GTAModel v4.2 Overview and Updates

 
GTAModel V4.2
 
TMG September Workshop (2021)
 
Travel Modelling Group
 
1
 
Documentation
 
Please visit 
 for the latest documentation for XTMF, GTAModel, and the
TMGToolbox.
http://tmg.utoronto.ca/doc
 
Travel Modelling Group
 
2
 
GTAModel V4.2 Overview
 
Travel Modelling Group
 
3
 
 
What’s new in V4.2
 
New parking cost model
New Pearson Airport model
Recalibrated Freight Model
Updated 2016 Network with
computed EROW speeds from
GTFS
Recalibrated Transit Assignment
model with separated bus EROW
Updated drive-access-transit
parking capacities
Updated NWS location choice
formulation
New Population synthesizer
MetroPop
 
Travel Modelling Group
 
4
 
Time Periods
 
Travel Modelling Group
 
5
 
Network Assignment
 
Network Assignments are done in
EMME 4.4
Road Assignment
Generating a Fare-Based Hyper-
network
Transit Assignment
Congested
Fare-based
Surface Transit Speed Updating
 
Travel Modelling Group
 
Road Assignment
 
All 5 time periods
Background public transit
demand
Computed automatically
Tolls for the 407
Upgraded to a multiclass
assignment allowing for the
assignment of freight
 
Travel Modelling Group
 
Fare-Based Hyper-network
 
Each transit operator is separated
into a virtual plane
Transfers are then modeled as
links between the planes
Created automatically based on a
fare template file
Allows for fare policy testing
Properly models transfers
Can include discounted transfers
 
Travel Modelling Group
 
Transit Assignment with Surface
Transit Speed Updating
 
Recalibrated model in V4.2 with:
Supports latest networks with TTF6,
Bus EROW segments
Exclusive right-of-way (EROW): stop-
to-stop speeds for each segment
STSU adjusts the bus and
streetcar speeds on SROW
segments
 
 
Travel Modelling Group
 
Compute Access Station
Models
 
Drive Access Transit
Tour-based logit
Attractiveness varies as available
parking changes
Passenger Access Transit
Trip-based logit model
Passenger Egress Transit
Trip-based logit model
 
Travel Modelling Group
 
Tour-Based Access Station Choice - Example
 
Driving to the Access Station
 
Returning from the Access
Station
 
Travel Modelling Group
 
11
 
 
Computing PoR-PoW
 
Separate models for each
Occupation and Employment
Status
Doubly-constrained gravity
model
Uses an auto, transit, and
distance based logit model for
friction
 
 
 
 
Travel Modelling Group
 
Computing PoR-PoS
 
Base-year proportional update by
zone
Separate categories for
elementary, high school, and
university aged persons
 
Travel Modelling Group
 
Assigning Work and School
Zones
 
Discrete zonal assignments
Done per person independently
Based on the corresponding pre-
computed model
 
Travel Modelling Group
 
Assign Driver Licenses
 
Binary logit model
Based on:
Demographics
Income level
Population density of home
Transit utility to work or school
Distance to work or school
 
Travel Modelling Group
 
Assign Cars to Household
 
Ordered-logit model
Based on:
Number of Adults
Number of Kids (<16)
Number of Full-time workers
Number of driver licenses
0, 1, 2, 3+
Household Income category
Population density of home zone
Job density of home zone
Average distance to work for home
zone
Average transit utility to work for home
zone
Average auto travel time to work for
home zone
 
Travel Modelling Group
 
TASHA
 
Generate projects
Combine projects into a personal
schedule and resolve conflicts
Generate trip chains, awaiting
mode assignments from Mode
Choice
 
Travel Modelling Group
 
Projects
 
Separate schedules for different
types of activity episodes
Work
School
Other
Market
Household Other
Household Market
Each project is generated using
the base year’s observed episode
rates
 
Travel Modelling Group
 
Activity Episodes
 
Purpose
Start-time
Duration
Location
 
Generated from observed TTS
distribution rates
 
Travel Modelling Group
 
Person Schedule
 
Used to combine all projects
If full-time employed work
episodes are scheduled first
otherwise school episodes
Household level projects are
scheduled next
Followed by Other and Market
Episodes are assigned locations
as they enter the person schedule
 
Travel Modelling Group
 
Time Prism Location Choice
 
Used to generate the location of
other, market, and work based
business episodes
Provides for a more feasible
schedule as the destination
choice-set is restricted by how far
the person is able to travel by
auto in the allotted schedule
Example on next slide
 
Travel Modelling Group
 
Time Prism Location Choice - Example
 
Travel Modelling Group
 
22
 
 
Generate Trip Chains
 
Generated from the person
schedule
Trips are generated to connect
the activity episodes
A home activity is added if they
are able to travel to home and
stay before going to their next
activity for a given duration
Trips are grouped from home to
home sequences into trip chains
 
Travel Modelling Group
 
Mode Choice
 
Tour-Based Probit Model
Household Level Constraints
Intra-household passenger
Multiple person categories
Student
Employed, by occupation
Professional
General
Sales
Manufacturing
Unemployed non-student
 
Travel Modelling Group
 
Auto (Driver)
 
Limited number of vehicles in the
household
Must have a driver’s license
Cars must return home at the end
of a trip chain
Value of time based on each
person’s category
 
Travel Modelling Group
 
Bicycle
 
Travel Modelling Group
 
Carpool
 
Inter-household passenger
Not taxi or vehicle for hire
Value of time based on each
person’s demographic category
 
Travel Modelling Group
 
Drive-Access Transit
 
Value of time based on each
person’s category
Gets a utility for each feasible
access station
Uses the expected level of service
for access stations
If the tour choses this mode a
discrete station is chosen from
the feasible stations for each
household iteration
 
 
Travel Modelling Group
 
Passenger
 
Intra-Household Facilitated
Trips
Broken into 3 legs
Driver to passenger
To Passenger’s Destination
To Driver’s Destination
Household utility must be
improved in order to be selected
Not available if the potential
passenger already has a vehicle
Value of time based on the
driver’s person category
 
Travel Modelling Group
 
Passenger Access / Egress
Transit (PAT / PET)
 
These modes does not require the
facilitation by a household
member, nor the allocation of a
household auto.
 
Travel Modelling Group
 
Rideshare
 
Intra-Household Joint Trip
Assigned if the tour’s
representative picks Auto
 
Travel Modelling Group
 
Vehicle For Hire (VFH)
 
Inter-household passenger
This mode serves as a grouping
for Taxi and Uber, Lift, and other
Vehicle for Hire services.
 
Travel Modelling Group
 
Walk-Access Transit
 
Uses the integrated congested
fare based transit network.
Different fare perceptions based
on the person’s class
Value of time based on each
person’s category
 
Travel Modelling Group
 
Walk-All Way
 
Travel Modelling Group
 
Mode Choice Algorithm
 
Compute all feasible tours
Compute tour-level utilities
Get highest utility tours for both
with auto and without
Complete discrete stations
choices
Resolve the household to assign
vehicles to tours and optimize
household utility
Find and apply passenger trips
given from drivers currently on
the road
Find and apply passenger trips
given from drivers currently at
home
 
 
 
 
Travel Modelling Group
 
External Model
 
Trips are added in by building a
demand matrix and then
applying a factor to scale based
future population divided by base
year population
Trips are separated into different
modes to be applied to the
appropriate demand matrices
 
Travel Modelling Group
 
Airport Model
 
Updated in V4.2
Generates both auto and transit
demand
Splits demand by time period,
GTHA residents / visitors, and
business / vacation
Has origin/destination location
and mode choice models
Fare sensitive
Results are added back into the
network demand matrices
 
Travel Modelling Group
 
Freight Model
 
Recalibrated using 2016 cordon
counts in V4.2
Three 3-step models for:
Light
Medium
Heavy
Special generators for:
Pearson International Airport
CN Brampton Intermodal Yard
CP Vaughan Intermodal Yard
 
Travel Modelling Group
 
Outputs
 
AssignedNetworks
Auto and Transit NWP’s with
assignment results
Demand
LoS Matrices
Microsim
Households
Persons
Trips
Trip modes
Trip stations
Facilitate passenger
 
Travel Modelling Group
 
Microsim - Households
 
Household ID
Home Zone
Expansion Factor
Number of Persons
Dwelling Type
Number of Vehicles
Income Class
 
Travel Modelling Group
 
Microsim - Persons
 
Household ID
Person ID
Age
Sex
License
Transit Pass
Ignore this for version 4.2
Employment Status
Occupation
Free parking at work
Ignore this for version 4.2
Student Status
Work Zone
School Zone
Weight (expansion factor)
 
Travel Modelling Group
 
Microsim - Trips
 
Household ID
Person ID
Trip ID
Origin Activity
Origin Zone
Destination Activity
Destination Zone
Weight (expansion factor)
 
Travel Modelling Group
 
Microsim – Trip Modes
 
Household ID
Person ID
Trip ID
Mode
Departure time
Arrival Time
Weight (number of times chosen)
 
* Times are in minutes from
midnight
 
Travel Modelling Group
 
Microsim – Trip Stations
 
Household ID
Person ID
Trip ID
Station ID
To / From Transit
auto2transit
transit2auto
Weight (number of times chosen)
Mode (New in XTMF 1.8+)
PAT/PET/DAT
 
* Times are in minutes from
midnight
 
Travel Modelling Group
 
Microsim – Facilitate
passenger
 
Household ID
Passenger Person ID
Passenger Trip ID
Driver Person ID
Driver Trip ID
-1 if the driver facilitated the trip by
going from home and returning back
Weight (number of times chosen)
 
Travel Modelling Group
 
MetroPop
 
A new population synthesizer
Allows you to generate new
synthetic households and persons
without going through TMG
Based on a hybrid IPU and
Monte Carlo simulation
Runs in 6 to 20 minutes, server
to laptop for a 2016 population
Able to generate inputs for
consumption by either
GTAModel or GGHM
Does not require an installation
to setup, just copy and run on
Windows 7+.
 
Travel Modelling Group
 
End Remarks
 
 
Travel Modelling Group
 
47
Slide Note
Embed
Share

This document provides an overview of GTAModel v4.2, including initial network assignment, load data analysis, access station modeling, mode choice considerations, and more. It highlights the new features and enhancements such as updated network assignments, transit models, freight modeling, and time periods classification. The document also discusses the tools and methodologies used for travel modeling and simulation, focusing on the latest advancements in the field.


Uploaded on Sep 01, 2024 | 0 Views


Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

E N D

Presentation Transcript


  1. GTAModel V4.2 TMG September Workshop (2021) Travel Modelling Group 1

  2. Documentation Please visit http://tmg.utoronto.ca/doc for the latest documentation for XTMF, GTAModel, and the TMGToolbox. Travel Modelling Group 2

  3. GTAModel V4.2 Overview Initial Network Assignment Load LoS Data Compute Access Station Models Compute Aggregate PoRPoW Compute Aggregate PoRPoS Load Household Assign Work/School Zones Assign Driver Licenses Assign Cars to Household TASHA Mode Choice External Model Network Assignment Airport Model Compute Next Iteration Freight Model Travel Modelling Group 3

  4. Initial Network Assignment What s new in V4.2 Load LoS Data New parking cost model New Pearson Airport model Recalibrated Freight Model Updated 2016 Network with computed EROW speeds from GTFS Recalibrated Transit Assignment model with separated bus EROW Updated drive-access-transit parking capacities Updated NWS location choice formulation New Population synthesizer MetroPop Compute Access Station Models Compute Aggregate PoRPoW Compute Aggregate PoRPoS Load Household Assign Work/School Zones Assign Driver Licenses Assign Cars to Household TASHA Mode Choice External Model Network Assignment Airport Model Compute Next Iteration Freight Model Travel Modelling Group 4 4

  5. Time Periods Time Period AM Mid Day PM Evening Overnight Start 6:00 AM 9:00 AM 3:00 PM 7:00 PM 0:00 AM End 9:00 AM 3:00 PM 7:00 PM 0:00 AM 6:00 AM Travel Modelling Group 5

  6. Initial Network Assignment Network Assignment Load LoS Data Network Assignments are done in EMME 4.4 Road Assignment Generating a Fare-Based Hyper- network Transit Assignment Congested Fare-based Surface Transit Speed Updating Compute Access Station Models Compute Aggregate PoRPoW Compute Aggregate PoRPoS Load Household Assign Work/School Zones Assign Driver Licenses Assign Cars to Household TASHA Mode Choice External Model Network Assignment Airport Model Compute Next Iteration Freight Model Travel Modelling Group 6

  7. Initial Network Assignment Road Assignment Load LoS Data All 5 time periods Background public transit demand Computed automatically Tolls for the 407 Upgraded to a multiclass assignment allowing for the assignment of freight Compute Access Station Models Compute Aggregate PoRPoW Compute Aggregate PoRPoS Load Household Assign Work/School Zones Assign Driver Licenses Assign Cars to Household TASHA Mode Choice External Model Network Assignment Airport Model Compute Next Iteration Freight Model Travel Modelling Group 7

  8. G Fare-Based Hyper-network Y GY T Each transit operator is separated into a virtual plane Transfers are then modeled as links between the planes Created automatically based on a fare template file Allows for fare policy testing Properly models transfers Can include discounted transfers T T G T Travel Modelling Group 8

  9. Transit Assignment with Surface Transit Speed Updating Recalibrated model in V4.2 with: Supports latest networks with TTF6, Bus EROW segments Exclusive right-of-way (EROW): stop- to-stop speeds for each segment STSU adjusts the bus and streetcar speeds on SROW segments Travel Modelling Group 9

  10. Compute Access Station Models Drive Access Transit Tour-based logit Attractiveness varies as available parking changes Passenger Access Transit Trip-based logit model Passenger Egress Transit Trip-based logit model Travel Modelling Group 10

  11. Tour-Based Access Station Choice - Example Driving to the Access Station Returning from the Access Station Travel Modelling Group 11

  12. Computing PoR-PoW Separate models for each Occupation and Employment Status Doubly-constrained gravity model Uses an auto, transit, and distance based logit model for friction Travel Modelling Group 12

  13. Computing PoR-PoS Base-year proportional update by zone Separate categories for elementary, high school, and university aged persons Travel Modelling Group 13

  14. Assigning Work and School Zones Discrete zonal assignments Done per person independently Based on the corresponding pre- computed model Travel Modelling Group 14

  15. Assign Driver Licenses Binary logit model Based on: Demographics Income level Population density of home Transit utility to work or school Distance to work or school Travel Modelling Group 15

  16. Assign Cars to Household Ordered-logit model Based on: Number of Adults Number of Kids (<16) Number of Full-time workers Number of driver licenses 0, 1, 2, 3+ Household Income category Population density of home zone Job density of home zone Average distance to work for home zone Average transit utility to work for home zone Average auto travel time to work for home zone Travel Modelling Group 16

  17. TASHA Generate projects Combine projects into a personal schedule and resolve conflicts Generate trip chains, awaiting mode assignments from Mode Choice Travel Modelling Group 17

  18. Projects Separate schedules for different types of activity episodes Work School Other Market Household Other Household Market Each project is generated using the base year s observed episode rates Travel Modelling Group 18

  19. Activity Episodes Purpose Start-time Duration Location Generated from observed TTS distribution rates Travel Modelling Group 19

  20. Person Schedule Used to combine all projects If full-time employed work episodes are scheduled first otherwise school episodes Household level projects are scheduled next Followed by Other and Market Episodes are assigned locations as they enter the person schedule Travel Modelling Group 20

  21. Time Prism Location Choice Used to generate the location of other, market, and work based business episodes Provides for a more feasible schedule as the destination choice-set is restricted by how far the person is able to travel by auto in the allotted schedule Example on next slide Travel Modelling Group 21

  22. Time Prism Location Choice - Example 1 4 7 Final Origin Destination 3 2 6 5 Travel Modelling Group 22

  23. Generate Trip Chains Generated from the person schedule Trips are generated to connect the activity episodes A home activity is added if they are able to travel to home and stay before going to their next activity for a given duration Trips are grouped from home to home sequences into trip chains Travel Modelling Group 23

  24. Mode Choice Tour-Based Probit Model Household Level Constraints Intra-household passenger Multiple person categories Student Employed, by occupation Professional General Sales Manufacturing Unemployed non-student Travel Modelling Group 24

  25. Auto (Driver) Limited number of vehicles in the household Must have a driver s license Cars must return home at the end of a trip chain Value of time based on each person s category Travel Modelling Group 25

  26. Bicycle A bicycle must leave from and return to home at the end of the trip-chain Uses network based distances for intra-zonal travel Uses 2 ???? distance for intra-zonal travel to approximate the 6 Travel Modelling Group 26

  27. Carpool Inter-household passenger Not taxi or vehicle for hire Value of time based on each person s demographic category Travel Modelling Group 27

  28. Drive-Access Transit Value of time based on each person s category Gets a utility for each feasible access station Uses the expected level of service for access stations If the tour choses this mode a discrete station is chosen from the feasible stations for each household iteration Travel Modelling Group 28

  29. Passenger Intra-Household Facilitated Trips Broken into 3 legs Driver to passenger To Passenger s Destination To Driver s Destination Household utility must be improved in order to be selected Not available if the potential passenger already has a vehicle Value of time based on the driver s person category Travel Modelling Group 29

  30. Passenger Access / Egress Transit (PAT / PET) These modes does not require the facilitation by a household member, nor the allocation of a household auto. Travel Modelling Group 30

  31. Rideshare Intra-Household Joint Trip Assigned if the tour s representative picks Auto Travel Modelling Group 31

  32. Vehicle For Hire (VFH) Inter-household passenger This mode serves as a grouping for Taxi and Uber, Lift, and other Vehicle for Hire services. Travel Modelling Group 32

  33. Walk-Access Transit Uses the integrated congested fare based transit network. Different fare perceptions based on the person s class Value of time based on each person s category Travel Modelling Group 33

  34. Walk-All Way On network distances for inter- zonal trips Intra-zonal trips use: 2 ???? 6 to approximate the distance Travel Modelling Group 34

  35. Mode Choice Algorithm Compute all feasible tours Compute tour-level utilities Get highest utility tours for both with auto and without Complete discrete stations choices Resolve the household to assign vehicles to tours and optimize household utility Find and apply passenger trips given from drivers currently on the road Find and apply passenger trips given from drivers currently at home Travel Modelling Group 35

  36. External Model Trips are added in by building a demand matrix and then applying a factor to scale based future population divided by base year population Trips are separated into different modes to be applied to the appropriate demand matrices Travel Modelling Group 36

  37. Airport Model Updated in V4.2 Generates both auto and transit demand Splits demand by time period, GTHA residents / visitors, and business / vacation Has origin/destination location and mode choice models Fare sensitive Results are added back into the network demand matrices Travel Modelling Group 37

  38. Freight Model Recalibrated using 2016 cordon counts in V4.2 Three 3-step models for: Light Medium Heavy Special generators for: Pearson International Airport CN Brampton Intermodal Yard CP Vaughan Intermodal Yard Travel Modelling Group 38

  39. Outputs AssignedNetworks Auto and Transit NWP s with assignment results Demand LoS Matrices Microsim Households Persons Trips Trip modes Trip stations Facilitate passenger Travel Modelling Group 39

  40. Microsim - Households Household ID Home Zone Expansion Factor Number of Persons Dwelling Type Number of Vehicles Income Class Travel Modelling Group 40

  41. Microsim - Persons Household ID Person ID Age Sex License Transit Pass Ignore this for version 4.2 Employment Status Occupation Free parking at work Ignore this for version 4.2 Student Status Work Zone School Zone Weight (expansion factor) Travel Modelling Group 41

  42. Microsim - Trips Household ID Person ID Trip ID Origin Activity Origin Zone Destination Activity Destination Zone Weight (expansion factor) Travel Modelling Group 42

  43. Microsim Trip Modes Household ID Person ID Trip ID Mode Departure time Arrival Time Weight (number of times chosen) * Times are in minutes from midnight Travel Modelling Group 43

  44. Microsim Trip Stations Household ID Person ID Trip ID Station ID To / From Transit auto2transit transit2auto Weight (number of times chosen) Mode (New in XTMF 1.8+) PAT/PET/DAT * Times are in minutes from midnight Travel Modelling Group 44

  45. Microsim Facilitate passenger Household ID Passenger Person ID Passenger Trip ID Driver Person ID Driver Trip ID -1 if the driver facilitated the trip by going from home and returning back Weight (number of times chosen) Travel Modelling Group 45

  46. MetroPop A new population synthesizer Allows you to generate new synthetic households and persons without going through TMG Based on a hybrid IPU and Monte Carlo simulation Runs in 6 to 20 minutes, server to laptop for a 2016 population Able to generate inputs for consumption by either GTAModel or GGHM Does not require an installation to setup, just copy and run on Windows 7+. Travel Modelling Group 46

  47. End Remarks Travel Modelling Group 47

Related


More Related Content

giItT1WQy@!-/#giItT1WQy@!-/#giItT1WQy@!-/#giItT1WQy@!-/#