TRANSIMS Current Status

1
TRANSIMS Current Status

• Frederick W. Ducca
• FHWA
• September 8, 2005

2
TRANSIMS Topics

• Status of Portland Study
• Lessons learned
• Challenges
• Other Applications
• Costs
• Open Source

3
Portland Study

• Track 1 Network Simulation
• Use networks travel demand from trip-based
  model
• Route and microsimulate individual trips for
  entire region
• Track 2 (Gen2) Full model
• Tour model replaces trip based demand
• Integrates tour model with network simulation

4
Conceptual Framework
Trip Tables
Attraction Balancing
Route Attributes
Population Synthesis
Activity Times and Locations
Mode Preference
Router
Microsimulator
Stabilization
Refine Modes
Change Activity Times or Patterns
5
Portland Size of Problem

• Regional simulation using existing network and
  trip tables (Track 1)
• 24 hours
• 4,900,000 trips
• 28,814 activity locations
• 1,370 signals
• 1260 zones
• 8375 links
• 250 transit routes

6
Portland Track 1

• Completed simulations
• Sensitivity testing
• Change signal timing
• Bridge closing
• System reliability
• 15 traffic growth
• Reasonable responses

7
Portland Track 2

• Activity-based tours with network microsimulation
• Simulate the activities and movements of
  individuals, vehicles and households
• Second by second over 24 hour period
• Model the location, duration and type of
  activities

8
Track 2 Current Status

• Calibrated and validated tour and destination
  choice models
• Specified mode choice model and partially
  calibrated
• Specified model feedback mechanisms

9
Lessons Learned

• Start with existing data
• Evolve from existing models
• Equilibration and convergence
• Different complexities
• New opportunities
• Testing and validation

10
Start with Existing Data

• Used available network and trip tables from
  Portland
• Utilities to translate and augment network
• Trip tables
• Pocket lanes
• Signal timing
• Transit schedules
• Do NOT need all streets

11
Start with Existing Data, But

• Need to refine translated networks
• Major or unique network elements
• Simulations are sensitive to data specification
  coding errors
• Local link problems become system problems due to
  queue spillback
• Developed utilities to diagnose and correct
  network problems

12
Evolve from existing models

• Track 1 used existing trip tables as input to
  microsimulator
• Added a layer on 4-step approach
• Track 2 used existing destination choice mode
  choice model elements

13
Equilibration and Convergence

• Route microsimulate whole trips
• Methods generally similar to iterative capacity
  restraint
• Three stage approach in Portland
• Use ICR-VDF for initial simulation
• Minimize path differences iterating between
  router microsimulation until gap criteria are
  met
• Iterate towards a Nash equilibrium by re-routing
  sets of individual travelers until gap criteria
  are met

14
New Opportunities

• All model components work with continuous time
• Different modeling tools
• Evaluate travel schedule, routes and simulation
  results for each traveler
• Feedback sets of travelers to refine and improve
  choices

15
Testing and validation

• Validate using available information
• ALL traditional validation measures
• New measures (ATR data)
• Sensitivity tests
• Demand, capacity, traffic control

16
Challenges

• How to communicate results
• Wealth of data
• How to analyze
• What to analyze
• How to visualize
• What do we tell people
• Future No-Build scenarios cant ignore high
  demand/capacity situation

17
Additional Applications

• New Jersey
• Middlesex and Monmouth Counties
• Congestion mobility
• Rutgers University with NJIT
• DC core
• White House vicinity
• Response to street closings
• FHWA

18
Cost Trends

• New study costs running significantly below
  original estimates
• Computer hardware much cheaper
• Use existing data and translation utilities
• Known problems and available fixes
• Must still use Linux

19
Computing Requirements
Cray supercomputer
Linux cluster
20
Open Source Licensing

• NASA Open Source Agreement
• Very similar to a well-known license
• Commercialization-friendly
• Initial products covered
• TRANSIMS source code docs
• TRANSIMS fundamentals textbook
• Application utilities
• Data sets

21
Open Source Information

• Open Source Institute
• http//www.opensource.org
• NASA open source agreement 1.3
• TMIP website
• http//tmip.fhwa.dot.gov/transims
• TRANSIMS listserv