enhancement and applied in subsequent project development steps?
- What MOEs and values should be used to confirm the feasibility/viability of tolled managed lanes to subsequently move a project forward to development?
- What MOEs and values should be used to confirm the express lanes' feasibility/viability? If not viable, how can future implementation of tolled managed lanes be preserved? For example, should HOV lanes could be implemented.
Topic Area #2 = Dynamic Traffic Assignment
- Is DTA current in use for PD&E (NEPA)/Design analysis in other regions?
- What level of acceptance/education is necessary for the incorporation of DTA at the institutional level?
- What standard scoping templates and guidelines are needed for managing a DTA project?
- What are the implications of NCHRP 255 or the FDOT Project Traffic Forecasting Handbook in the application of DTA?
- What are appropriate DTA validation standards?
- Identify data shortcomings of DTA.
- How can toll choice model expressions be incorporated in DTA?
- Can DTA be used to calculate an equilibrium toll?
Topic Area #3 = Microsimulation
- What should be the source of the decision and pricing model inputs for the VISSIM and CORSIM managed lane modules?
- How can VISSIM or CORSIM be modified to model reliability?
- What is the basis for adjusting the intercept values in VISSIM?
- Should preference be giving to multiple segment travelers in VISSIM?
- What microsimulation MOE's are acceptable in oversaturated conditions?
- What modeling time-slice is preferred: 15 minutes or 1 hour? What are the shortcomings of a lower time-slice?
The panel's discussion of these questions is presented in Section 6.0 of this report.