Regional Application of 24-Hour Dynamic Transit Assignment

1
Regional Application of 24-Hour Dynamic Transit
Assignment
12th TRB Transportation Planning Application
Conference, Houston TX, May 2009
Wolfgang Scherr, PTV Ian Fisher, TransLink
2
Overview

• Transit planning in Metro Vancouver and current
  capacity issues
• Dynamic transit assignment the method
• Model development
• Application in service planning and operations
  planning
• Conclusions and future directions

3
Metro Vancouver

• 21 cities
• 2.2 million residents, 3.2 million by 2040
• 294 million transit boardings (unlinked) in 2007
• 78 annual transit rides (linked) per resident
• Major gateway for sea, air and rail freight
• No urban freeways/motorways
• Prior to 1999 no regional transportation agency

4
TransLink Mandate - Integration of
Intelligent Transportation Systems
Public Transport
Transportation Demand Management
Regional Cycling
Roads Bridges
Vehicle Emissions Testing
5
Intermodal Transit Network

• ALRT (SkyTrain)
• Commuter rail (West Coast Express)
• Passenger ferry (SeaBus)
• Bus (includes trolleybus)
• B-Line (frequent, limited stop)
• Community shuttle (minibus)
• BRT and LRT (planned)

6
Rapid Transit Projects in the Region
Not pictured BRT projects
7
Current SkyTrain Capacity Issues

• Picture Broadway Station (SkyTrain), inbound
  boarding, AM peak.
• SkyTrain capacity major subject for application
  of this model

8
Overview of the Regional Transit Model (RTM)

• 24-hour operations model (supply), average
  weekday
• 24-hour passenger model (demand)
• Dynamic assignment
• Fixed OD matrices, 800 zones, interface to
  regional four-stage model
• Platform VISUM
• Applications
• Operations planning routes and patterns, fleet
  requirements, schedule development
• Operations analysis for new rapid transit
  corridors
• Service planning transit area plans, bus network
  planning
• Visualizations and animations that are
  comprehensible to lay people
• History
• Phase A (2006) rail-only model
• Phase B (2008) regional model of all modes

9
The Dynamic Assignment Method

• VISUMs Timetable-based assignment
• Schedule-based method (as opposed to
  headway-based)
• Path-builder uses operations schedule data
• The supply model based on individual vehicle runs
• Assumption deterministic flow of vehicles and
  trains according to schedule
• Multiple paths and multiple connections are built
  using branchbound
• Naturally a time-dynamic method
• Travel conditions vary during the day
• Availability of paths varies during the day
• Travel time of paths varies during the day
• Path choice
• Box-Cox transformed LOGIT
• Impedance IVT OVT NT Delta(early, late)
  correction for similarity of connections
• Vancouver RTM capacity constraints not included

10
Schedule-Based Assignment - Schematic
11
The 24-Hour Dynamic Assignment

• 30-minute time-dynamic passenger volumes 2011,
  colour-coded by v/c-ratio

12
Model Development Supply Side

• Integration of existing operations data
• Network topology
• Nodes and links (NAVTEQ street network)
• Stops, stations (resolution as in APC/AVL)
• Schedule
• Data lines/routes/patterns, run times, runs
  with departure and arrival times
• Existing schedule imported
• Future schedule edited
• Fleet model
• Vehicle types, capacities, consists
• Procedure for block building (vehicle rotations)

13
Model Development Demand Side

• Counts
• Station boarding counts good 24-hour coverage
  for SkyTrain
• Link loading counts ferry, commuter rail,
  specific SkyTrain links
• Automated Passenger Count System (APC)
• All bus lines except community shuttles
• Mainly used total route boardings in hourly
  increments over 24-hours
• APC evaluation tool was still being calibrated
  itself during our model calibration
• OD matrices
• Four segments AM, Midday, PM, Evening
• Existing demand derived with matrix calibration
  and matrix crafting methods
• Future demand
• Link to regional four stage model
• Future OD demand existing OD demand delta
  forecast
• Departure time distributions
• Derived from counts
• Adjusted to meet real-world time distributions
  all over the region

14
Model Development Departure Time Distribution

• Morning peak of demand dependent on distance
  from centre
• Afternoon peak no significant difference
  observed between markets

15
Calibration of Time-Dynamics Model versus Counts

• Rail stations centre versus sub-urban

16
Calibration of Time-Dynamics Model versus Counts

• Suburban bus lines (aggregated APC data for
  several routes)

17
Applications (1) Operations Scenarios SkyTrain

• Study
• Develop and analyze multiple scenarios to operate
  SkyTrain with existing fleet
• Result
• The present operations are optimal with regard to
  capacity
• Analysis
• Use of aggregated and time-disagregated
  performance measures

18
Other Applications (2)

• Other SkyTrain Studies
• Future maximal capacity for the critical segment
  BW-WF
• Fleet strategy it is desirable to add 3-car
  units to the fleet (now only 2-car units)
• Operations concepts for future extensions of the
  SkyTrain system (UBC Line, Evergreen Line)
• Bus service planning
• Bus network planning in the context of a new rail
  line
• The Canada Line (ALRT) will open in 8/2009
• Bus realignments, bus schedules, bus fleet usage
  
• System Analysis
• 2011 short-term forecast interaction of the
  existing SkyTrain Lines with the new Canada Line

19
Conclusions (1)

• Transit supply varies over time
• The use of dynamic tools seems natural in transit
  planning.
• A lesson learned
• It is good to start with a sub-network and later
  expand to entire region
• Early applications allow
• Planners to learn the tool, the data
• Stakeholders to give feedback and formulate
  requirements for the final model
• Benefits of 24-hour dynamic transit assignment
• Better understanding how demand builds up at
  congested spots
• Analysis of peak and off-peak operations
• Visualizations of time-dynamic supply and demand

20
One Benefit of Dynamic Assignment Visualization

• DA can help to visualize and explain truths about
  congestion
• Congestion is of short duration - in a 24-hour
  perspective
• It is challenging to provide capacity as flexibly
  as demand varies
• The demand peak has shifted and the rules of
  operations have to adjust

21
Future Perspectives

• Next studies
• Contribute operations analysis to three major
  strategic studies
• Expo Line Upgrade study
• UBC Line rapid transit study
• Strategic Network Review
• Analysis of BRT projects
• Area transit plans
• Future model improvements
• Calibration of in-station transfer volumes
  (recent surveys)
• Adding capacity constraints to the assignment
  model
• Holding passengers in stations if departing train
  over capacity
• Diverting passengers to alternative paths
• Changing time of departure (peak spreading)
• Research on time-of-day distribution of demand

22
Acknowledgements

• Peter Klitz, TransLink
• Graeme Masterton, TransLink
• Cherry Xiong, PTV
• Kean Lew , PTV
• Anett Ehlert , PTV
• Fang Yuan , PTV
• Juergen Beyer , PTV

23
Thank You
For more information, please contact
Wolfgang Scherr PTV 302-654-4384wscherr_at_ptvameric
a.com www.ptvamerica.com
Ian FisherTransLink604-453-3058ian_fisher_at_trans
link.bc.cawww.translink.bc.ca