HIGHWAY

1
HIGHWAY RANKING MODEL
HIGHWAY RANKING MODEL
Based on Future Congestion and Its Relief Per
Cost (Congestion Relief is only one of many
criteria for selecting projects for construction)
Bill Mann, P.E. Transportation Planning
Section Northern Virginia District July 25, 2006
2
HIGHWAY RANKING MODEL
PURPOSE

• No standardized model exists to objectively rank
  highway improvements.
• VDOT has developed a method to do this ranking.
• The method piggybacks off MPO traffic forecasting
  models to rank highway projects.
• This tool is just one objective tool in the final
  ranking process that addresses gridlock.

3
HIGHWAY RANKING MODEL
WHY RANK IMPROVEMENTS?

• There is a funding shortfall of 15 billion to
  meet
• the needs of the 2030 Transportation Plan for
  the
• Northern Virginia District.
• TransAction 2030, January, 2006

4
HIGHWAY RANKING MODEL

• INPUTS (Nothing New is Required)
• Areas Future Highway Networks (Transit networks
  may be added in later model versions.)
• Areas Future Highway Trips
• Improvement Costs
• OUTPUTS (Two Ways of Measuring Benefit/Cost
  Ratios)
• Option 1 - Ranking of Improvements by Vehicle
  Hours of Delay (VHD) Reductions (above LOS _) Per
  Cost
• Option 2 - Ranking of Improvements by Vehicle
  Miles of Travel (VMT) (above LOS _) Reductions
  Per Cost
• Note Thresholds for both options are user
  defined (LOS C, D, E, F, G)

5
HIGHWAY RANKING MODEL

• CODING REQUIRED FOR RANKING
• (Took about two weeks for Northern Virginia)
• For each network link affected, code
• -- ID Number
• -- New Lanes (if needed)
• -- New Functional Class Code (if needed)
• Provide table of construction costs for each
  improvement ID
• ______________________________________________
• Note Once all CLRP elements are coded, one
  could rank selected elements for any forecast
  year.

6
HIGHWAY RANKING MODEL
MODELING PROCESS (Assumes 70 Improvements as
Identified in the CLRP)

• Phase I (First Iteration 70 Individual
  Assignments)
• Calculate VHD for Base Network (with no
  improvements).
• Add Improvement No. 1 to Base and calculate VHD
  Reductions with this Improvement.
• Delete Improvement No. 1 and add Improvement No.
  2 to Base Calculate VHD Reductions with this
  Improvement.
• Repeat Step 3 until all 70 improvements are
  tested.
• Rank all 70 improvements based on their VHD
  Reductions/Cost.

7
HIGHWAY RANKING MODEL
MODELING PROCESS (Contd.)

• Phase II (Second Iteration 69 Individual
  Assignments)
• Add the highest-ranked Improvement from Phase I
  to the Base Network to get a new Permanent Base.
• Calculate VHD for the new Permanent Base.
• Repeat Steps 2 through 5.
• Remaining 68 Phases
• Do 68 Individual Assignments, then 67, 66, etc.
• Repeat Steps 6 through 8 until all 70
  improvements are ranked.
• Summarize results.

8
HIGHWAY RANKING MODEL
MODEL OUTPUTS
Phase I Output shows for each improvement when
compared to the Base the following (tested
individually) -- UserBenefits 1) Vehicle hours
of delay reductions (VHDR) or 2) VMT at LOS F
reductions (VMTR) -- Construction Costs -- User
Benefit/Cost Ratio expressed as VHDR/C or
VMTR/C ___________________________________________
_____________________________ Phase II Output
shows for each improvement as it is
incrementally added to Base (synergistic effect)
what its contribution is relative to the entire
plan -- User Benefits (VHDR or VMTR) --
Construction Costs -- User Benefit/Cost Ratio
9
HIGHWAY RANKING MODEL
ADVANTAGE OF MODELING IMPROVEMENTS

• Improvement A and Improvement B might both be
• excellent improvements by themselves based on
• gridlock relief per cost.
• If A is implemented first, the regional benefits
  of B
• might diminish and may not reduce regional
• congestion any further on the regional system
  until
• post-2010/2030.
• This process most likely spreads the ranking of
• improvements around the region, irrespective of
  
• geography or political boundaries.

10
HIGHWAY RANKING MODEL
WHY HASNT THIS BEEN DONE BEFORE?

• It is cost prohibitive without new software.
• One would have to code thousands of computer
  networks
• and run thousands of computer traffic
  assignments to
• rank 70 improvement projects in CLRP (e.g., 70
  69 68
• 67 66 65 etc.) 2485 runs.
• Coding and running 3 networks and 3 traffic
  assignments
• per week in sequential order is all one could
  do with
• traditional modeling techniques.
• This approach would take years to accomplish
  manually.

11
HIGHWAY RANKING MODEL
EXECUTION TIME WITH NEW VDOT MODEL
-- 6 days of computer time to rank 70
improvement projects in Northern Virginia
CLRP using a PC with Intel 3.06 GHz
processor.
12
HIGHWAY RANKING MODEL
THE NEXT SLIDES SHOW REAL SAMPLE RESULTS

• VDOTs Central Office selected 52 Northern
• Virginia District arterial candidates for
  possible
• inclusion in the next Six-Year Plan.
• These were run through the Highway Ranking
• Model using 2005 trips from MWCOG.
• (We could have used 2010 trips.)
• MWCOGMetropolitan Washington Council of
  Governments

13
HIGHWAY RANKING MODEL
14
HIGHWAY RANKING MODEL
RANKING RESULTS
EXAMPLE Project 24 reduces 347 vehicle hours of
delay (daily) per construction cost ( millions).
Final Output showed that after adding the top 16
improvements, the bottom 36 did little to reduce
VHD.
15
HIGHWAY RANKING MODEL
COST EFFECTIVENESS OF IMPROVEMENTS
NOTES 1. This is total cost () divided by the
vehicle hours of delay saved assuming an interest
rate of 5 per year for annualizing capital
investments. 2. Delay is measured if LOS D or
worse. LOS is user defined.
16
HIGHWAY RANKING MODEL
WHO IS AFFECTED BY IMPROVEMENTS?

• COLLECTOR ROADWAYS. Users living in the
• immediate area and those using the road and
• parallel roads.
• FREEWAYS. This can affect users from parallel
• roads as far away as 10 miles.

17
HIGHWAY RANKING MODEL
WHERE DOES ALL THE NEW TRAFFIC COME FROM?

• In modeling (and in the real world) a roadway
• widening could result in a LOS F after the
  widening
• with great benefits to the road and on parallel
  roads.
• Someone may ask Where did all this new
  traffic
• come from? It came from the three Ss
• -- mode shift,
• -- time shift, and
• -- diversion shift.

18
HIGHWAY RANKING MODEL
MODEL OUTPUT CORRELATES WITH TTI DATA FOR 2003
NETWORK

• Texas Transportation Institute (TTI) estimates
  69 hours
• of delay per year, per peak period traveler in
  2003.
• Source Washington Post, May 10, 2005
• VDOT/COG model estimates 108 hours of delay per
  
• year, per peak period traveler in Northern
  Virginia.
• ________________________________________________
• Explanation of Differences
• TTI measures delay on Freeways and Major
  Arterials only.
• VDOT measures delays on all roadways, including
  Minor
• Arterials and Collectors.

19
HIGHWAY RANKING MODEL
RECOMMENDED APPROACH FOR IMPROVED
TRANSPORTATION PLANNING PROCESS

• Rank all plan elements (Comp, CLRP, SYIP)
• using future land uses, trips and networks by
• forecast year.
• For those at bottom half of priority list,
  investigate
• to see why they are low in the ranking. What
  could
• be changed to improve their ranking?
• Take results to decision makers for plan
  updates.
• This is an ongoing, continuous, iterative
  process.

20
HIGHWAY RANKING MODEL
EXAMPLE OF IMPROVED TRANSPORTATION PLANNING
PROCESS
(Post Processing)
ROUTE Z
SUBURBS
A
B
C
ACTIVITY CENTER

• Project Widen Route Z from 4 lanes to 6 lanes
  grade separate 3 intersections.
• Ranking Model Results Low
• Analysis Delete interchanges.
• Results Less VHDR, less cost, but benefit/cost
  ratio improves.
• Plan Update Widen roadway now but postpone
  building interchanges until after 2020/2030 when
  LOS drops to F/G based on forecast 2020/2030 land
  uses.

21
HIGHWAY RANKING MODEL
OUTPUT OPTIONS

• Rank all improvements or only selected
  improvements.
• Rank selected Functional Classification
  improvements
• (Freeways, Major Arterials, and/or Collectors).
  
• Example Rank only primaries assuming
• -- All planned secondary and freeway improvements
  
• are built, or
• -- All planned secondary and freeway improvements
  
• are not built.

22
HIGHWAY RANKING MODEL
OUTPUT OPTIONS (Contd.)

• Select threshold for LOS delays.
• Rank improvements based on
• -- VHD Reductions
• -- VHD Reduction/Cost
• -- VMT at LOS F Reductions
• -- VMT at LOS F Reductions/Cost
• Select small or large area or specific
  jurisdictions
• for comparing VHD Reductions.

23
HIGHWAY RANKING MODEL
CONCLUSION
With new and better computer models, we can now
help decision makers

• Maximize area-wide traveler benefits at minimum
• cost
• Develop more cost effective transportation
  plans

24
HIGHWAY RANKING MODEL
Thank You! Questions? Comments ?
25
HIGHWAY RANKING MODEL
CONTACT INFORMATION
HIGHWAY RANKING MODEL
Bill Mann Bill.Mann_at_VDOT.Virginia.gov 703-383-2211
Mazen Dawoud Mazen.Dawoud_at_VDOT.Virginia.gov 703-38
3-2229