From Academia to Application: Results from the Development of the First Accessibility-Based Model...

From Academia to Application:From Academia to Application:

Results from the Development of the Results from the Development of the First Accessibility-Based ModelFirst Accessibility-Based Model

From Academia to Application:From Academia to Application:

Results from the Development of the Results from the Development of the First Accessibility-Based ModelFirst Accessibility-Based Model

Mike Conger, P.E.Mike Conger, P.E.Knoxville Regional Transportation Planning OrganizationKnoxville Regional Transportation Planning Organization

Vince Bernardin, Jr., Ph.D.Vince Bernardin, Jr., Ph.D.Bernardin, Lochmueller & Associates, Inc.Bernardin, Lochmueller & Associates, Inc.

Mike Conger, P.E.Mike Conger, P.E.Knoxville Regional Transportation Planning OrganizationKnoxville Regional Transportation Planning Organization

Vince Bernardin, Jr., Ph.D.Vince Bernardin, Jr., Ph.D.Bernardin, Lochmueller & Associates, Inc.Bernardin, Lochmueller & Associates, Inc.

Background & Motivation Overview of Model Design Progress Update Initial Results

Contributions to the State-of-the-Art New Policy Variables and Sensitivity

Agenda

Background & MotivationBackground & Motivation Overview of Model Design Progress Update Initial Results


Agenda

Knoxville Area Population: 863,000 Area: 3,425 sq. mi. Network: 6,626 lane miles TAZ: 1019 zones

Background

Knoxville Area Multinucleated

Knoxville Maryville Oak Ridge Smoky Mountains

/ Gatlinburg, etc.

Background

Knoxville Area Multinucleated Topography

Background

2000 Household Travel Survey

2002 New Trip-based Model

2005 Model Peer-Review

2007 External Cordon Line Survey

2008 Land Use Model (ULAM)

2008 Household Travel Survey

2008 Transit On-board Survey

2009 New Hybrid Trip/Tour-based Model

Background

Model Peer Review Noted poor distribution and k-factors

Policy and Planning Interests Built environment / land use interactions Importance of transit and walking Future tolling / pricing scenarios?

Motivation

Hybrid Trip/Tour-based Model promised: Improved fundamentals of travel behavior Sensitivity to new planning / policy issues Reasonable model run times Reasonable development costs and

timeframe

New Model

Background & Motivation Overview of Model DesignOverview of Model Design Progress Update Initial Results


Agenda

Methodology1. A hybrid disaggregatedisaggregate / aggregateaggregate system

To maximize model fidelity and minimize run time

A New, Alternative Model Design

Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice

Activity / TourGeneration

PopulationSynthesizer

Accessibility

Travel Times

LinkFlows

Departure TimeChoice

Variables Models

Disa

ggregate M

odelsA

ggregate M

odels

Vehicle AvailabilityChoice



2.2. Disaggregate vehicle & tour mode choicesDisaggregate vehicle & tour mode choices

3.3. Departure time choiceDeparture time choice


Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels




2.2. Disaggregate vehicle & tour mode choiceDisaggregate vehicle & tour mode choice


4. Feedback of ACCESSIBILITYACCESSIBILITY as well as travel time To introduce sensitivity to ‘lower level’ choices in

‘upper level’ decisions


Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels




2.2. Disaggregate tour mode choiceDisaggregate tour mode choice


4. Feedback of ACCESSIBILITYACCESSIBILITY as well as travel time To introduce sensitivity to ‘lower level’ choices in

‘upper level’ decisions

5. A ‘double destination choicedouble destination choice’ framework Produce trips consistent w/ tours & trip-chaining behavior


Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels


Background & Motivation Overview of Model Design Progress UpdateProgress Update Initial Results


Agenda

Since development began in earnest in January: All model estimation is complete Model implementation is nearly complete Model validation is currently underway

Progress

Background & Motivation Overview of Model Design Progress Update Initial ResultsInitial Results

Contributions to the State-of-the-ArtContributions to the State-of-the-Art New Policy Variables and Sensitivity

Agenda

Choice Hierarchy

In traditional four-step models, mode choice was modeled conditional on (after) destination choice (due to a preoccupation with choice riders and commuting).

Instead, we modeled stop location or destination choice conditional on (after) mode choice

March 12, 2009

Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels


Choice Hierarchy

In traditional four-step models, mode choice was modeled conditional on (after) destination choice (due to a preoccupation with choice riders and commuting).

Instead, we modeled stop location or destination choice conditional on (after) mode choice We sequentially estimated combined (nested logit) combined (nested logit)

mode and stop location (and sequence) choice modelsmode and stop location (and sequence) choice models And all the logsum / nesting parameters were in the

acceptable ranges without using constraintswithout using constraints, suggesting that this may be the correct choice hierarchy

Choice Hierarchy

This reverse choice hierarchy reflects the fact that many travelers are more likely to change many travelers are more likely to change destinations than switch modesdestinations than switch modes Even for work tours, the data suggests that in

Knoxville, people are more likely to change jobs than change their travel mode to work

This may not be as unreasonable as it seems, considering captive riderscaptive riders, dependent on the bus to get to work

Imposing the traditional hierarchy traditional hierarchy may be a source of “optimism bias” “optimism bias” in transit forecasts

The new model incorporates travelers’ tendency to group stops together into convenient chains

This convenience of locations to other stops is measured by special accessibilityaccessibility variables included in stop location (destination) choice

Trip-Chaining

Independence

In traditionaltraditional models, two equidistantequidistant, equal-sizeequal-size destinations are equally probableequally probable.

ComplementarityIn the Knoxville model, the more accessiblemore accessible one is more probable -

because you have to go a nearby destination anyway, and so it’s convenient.

Higher accessibility means the expected costexpected cost of a possible

subsequent trip is lower.

The new model incorporates travelers’ tendency to group stops together into convenient chains

This convenience of locations to other stops is measured by special accessibilityaccessibility variables included in stop location (destination) choice The models are extensions of Fotheringham’s

Competing Destinations (CD) model Incorporating recent research by Bernardin,

Koppelman and Boyce (2009)

Trip-Chaining

The model uses a “double destination choice” framework Generating trips guaranteed to be consistent

with tours Without generating the tours, themselves Allowing fast run times!

Touring

Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels


Travel Cost Elasticity Found elasticities of out-of-home activities with

respect to accessibility of 0.13 - 0.16 Lower tour-making by residents of ruralrural (lower-

accessibility) areas, Decreased tour/stop-making in response to

congestioncongestion (decreased accessibility), Induced tour/stop-making in response to added added

network capacitynetwork capacity (increased accessibility), Induced tour/stop-making in response to new land new land

use developmentsuse developments in other nearby zones (increased accessibility)

Network

TAZ

Flow Averaging

TrafficAssignment

HOV and TollChoices

Stop SequenceChoice

Stop LocationChoice

Tour ModeChoice



Accessibility

Travel Times

LinkFlows


Variables Models

Disa

ggregate M

odelsA

ggregate M

odels


Residence Effects on Trip Length

When people choose their residence location, they also choose how far they are willing to travel.

We allowed travelers’ willingness-to-travel, and hence, trip lengths to vary as a function of the accessibility of their residence location The willingness-to-travel of residents of the most urban most urban

(most accessible) areas was about 10% lower 10% lower than the regional average

The willingness-to-travel of residents of the most rural most rural (least accessible) areas was about 200% higher 200% higher or twice the regional average for most activity types

Cost Elasticity from Accessibility

Including accessibility in both activity generation and stop location choice reflects fewerfewer, but longerlonger ruralrural tours; more shorter urban more shorter urban tours

Background & Motivation Overview of Model Design Progress Update Initial ResultsInitial Results

Contributions to the State-of-the-Art New Policy Variables and SensitivityNew Policy Variables and Sensitivity

Agenda

Auto Availability Each individual household chooses how

many vehicles to own / lease Disaggregate ordered response logit model Vehicle ownership levels respond to

Demographics (household size, income, number of workers, students, etc.)

Gas PricesGas Prices Transit AvailabilityTransit Availability Urban Design Urban Design (intersection approach density

measuring grid vs. cul-de-sac network design)

No Veh

1 Veh

Nest

Nest

Root

2 Veh

Nest

3 Veh

4+ Veh

Tour & Stop Generation

Workers Non-Workers Students Seniors Vehicles Income Gas Price Access-

ibility

Work Tours + - +

Work Stops + - + - +

Non-UT Univ Stops + + - - +

Other Stops + - + - + - +

School Tours + +

School Stops + +

Other Stops + + +

Other Tours + + + +

Short Maintenance Stops + + + +

Long Maintenance Stops + - + + + +

Discretionary Stops + + + - +

Factors Affecting Mode Choice

Level of Service Costs Demographics Built Environment

Accessibility by mode

Distance to UT

% of TAZ Near Bus Gas Price Bus Fare Workers Students Senior

HH IncomeVehicles per

Person

Percent Sidewalks

Activity Diversity

Intersection Density

Work Tours Auto + - + + + + + - - Bus + + - - - + - - -

Walk + + + + - - - + + UT Tours

Auto + - - + - Bus - + + - -

Walk - - - - +

School Tours Auto + - + + Bus + + - + Walk + + - -

School Bus + + - -

Other Tours Auto + - + + + + + - -Bus + + - + - - - - -

Walk + + + - + - - + +

Stop Location Choice

Consistency with tourstours and trip-chaining trip-chaining behavior Reduced aggregation bias Improved policy sensitivitypolicy sensitivity Reasonable run times run times and development costsdevelopment costs!

Advantages of the New Approach

Thank You!

From Academia to Application: Results from the Development of the First Accessibility-Based Model...

Documents

Transcript of From Academia to Application: Results from the Development of the First Accessibility-Based Model...