Comprehensive Transit Network Implementation Workshop
undefined
LESSON
7
FSUTMS TRANSIT IMPLEMENTATION
1
LESSON GOALS
•
In this lesson, we will discuss:
•
Transit Network Coding
•
Mode Choice
•
Transit Assignment
•
FTA New Starts Considerations
and STOPS
FSUTMS COMPREHENSIVE MODELING WORKSHOP
2
PART A
TRANSIT OVERVIEW
FSUTMS COMPREHENSIVE MODELING WORKSHOP
3
OUTLINE
•
Transit Network Coding:
•
Overview
•
Highway Network Modifications for Transit
•
FSUTMS Transit Standards
FSUTMS COMPREHENSIVE MODELING WORKSHOP
4
OVERVIEW
•
What is a transit network?
•
Transit networks are a system of existing or proposed transit facilities in a format
for network-based travel demand modeling.
•
A transit network represents individual transit lines in a system of links and nodes.
•
Transit links and nodes define the transit line stops or stations and the distances
between them.
•
Connections onto and off of the transit network are also represented in a series
of pedestrian and vehicular access links.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
5
OVERVIEW (CONT’D)
•
What are the differences between transit stops and transit
stations?
•
Transit stops occur ON the highway network and usually along bus routes.
•
Transit stations usually occur OFF the highway network along light rail,
heavy rail, or rapid rail facilities. Bus park-n-ride lots are also usually
modeled as stations.
•
Transit stops and stations are notated by positive node numbers.
•
Non-stops are notated by negative node numbers (“-”) in the line file.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
6
OVERVIEW (CONT’D)
•
Rail station data includes parking
lot capacity, parking costs, and
walk times needed to generate
auto access connectors.
•
Olympus does not include any rail
modes but rail modes can be
added relatively easily.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
7
OVERVIEW (CONT’D)
•
What constitutes a mode?
•
A mode is represented by the type of transit or transit access. This can include
walk access, automobile access, local bus, express bus, rail, etc.
•
What does headway mean?
•
Headway can be defined as the interval of time between the arrivals of a
vehicle on a particular route at a particular stop. (For example, if a bus from
Route 2 leaves a particular bus stop and 30 minutes pass before another bus
from Route 2 arrives at the same stop, Route 2 is said to have a headway of 30
minutes).
FSUTMS COMPREHENSIVE MODELING WORKSHOP
8
HIGHWAY NETWORK MODIFICATIONS FOR
TRANSIT
•
Three elements should be represented in a highway network for transit:
•
Transit-only links
•
Code links on a highway network
•
Coded with Facility Type 69 and special fields
•
Micro-coded stations
•
Separate rail and bus platforms
•
Individual links connect station nodes to network/bus stop nodes
•
Station data
•
Park-ride and fixed-guideway station information kept on nodes
FSUTMS COMPREHENSIVE MODELING WORKSHOP
9
HIGHWAY NETWORK MODIFICATIONS FOR
TRANSIT (CONT’D)
Mandatory link fields in the highway network
FSUTMS COMPREHENSIVE MODELING WORKSHOP
10
HIGHWAY NETWORK MODIFICATIONS FOR
TRANSIT (CONT’D)
Mandatory node fields in the highway network
FSUTMS COMPREHENSIVE MODELING WORKSHOP
11
FSUTMS TRANSIT STANDARDS
•
Model Structure and Procedures:
•
The Transit Network Model is built during the fourth step in the FSUTMS
model chain.
•
The Transit Network Model utilizes Voyager’s PUBLIC TRANSPORT
program.
•
The model first reads input files (including highway network-related files,
transit route files, transit system data, transit factors, and station data).
•
The process outputs appropriate transit networks.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
12
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Model Structure and Procedures (Cont’d)
•
Any network produced by this process can be displayed by clicking the
appropriate file box on the model flow chart in Cube.
•
Networks can be edited either by modifying the route line files in a text
editor or visually through the Cube interface similar to the highway
network prior to building.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
13
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Input File Standards:
•
One network for each alternative (route file)
•
Cube Voyager LIN format
•
Use HEADWAY[i] to represent different time periods. Up to five periods
are supported on a single transit network.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
14
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Input File Names:
•
Routes - TROUTE_yya.lin
•
Stations - STATDATA_CSV.dat (based on station data kept in the nodes of
highway network)
•
Unloaded Highway Network - UNLOADED_ayy.net
•
Node Coordinates - NODES.csv
•
Turn Penalties/Prohibitors - TURN_yya.pen
FSUTMS COMPREHENSIVE MODELING WORKSHOP
15
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Transit Networks:
•
Input transit network data no longer require separate peak and off-peak
files thanks to Voyager headway parameters.
•
Separate transit networks are output (NTLEGS) for each period in the
model (peak and off-peak periods).
•
Peak period transit networks are typically designated as the “PK” network
(including both morning and evening peak periods).
•
Off-peak period transit networks are typically designated as
the “OP”
network (including all time periods not included in peak periods).
FSUTMS COMPREHENSIVE MODELING WORKSHOP
16
FSUTMS TRANSIT STANDARDS (CONT’D)
•
TROUTE _yya.lin
•
Fields:
•
LINE NAME=: In Cube, lines are identified by a unique name, not a number.
•
ONEWAY: One-way versus two-way directionality is signified as a true (T) / false (F)
switch.
•
HEADWAY[1]: Peak period headway in minutes
•
HEADWAY[2]: Off-peak period headway in minutes
FSUTMS COMPREHENSIVE MODELING WORKSHOP
17
FSUTMS TRANSIT STANDARDS (CONT’D)
•
TROUTE_yya.lin
•
Fields (Cont’d):
•
Mode: Modes are now designated
by two-digit codes.
•
Operator: Operator of the line
individually identified with a
single-digit code.
•
N: Node sequence indicating the
routing of the transit line.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
18
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Output Files:
•
Peak Period Transit Network =
TransitPK.net
•
Off-Peak Period Transit Network
= TransitOP.net
FSUTMS COMPREHENSIVE MODELING WORKSHOP
19
FSUTMS TRANSIT STANDARDS (CONT’D)
•
Output Transit Binary Network Field
Names:
•
Lines:
•
Line Name- NAME, LONGNAME
•
Directionality- ONEWAY
•
Travel Time Factor- TIMEFAC[1],
TIMEFAC[2]
•
Circulator- CIRCULAR
•
Headways- HEADWAY[1], HEADWAY[2]
•
Transit Mode- MODE
•
Company- OPERATOR
•
Nodes- N (non-stops indicated with “-”)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
20
END OF LESSON 7, PART A
PLEASE PROCEED TO GUIDED EXERCISE
29
FSUTMS COMPREHENSIVE MODELING WORKSHOP
21
PART B
MODE CHOICE
FSUTMS COMPREHENSIVE MODELING WORKSHOP
22
OUTLINE
•
Mode Choice:
•
Overview
•
Standards
•
Highway Only Mode Choice-Occupancy Model
•
Transit Mode Choice Models
•
Mode Choice Input Files
FSUTMS COMPREHENSIVE MODELING WORKSHOP
23
OVERVIEW
•
Mode choice models represent travel decisions about
which vehicular mode to use as a function of:
•
Level-of-Service (LOS) characteristics of the mode - may
include highway operating cost, auto occupancy factors, and
value of time.
•
Traveler and household characteristics - may include auto
availability and income.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
24
OVERVIEW (CONT’D)
•
Mode choice is actually applied in two distinctively different ways:
•
For Highway-Only applications, Mode Choice is simply an auto occupancy
model.
•
For MPOs with transit models, Mode Choice determines the split among highway
and transit modes of travel.
•
Most mode choice models in Florida previously used FORTRAN programs (the
exceptions are the Northeast Florida Regional Planning Model (NERPM), the
Florida Statewide Model (Long Distance Passenger and FreightSIM) and
Olympus which use Cube Voyager’s XCHOICE program).
FSUTMS COMPREHENSIVE MODELING WORKSHOP
25
FSUTMS STANDARDS
•
Model Structure and Procedures
•
Two alternate structures are Highway-Only auto occupancy models and
Nested Logit transit mode choice models.
•
Highway-Only mode choice models use auto occupancy rates to convert
person trips to vehicle trips.
•
Nested Logit mode choice models allocate trips between auto and transit
modes.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
26
FSUTMS STANDARDS (CONT’D)
•
Model Structure and Procedures (Cont’d)
•
Mode Choice accomplishes several tasks in building auto and/or transit
vehicle trip tables:
•
Performs a set of trip table manipulations.
•
Allocates trips among a variety of highway and transit modes (Nested Logit
only).
•
Applies vehicle occupancy rates.
•
Balances and factors highway trip tables.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
27
FSUTMS STANDARDS (CONT’D)
•
Model Structure and Procedures (Cont’d)
•
Combines all trip tables into one file:
•
In Highway-Only models, the typical seven-purpose person-trip table output
from DISTRIBUTION is converted into a single-purpose vehicle-trip table.
•
In Transit models, multi-purpose vehicle-trip tables are produced which
subdivide auto trips by vehicle occupancy category, and transit trips by
transit service and access.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
28
TYPES OF MODE CHOICE MODELS
•
Highway-Only Mode Choice:
•
Occupancy Model – Converts person trips to vehicle trips using auto
occupancy factors.
•
Transit Mode Choice:
•
Multinomial Logit Model – Splits trips among a variety of auto and transit
modes (no longer used in Florida).
•
Nested Logit Model – Splits trips among a series of modes and submodes
(auto and transit are primary modes with submodes under each).
FSUTMS COMPREHENSIVE MODELING WORKSHOP
29
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL
•
Overview of Occupancy Model:
•
Converts person trips to vehicle trips by applying auto occupancy factors for each
purpose.
•
Adds all vehicle trips into a single purpose for the assignment process.
•
Transposes productions and attractions to origins and destinations:
•
50/50 split between zone pairs
•
balanced trip table needed to reflect travel over 24-hour period
•
Auto occupancy rates or auto occupancy factors can generally be designated in the
Scenario Manager Catalog Keys for Cube models.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
30
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL (CONT’D)
•
Standard Data Requirements for Highway-Only Mode Choice
Models:
•
Auto occupancy factors by person trip purpose
•
Person trip tables (HB Work, HB Shop, HB SocRec, HB Other, NHB)
•
Commercial vehicle trip table
•
IE vehicle trip table
•
EE vehicle trip table
FSUTMS COMPREHENSIVE MODELING WORKSHOP
31
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL (CONT’D)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
32
Auto Occupancy Factors
for pre-assignment mode
choice in Distribution
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL (CONT’D)
•
Data Requirements: (Cont’d)
•
Auto Occupancy Rate (AOR)
•
AOR = Total Persons/Total Vehicles (for example 1.10)
•
Auto Occupancy Factor (AOFAC)
•
AOF = 1.0 / Auto Occupancy Rate (for example 0.909)
•
By Trip Purpose
•
Home-Based Purposes
•
NonHome-Based Purpose
•
For Vehicle Purposes the AOFAC = 1.0
FSUTMS COMPREHENSIVE MODELING WORKSHOP
33
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL (CONT’D)
•
Data Requirements: (Cont'd)
•
Where do you find auto occupancy factors?
•
Survey Data
•
Household Travel Characteristics Surveys
•
Roadside Travel Surveys (corridor/subarea specific)
•
Census Journey-to-Work/ACS Data
•
Home-Based Work Only
•
Borrowed from other similar areas
•
National Household Travel Survey
•
NCHRP 716 Report
•
Set during model validation and not modified in future years
FSUTMS COMPREHENSIVE MODELING WORKSHOP
34
HIGHWAY-ONLY MODE CHOICE – OCCUPANCY
MODEL (CONT’D)
•
How are AOFACs entered?
•
Auto Occupancy Factors are listed for each trip purpose in
the Catalog Keys of the Cube Scenario Manager for
Highway-Only models.
•
Listed in the Keys as follows:
•
AOFAC_HBW
•
AOFAC_HBSH
•
AOFAC_HBSR
•
AOFAC_HBO
•
AOFAC_NHB
FSUTMS COMPREHENSIVE MODELING WORKSHOP
35
TRANSIT MODE CHOICE MODELS
•
Multinomial Logit Model (previously used):
•
Multinomial Logit models split trips among a variety of auto and transit
modes.
•
All modes compete equally in the absence of nesting structures.
•
Florida is discontinuing the use of both Multinomial and Binomial (only auto
and transit split) Logit models.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
36
TRANSIT MODE CHOICE MODELS (CONT’D)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
37
Basic Logit Equation
(Applies to Multinomial Logit
and
Nested Logit Models)
Where:
M
t
= Mode Split to Transit Mode
μ
(m)
= Utility of mode (m)
TRANSIT MODE CHOICE MODELS (CONT’D)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
38
Multinomial Logit Model Example
TRANSIT MODE CHOICE MODELS (CONT’D)
•
Nested Logit Model:
•
Computes mode choice similar to multinomial mode choice, but further
disaggregates into submodes.
•
Nested Logit models split trips among a series of modes and submodes.
•
Allocates trips to primary modes (transit and auto) first, and then allocates
primary mode shares to submodes (e.g., for transit, walk vs. auto access;
for auto, drive alone vs. shared ride).
•
Current state-of-the-art practice.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
39
TRANSIT MODE CHOICE MODELS (CONT’D)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
40
Nested Logit Model Example
TRANSIT MODE CHOICE MODELS (CONT’D)
•
What are the differences between Multinomial and Nested Logit
Models?
•
The Nested Logit Model provides a more realistic modeling relationship
among possible modes.
•
The Nested Logit Model minimizes violation of the Independence of
Irrelevant Alternatives (IIA).
•
The Nested Logit Model produces more accurate forecasts than the
Multinomial Model with submodes independent from one another.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
41
QUIZ #11
•
What are AOFACs?
•
You need to compute the mode shares of car, red bus, and
blue bus. What type of model will you use?
FSUTMS COMPREHENSIVE MODELING WORKSHOP
42
QUIZ #11 – ANSWERS
•
What are AOFACs?
Auto Occupancy Factors
•
You need to compute the mode shares of car, red bus, and
blue bus. What type of model will you use?
Nested Logit
FSUTMS COMPREHENSIVE MODELING WORKSHOP
43
OLYMPUS MODE CHOICE MODEL INPUT FILES
•
The Mode Choice Model uses the following files as input:
•
Person Trips - PSNTRIPS_ayy.mat
•
External Trip Table - EETABLE_ayy.mat
•
Free-Flow Highway Skims - FREESKIM_ayy.mat
•
Restrained Highway Skims - CONGSKIM_ayy.mat
•
Off-Peak Transit Skims - TSKIMOP_ayy.mat
•
Peak Transit Skims - TSKIMPK_ayy.mat
•
Parking Cost - ZONEDATA_yya.dbf
•
Mode Choice Coefficients - MC_COEFFICIENTS.CSV
•
Mode Choice Targets - MC_TARGETS.CSV
•
Mode Choice Constants - MC_CONSTANTS.CSV
FSUTMS COMPREHENSIVE MODELING WORKSHOP
44
FSUTMS STANDARD FILES
•
The Mode Choice Transit Model generates the following
output files:
•
Print File - FINALTABLES.PRN
•
Vehicle Trips - HWYTRIPS_ayy.MAT
•
Transit Trips - TRNTRIPS_ayy.MAT
FSUTMS COMPREHENSIVE MODELING WORKSHOP
45
MODE CHOICE MODELS
•
Transit vs. Non-Transit Applications:
•
Transit mode choice models are needed when considering major transit capacity options
such as fixed-guideways or Bus Rapid Transit (BRT).
•
Highway-Only models are acceptable where transit is non-existent or limited to all-day
local bus service.
•
If a model is validated with transit, the Highway-Only application should not be used
because of the fundamental difference in auto occupancy procedures.
•
Some models have been validated with both transit and highway-only versions.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
46
STANDARD RESULTS
•
Desired Reporting for Mode Choice Models:
•
Formatted listing of model coefficients and constants.
•
Diagram of nesting structure.
•
Number and percent of trips by mode, sub-mode, and purpose.
•
Mode share for major activity centers.
•
Estimated auto occupancy by purpose.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
47
END OF LESSON 7, PART B
FSUTMS COMPREHENSIVE MODELING WORKSHOP
48
PART C
TRANSIT ASSIGNMENT
FSUTMS COMPREHENSIVE MODELING WORKSHOP
49
OUTLINE
•
Transit Assignment:
•
Overview
•
Review of FSUTMS Standards
•
Look at Data Requirements
•
Standards Results
•
Displaying Results
FSUTMS COMPREHENSIVE MODELING WORKSHOP
50
OVERVIEW
•
Transit assignment models load transit trips to the least cost route
(or path) between each pair of zones.
•
Transit assignment loads transit trip tables to transit networks and
develops a loaded transit network database.
•
Transit assignment is not executed when running FSUTMS Highway-
Only (non-transit) models.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
51
FSUTMS STANDARDS
•
Model Structure and Procedures:
•
Utilizes Cube Voyager Public Transport Program.
•
Olympus includes two Public Transport steps in the Assignment
phase:
•
Load Peak Transit
•
Load Off-Peak Transit
FSUTMS COMPREHENSIVE MODELING WORKSHOP
52
FSUTMS STANDARDS (CONT’D)
•
Input files:
•
Transit networks:
•
Peak period transit network - TransitPK.net
•
Off-peak period transit network - TransitOP.net
•
Transit trip table - TRNTRIPS_ayy.mat
•
Transit paths:
•
Peak period walk to transit paths - TPATHPK1_ayy.rte
•
Peak period parking to transit paths - TPATHPK2_ayy.rte
•
Off-Peak period walk to transit paths - TPATHOP1_ayy.rte
•
Off-Peak period parking to transit paths - TPATHOP2_ayy.rte
FSUTMS COMPREHENSIVE MODELING WORKSHOP
53
FSUTMS STANDARDS (CONT’D)
•
Output files:
•
Loaded peak period transit
network - TRNWLKPK_ayy.net
•
Loaded off-peak period
transit network -
TRNWLKOP_ayy.net
FSUTMS COMPREHENSIVE MODELING WORKSHOP
54
DATA REQUIREMENTS
•
No user-supplied data is required for transit assignment.
•
Data generated during previous model steps used in transit
assignment include:
•
Unloaded transit networks
•
Transit paths
•
Transit trip table
FSUTMS COMPREHENSIVE MODELING WORKSHOP
55
STANDARD RESULTS
•
Reporting Transit Assignment:
•
Boardings and alightings by stop and direction.
•
Estimated number of transfers by mode and route.
•
Peak period vehicle requirements.
•
Transit station loading report in tabular and graphic format.
•
Estimated ridership by mode and route manually compared to observed
values.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
56
DISPLAYING RESULTS
•
Tabular Data:
•
Transit ridership by bus route
•
Transit ridership by mode
•
Total transit ridership
•
Mapping:
•
Transit ridership flow
•
Transit boardings and
alightings
FSUTMS COMPREHENSIVE MODELING WORKSHOP
57
DISPLAYING RESULTS (CONT’D)
Sample Output for Olympus
Model – Boardings and
Alightings
FSUTMS COMPREHENSIVE MODELING WORKSHOP
58
END OF LESSON 7, PART C
PLEASE PROCEED TO GUIDED EXERCISES
30
, 31, 32, AND 33
FSUTMS COMPREHENSIVE MODELING WORKSHOP
59
PART D
FTA NEW STARTS CONSIDERATIONS
FSUTMS COMPREHENSIVE MODELING WORKSHOP
60
OUTLINE
•
FTA STOPS (Simplified Trips-on-Project Software)
•
FTA New Starts:
•
Role of Models like STOPS for FTA New Starts
•
Lessons about travel models
•
Lessons about forecasting
NOTE
: The information presented in this section is derived from the FTA New Starts workshops
held in 2006, 2007, and 2008.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
61
FTA STOPS
•
STOPS is a limited implementation of the conventional four-step travel
demand model used to quantify the measures used by FTA to evaluate and
rate projects.
•
STOPS replaces the standard trip generation and trip distribution steps with
the Census Transportation Planning Products (CTPP) tabulations to describe
overall travel markets.
•
It also replaces the traditional coded transit network with standard transit-
services data in the General Transit Feed Specification (GTFS) format.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
62
FTA STOPS SOFTWARE (CONT’D)
•
This forecasting tool is recommended for use by New Starts project
sponsors.
•
STOPS produces all reporting needed by project sponsors to
review its ridership forecasts in detail and to support grant
applications to the FTA New Starts program.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
63
STOPS VERSUS FSUTMS
•
STOPS - Based off Journey to Work (CTPP)
•
FSUTMS – Broader Range of Survey Inputs including CTPP
•
STOPS – Uses GTFS (General Transit Feed Specification )
•
FSUTMS – May use GTFS or other Modeler Input Network Attributes
•
STOPS – Developed Primarily for Fixed Guideway Transit Analysis
•
FSUTMS – Broader Range of Analysis
•
Local Bus, Express Bus, Other modes (CAV?)
•
STOPS – Easier to Calibrate
•
FSUTMS – Harder to Calibrate but wider range of analysis possible
FSUTMS COMPREHENSIVE MODELING WORKSHOP
64
STOPS VS FSUTMS
•
Need to use both tools
•
Dependent on requirements of analysis
•
For specific new start fixed guideway studies
•
Use STOPS
•
For general understanding of transit interactions
•
Use FSUTMS
•
More complete picture of effects of transit on transportation system performance
•
Provides more regional context
•
More likely to accommodate new developments (CAV, TNC etc.)
FSUTMS COMPREHENSIVE MODELING WORKSHOP
65
THE ROLE OF MODELS FOR FTA NEW STARTS
•
Retrieve transit ridership
•
Models should provide insights
•
Useful forecasts must include derived insights into:
•
Nature of the problem(s) for specific travel markets.
•
Ability of the alternatives to improve transit service.
•
Ridership response for specific travel markets.
•
Benefits accruing to those markets and others.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
66
WHY FOCUS ON INSIGHTS?
•
Provides better context for data on transit riders.
•
Helps set standards for travel models:
•
Few FTA specifications on model properties.
•
New focus on ability to provide insights for decisions.
•
Helps with quality control on forecasts:
•
Easy to put a bad number in a table.
•
Harder to include a bad number in a coherent story.
•
Focuses on information for decision-making.
•
Sets stage for evaluation of success.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
67
IMPLEMENTATION OF EMPHASIS ON INSIGHTS
•
Data and model testing:
•
What are the major transit markets on the system?
•
How well does the model set grasp those markets?
•
“Case” for each proposed project:
•
Required part of information for FTA rating of project.
•
What key things do we think the project will accomplish?
•
Before-and-After Studies:
•
What key things did the project actually accomplish?
•
How well did our forecasts anticipate those things?
FSUTMS COMPREHENSIVE MODELING WORKSHOP
68
LESSONS ABOUT TRAVEL MODELS
•
Analytical reporting is absolutely crucial:
•
Tradition: aggregate summaries
“black box”.
•
FTA response:
•
Standard set of analytical summaries with FTA software.
•
Routine part of project information for FTA review.
•
Outcome: key properties of forecasts are clearer:
•
For testing of travel models and quality control of forecasts.
•
For understanding and refinement of alternatives.
•
For selecting a project and making a case for funding.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
69
LESSONS ABOUT TRAVEL MODELS (CONT’D)
•
Data are absolutely crucial:
•
Tradition: no data requirements
little data.
•
FTA response (requirement effective May 2009):
•
Project proposals supported by “tested” models.
•
“Tested” effectively against data on ridership patterns.
•
Outcomes:
•
Much better understanding of key ridership markets
•
More rigorous testing of models.
•
Better detection of errors in travel models, transit components, and other model components.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
70
LESSONS ABOUT TRAVEL MODELS (CONT’D)
•
Useful testing is more than data-matching:
•
Tradition: aggregate tests, X-section data, factoring
•
FTA response: meaningful testing that includes:
•
Identification of key transit travel markets in current data
•
Focused testing on model’s grasp of key markets
•
Detection and correction of actual sources of error
•
Tests over (1) time and (2) transit system changes
•
Outcomes:
•
Cleaner models
•
More plausible predictions of project benefits
FSUTMS COMPREHENSIVE MODELING WORKSHOP
71
LESSONS ABOUT TRAVEL MODELS (CONT’D)
•
Testing: an example “based on fact”:
•
Existing rail line with 10,000 riders.
•
Models calibrated to match counts.
•
Subsequent survey of riders.
•
Riders with origins and destinations near the line:
•
Models: 15% (so line is a collector/distributor).
•
Data: 85% (so line is really a local-area circulator).
•
Question: How useful was the “calibrated” model?
•
Answer: Not useful at all for similar rail lines.
FSUTMS COMPREHENSIVE MODELING WORKSHOP
72
LESSONS ABOUT FORECASTING
•
Quality control
•
Peer comparisons
•
Analysis of uncertainties
•
Post-implementation assessments
FSUTMS COMPREHENSIVE MODELING WORKSHOP
73
LESSONS ABOUT FORECASTING (CONT’D)
•
Quality control (QC):
•
All forecasts have errors; question is “How big?”
•
Tradition: little QC
lots of errors, big and small
•
FTA response:
•
Standard analytical reports/graphics for project forecasts
•
Routine review by FTA staff
questions
resolutions
•
Outcomes:
•
More attention to details by forecasters, project sponsors
•
Much more plausible forecasts; fewer obvious errors
FSUTMS COMPREHENSIVE MODELING WORKSHOP
74
LESSONS ABOUT FORECASTING (CONT’D)
•
Analysis of uncertainties:
•
The only thing certain in forecasts is uncertainty.
•
Tradition: single-number forecasts for 25 years out
•
New FTA requirement:
•
Analysis of uncertainties
sources of potential error
•
Forecasts as a range: lower – most likely – upper
•
Discussion of specific sources/magnitude of impacts
•
Outcomes (anticipated):
•
Truth in forecasting
better information for decisions
•
Realistic presentation of insights, not just numbers
FSUTMS COMPREHENSIVE MODELING WORKSHOP
75
LESSONS ABOUT FORECASTING (CONT’D)
•
Post-implementation assessments:
•
Learn by doing – but learning takes effort.
•
Tradition: no follow-up evaluation of projects
•
New FTA requirement for every funded project:
•
Before-and-After comparisons
•
Predicted-versus-Actual comparisons
•
Outcomes (anticipated):
•
Better understanding of projects and their benefits
•
Better tools for prediction of benefits
•
Biggest missed opportunity with the program
FSUTMS COMPREHENSIVE MODELING WORKSHOP
76
Summarized in annual FTA
report to Congress
SUMMARY
•
This concludes Lesson 7
•
We covered:
•
Coding a Transit Line
•
Types of Mode Choice Models
•
Retrieving Transit Ridership
•
FTA New Starts Considerations and STOPS
FSUTMS COMPREHENSIVE MODELING WORKSHOP
77
END OF LESSON 7, PART D
FSUTMS COMPREHENSIVE MODELING WORKSHOP
78
This workshop covers various aspects of transit network implementation, including transit network coding, mode choice, assignment considerations, and network distribution. It discusses the differences between transit stops and stations, rail station data, and modes of transit access. The workshop also explains the concept of headway in transit operations.
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.If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
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.
E N D
Presentation Transcript
LESSON 7 FSUTMS TRANSIT IMPLEMENTATION 1
START LESSON GOALS Generation In this lesson, we will discuss: Transit Network Coding Mode Choice Transit Assignment FTA New Starts Considerations and STOPS Network Distribution Transit Mode Split Assignment Reporting END FSUTMS COMPREHENSIVE MODELING WORKSHOP 2
PART A TRANSIT OVERVIEW FSUTMS COMPREHENSIVE MODELING WORKSHOP 3
OUTLINE Transit Network Coding: Overview Highway Network Modifications for Transit FSUTMS Transit Standards FSUTMS COMPREHENSIVE MODELING WORKSHOP 4
OVERVIEW What is a transit network? Transit networks are a system of existing or proposed transit facilities in a format for network-based travel demand modeling. A transit network represents individual transit lines in a system of links and nodes. Transit links and nodes define the transit line stops or stations and the distances between them. Connections onto and off of the transit network are also represented in a series of pedestrian and vehicular access links. FSUTMS COMPREHENSIVE MODELING WORKSHOP 5
OVERVIEW (CONTD) What are the differences between transit stops and transit stations? Transit stops occur ON the highway network and usually along bus routes. Transit stations usually occur OFF the highway network along light rail, heavy rail, or rapid rail facilities. Bus park-n-ride lots are also usually modeled as stations. Transit stops and stations are notated by positive node numbers. Non-stops are notated by negative node numbers ( - ) in the line file. FSUTMS COMPREHENSIVE MODELING WORKSHOP 6
OVERVIEW (CONTD) Rail station data includes parking lot capacity, parking costs, and walk times needed to generate auto access connectors. Olympus does not include any rail modes but rail modes can be added relatively easily. FSUTMS COMPREHENSIVE MODELING WORKSHOP 7
OVERVIEW (CONTD) What constitutes a mode? A mode is represented by the type of transit or transit access. This can include walk access, automobile access, local bus, express bus, rail, etc. What does headway mean? Headway can be defined as the interval of time between the arrivals of a vehicle on a particular route at a particular stop. (For example, if a bus from Route 2 leaves a particular bus stop and 30 minutes pass before another bus from Route 2 arrives at the same stop, Route 2 is said to have a headway of 30 minutes). FSUTMS COMPREHENSIVE MODELING WORKSHOP 8
HIGHWAY NETWORK MODIFICATIONS FOR TRANSIT Three elements should be represented in a highway network for transit: Transit-only links Code links on a highway network Coded with Facility Type 69 and special fields Micro-coded stations Separate rail and bus platforms Individual links connect station nodes to network/bus stop nodes Station data Park-ride and fixed-guideway station information kept on nodes FSUTMS COMPREHENSIVE MODELING WORKSHOP 9
HIGHWAY NETWORK MODIFICATIONS FOR TRANSIT (CONT D) Mandatory link fields in the highway network FSUTMS COMPREHENSIVE MODELING WORKSHOP 10
HIGHWAY NETWORK MODIFICATIONS FOR TRANSIT (CONT D) Mandatory node fields in the highway network FSUTMS COMPREHENSIVE MODELING WORKSHOP 11
FSUTMS TRANSIT STANDARDS Model Structure and Procedures: The Transit Network Model is built during the fourth step in the FSUTMS model chain. The Transit Network Model utilizes Voyager s PUBLIC TRANSPORT program. The model first reads input files (including highway network-related files, transit route files, transit system data, transit factors, and station data). The process outputs appropriate transit networks. FSUTMS COMPREHENSIVE MODELING WORKSHOP 12
FSUTMS TRANSIT STANDARDS (CONTD) Model Structure and Procedures (Cont d) Any network produced by this process can be displayed by clicking the appropriate file box on the model flow chart in Cube. Networks can be edited either by modifying the route line files in a text editor or visually through the Cube interface similar to the highway network prior to building. FSUTMS COMPREHENSIVE MODELING WORKSHOP 13
FSUTMS TRANSIT STANDARDS (CONTD) Input File Standards: One network for each alternative (route file) Cube Voyager LIN format Use HEADWAY[i] to represent different time periods. Up to five periods are supported on a single transit network. FSUTMS COMPREHENSIVE MODELING WORKSHOP 14
FSUTMS TRANSIT STANDARDS (CONTD) Input File Names: Routes - TROUTE_yya.lin Stations - STATDATA_CSV.dat (based on station data kept in the nodes of highway network) Unloaded Highway Network - UNLOADED_ayy.net Node Coordinates - NODES.csv Turn Penalties/Prohibitors - TURN_yya.pen 15 FSUTMS COMPREHENSIVE MODELING WORKSHOP
FSUTMS TRANSIT STANDARDS (CONTD) Transit Networks: Input transit network data no longer require separate peak and off-peak files thanks to Voyager headway parameters. Separate transit networks are output (NTLEGS) for each period in the model (peak and off-peak periods). Peak period transit networks are typically designated as the PK network (including both morning and evening peak periods). Off-peak period transit networks are typically designated as the OP network (including all time periods not included in peak periods). 16 FSUTMS COMPREHENSIVE MODELING WORKSHOP
FSUTMS TRANSIT STANDARDS (CONTD) TROUTE _yya.lin Fields: LINE NAME=: In Cube, lines are identified by a unique name, not a number. ONEWAY: One-way versus two-way directionality is signified as a true (T) / false (F) switch. HEADWAY[1]: Peak period headway in minutes HEADWAY[2]: Off-peak period headway in minutes FSUTMS COMPREHENSIVE MODELING WORKSHOP 17
FSUTMS TRANSIT STANDARDS (CONTD) TROUTE_yya.lin Fields (Cont d): Mode: Modes are now designated by two-digit codes. Operator: Operator of the line individually identified with a single-digit code. N: Node sequence indicating the routing of the transit line. FSUTMS COMPREHENSIVE MODELING WORKSHOP 18
FSUTMS TRANSIT STANDARDS (CONTD) Output Files: Peak Period Transit Network = TransitPK.net Off-Peak Period Transit Network = TransitOP.net FSUTMS COMPREHENSIVE MODELING WORKSHOP 19
FSUTMS TRANSIT STANDARDS (CONTD) Output Transit Binary Network Field Names: Lines: Line Name- NAME, LONGNAME Directionality- ONEWAY Travel Time Factor- TIMEFAC[1], TIMEFAC[2] Circulator- CIRCULAR Headways- HEADWAY[1], HEADWAY[2] Transit Mode- MODE Company- OPERATOR Nodes- N (non-stops indicated with - ) FSUTMS COMPREHENSIVE MODELING WORKSHOP 20
END OF LESSON 7, PART A PLEASE PROCEED TO GUIDED EXERCISE 29 FSUTMS COMPREHENSIVE MODELING WORKSHOP 21
PART B MODE CHOICE FSUTMS COMPREHENSIVE MODELING WORKSHOP 22
OUTLINE Mode Choice: Overview Standards Highway Only Mode Choice-Occupancy Model Transit Mode Choice Models Mode Choice Input Files FSUTMS COMPREHENSIVE MODELING WORKSHOP 23
OVERVIEW Mode choice models represent travel decisions about which vehicular mode to use as a function of: Level-of-Service (LOS) characteristics of the mode - may include highway operating cost, auto occupancy factors, and value of time. Traveler and household characteristics - may include auto availability and income. FSUTMS COMPREHENSIVE MODELING WORKSHOP 24
OVERVIEW (CONTD) Mode choice is actually applied in two distinctively different ways: For Highway-Only applications, Mode Choice is simply an auto occupancy model. For MPOs with transit models, Mode Choice determines the split among highway and transit modes of travel. Most mode choice models in Florida previously used FORTRAN programs (the exceptions are the Northeast Florida Regional Planning Model (NERPM), the Florida Statewide Model (Long Distance Passenger and FreightSIM) and Olympus which use Cube Voyager s XCHOICE program). FSUTMS COMPREHENSIVE MODELING WORKSHOP 25
FSUTMS STANDARDS Model Structure and Procedures Two alternate structures are Highway-Only auto occupancy models and Nested Logit transit mode choice models. Highway-Only mode choice models use auto occupancy rates to convert person trips to vehicle trips. Nested Logit mode choice models allocate trips between auto and transit modes. FSUTMS COMPREHENSIVE MODELING WORKSHOP 26
FSUTMS STANDARDS (CONTD) Model Structure and Procedures (Cont d) Mode Choice accomplishes several tasks in building auto and/or transit vehicle trip tables: Performs a set of trip table manipulations. Allocates trips among a variety of highway and transit modes (Nested Logit only). Applies vehicle occupancy rates. Balances and factors highway trip tables. FSUTMS COMPREHENSIVE MODELING WORKSHOP 27
FSUTMS STANDARDS (CONTD) Model Structure and Procedures (Cont d) Combines all trip tables into one file: In Highway-Only models, the typical seven-purpose person-trip table output from DISTRIBUTION is converted into a single-purpose vehicle-trip table. In Transit models, multi-purpose vehicle-trip tables are produced which subdivide auto trips by vehicle occupancy category, and transit trips by transit service and access. FSUTMS COMPREHENSIVE MODELING WORKSHOP 28
TYPES OF MODE CHOICE MODELS Highway-Only Mode Choice: Occupancy Model Converts person trips to vehicle trips using auto occupancy factors. Transit Mode Choice: Multinomial Logit Model Splits trips among a variety of auto and transit modes (no longer used in Florida). Nested Logit Model Splits trips among a series of modes and submodes (auto and transit are primary modes with submodes under each). FSUTMS COMPREHENSIVE MODELING WORKSHOP 29
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL Overview of Occupancy Model: Converts person trips to vehicle trips by applying auto occupancy factors for each purpose. Adds all vehicle trips into a single purpose for the assignment process. Transposes productions and attractions to origins and destinations: 50/50 split between zone pairs balanced trip table needed to reflect travel over 24-hour period Auto occupancy rates or auto occupancy factors can generally be designated in the Scenario Manager Catalog Keys for Cube models. FSUTMS COMPREHENSIVE MODELING WORKSHOP 30
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL (CONT D) Standard Data Requirements for Highway-Only Mode Choice Models: Auto occupancy factors by person trip purpose Person trip tables (HB Work, HB Shop, HB SocRec, HB Other, NHB) Commercial vehicle trip table IE vehicle trip table EE vehicle trip table FSUTMS COMPREHENSIVE MODELING WORKSHOP 31
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL (CONT D) Auto Occupancy Factors for pre-assignment mode choice in Distribution FSUTMS COMPREHENSIVE MODELING WORKSHOP 32
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL (CONT D) Data Requirements: (Cont d) Auto Occupancy Rate (AOR) AOR = Total Persons/Total Vehicles (for example 1.10) Auto Occupancy Factor (AOFAC) AOF = 1.0 / Auto Occupancy Rate (for example 0.909) By Trip Purpose Home-Based Purposes NonHome-Based Purpose For Vehicle Purposes the AOFAC = 1.0 FSUTMS COMPREHENSIVE MODELING WORKSHOP 33
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL (CONT D) Data Requirements: (Cont'd) Where do you find auto occupancy factors? Survey Data Household Travel Characteristics Surveys Roadside Travel Surveys (corridor/subarea specific) Census Journey-to-Work/ACS Data Home-Based Work Only Borrowed from other similar areas National Household Travel Survey NCHRP 716 Report Set during model validation and not modified in future years FSUTMS COMPREHENSIVE MODELING WORKSHOP 34
HIGHWAY-ONLY MODE CHOICE OCCUPANCY MODEL (CONT D) How are AOFACs entered? Auto Occupancy Factors are listed for each trip purpose in the Catalog Keys of the Cube Scenario Manager for Highway-Only models. Listed in the Keys as follows: AOFAC_HBW AOFAC_HBSH AOFAC_HBSR AOFAC_HBO AOFAC_NHB FSUTMS COMPREHENSIVE MODELING WORKSHOP 35
TRANSIT MODE CHOICE MODELS Multinomial Logit Model (previously used): Multinomial Logit models split trips among a variety of auto and transit modes. All modes compete equally in the absence of nesting structures. Florida is discontinuing the use of both Multinomial and Binomial (only auto and transit split) Logit models. FSUTMS COMPREHENSIVE MODELING WORKSHOP 36
TRANSIT MODE CHOICE MODELS (CONTD) Basic Logit Equation (Applies to Multinomial Logit and Nested Logit Models) ?(? ? ??= ?(? ? ? = ? Where: Mt = Mode Split to Transit Mode (m) = Utility of mode (m) FSUTMS COMPREHENSIVE MODELING WORKSHOP 37
TRANSIT MODE CHOICE MODELS (CONTD) Multinomial Logit Model Example CHOICE AUTO TRANSIT Drive Alone Shared Ride Walk Access Auto Access FSUTMS COMPREHENSIVE MODELING WORKSHOP 38
TRANSIT MODE CHOICE MODELS (CONTD) Nested Logit Model: Computes mode choice similar to multinomial mode choice, but further disaggregates into submodes. Nested Logit models split trips among a series of modes and submodes. Allocates trips to primary modes (transit and auto) first, and then allocates primary mode shares to submodes (e.g., for transit, walk vs. auto access; for auto, drive alone vs. shared ride). Current state-of-the-art practice. FSUTMS COMPREHENSIVE MODELING WORKSHOP 39
TRANSIT MODE CHOICE MODELS (CONTD) Nested Logit Model Example CHOICE AUTO TRANSIT Drive Alone Shared Ride Walk Access Auto Access FSUTMS COMPREHENSIVE MODELING WORKSHOP 40
TRANSIT MODE CHOICE MODELS (CONTD) What are the differences between Multinomial and Nested Logit Models? The Nested Logit Model provides a more realistic modeling relationship among possible modes. The Nested Logit Model minimizes violation of the Independence of Irrelevant Alternatives (IIA). The Nested Logit Model produces more accurate forecasts than the Multinomial Model with submodes independent from one another. FSUTMS COMPREHENSIVE MODELING WORKSHOP 41
QUIZ #11 What are AOFACs? You need to compute the mode shares of car, red bus, and blue bus. What type of model will you use? FSUTMS COMPREHENSIVE MODELING WORKSHOP 42
QUIZ #11 ANSWERS What are AOFACs? Auto Occupancy Factors You need to compute the mode shares of car, red bus, and blue bus. What type of model will you use? Nested Logit FSUTMS COMPREHENSIVE MODELING WORKSHOP 43
OLYMPUS MODE CHOICE MODEL INPUT FILES The Mode Choice Model uses the following files as input: Person Trips - PSNTRIPS_ayy.mat External Trip Table - EETABLE_ayy.mat Free-Flow Highway Skims - FREESKIM_ayy.mat Restrained Highway Skims - CONGSKIM_ayy.mat Off-Peak Transit Skims - TSKIMOP_ayy.mat Peak Transit Skims - TSKIMPK_ayy.mat Parking Cost - ZONEDATA_yya.dbf Mode Choice Coefficients - MC_COEFFICIENTS.CSV Mode Choice Targets - MC_TARGETS.CSV Mode Choice Constants - MC_CONSTANTS.CSV FSUTMS COMPREHENSIVE MODELING WORKSHOP 44
FSUTMS STANDARD FILES The Mode Choice Transit Model generates the following output files: Print File - FINALTABLES.PRN Vehicle Trips - HWYTRIPS_ayy.MAT Transit Trips - TRNTRIPS_ayy.MAT FSUTMS COMPREHENSIVE MODELING WORKSHOP 45
MODE CHOICE MODELS Transit vs. Non-Transit Applications: Transit mode choice models are needed when considering major transit capacity options such as fixed-guideways or Bus Rapid Transit (BRT). Highway-Only models are acceptable where transit is non-existent or limited to all-day local bus service. If a model is validated with transit, the Highway-Only application should not be used because of the fundamental difference in auto occupancy procedures. Some models have been validated with both transit and highway-only versions. FSUTMS COMPREHENSIVE MODELING WORKSHOP 46
STANDARD RESULTS Desired Reporting for Mode Choice Models: Formatted listing of model coefficients and constants. Diagram of nesting structure. Number and percent of trips by mode, sub-mode, and purpose. Mode share for major activity centers. Estimated auto occupancy by purpose. FSUTMS COMPREHENSIVE MODELING WORKSHOP 47
END OF LESSON 7, PART B FSUTMS COMPREHENSIVE MODELING WORKSHOP 48
PART C TRANSIT ASSIGNMENT FSUTMS COMPREHENSIVE MODELING WORKSHOP 49
OUTLINE Transit Assignment: Overview Review of FSUTMS Standards Look at Data Requirements Standards Results Displaying Results FSUTMS COMPREHENSIVE MODELING WORKSHOP 50
