Tour Mode Choice Model - Comprehensive Documentation ¶

Overview ¶

The Tour Mode Choice Model is one of the most critical components in CT-RAMP, determining the primary transportation mode for each tour. This model directly influences network loading, environmental impacts, and policy sensitivity of the travel demand system.

UEC Framework - Mathematical framework underlying mode choice calculations
Trip Mode Choice - Trip-level mode choice decisions within tours
Value of Time Analysis - Income-stratified time value assignments used in mode choice utilities

Model Purpose ¶

Primary Function: Select the transportation mode for each generated tour based on level-of-service, demographic characteristics, built environment, and policy variables.

Key Decisions:

Primary transportation mode for entire tour
Access/egress modes for transit tours
Vehicle occupancy for shared ride modes
Parking strategy for auto modes
Cost and time trade-offs between alternatives

Behavioral Foundation ¶

Mode Choice Theory ¶

Utility Maximization: Travelers choose the mode that provides the highest utility (satisfaction) given:

Level-of-Service: Travel time, cost, reliability, comfort
Personal Preferences: Past experience, cultural factors, lifestyle
Constraints: Vehicle availability, physical ability, time budget
Built Environment: Density, walkability, transit service quality

Hierarchical Decision Structure: Mode choice involves nested decisions:

Motorized vs. Non-Motorized: Basic mobility choice
Auto vs. Transit: Within motorized modes
Specific Mode: Detailed alternative selection

Mode Alternatives Structure ¶

CT-RAMP typically includes 17 mode alternatives organized hierarchically:

Auto Modes (Alternatives 1-9) ¶

Drive Alone Modes:

1: Drive Alone - Free parking
2: Drive Alone - Pay parking
3: Drive Alone - Park at transit for return trip

Shared Ride 2 Person (HOV2):

4: Shared Ride 2 - Free parking
5: Shared Ride 2 - Pay parking
6: Shared Ride 2 - Park at transit for return trip

Shared Ride 3+ Person (HOV3+):

7: Shared Ride 3+ - Free parking
8: Shared Ride 3+ - Pay parking
9: Shared Ride 3+ - Park at transit for return trip

Non-Motorized Modes (Alternatives 10-11) ¶

Active Transportation:

10: Walk - All access by walking
11: Bike - All access by bicycle

Transit Modes (Alternatives 12-14) ¶

Public Transit:

12: Walk-Transit-Walk - Access/egress by walking
13: Park & Ride - Drive to transit, transit to destination
14: Kiss & Ride - Dropped off at transit by household member

New Mobility Modes (Alternatives 15-17) ¶

On-Demand Transportation:

15: Taxi - Traditional taxi service
16: Transportation Network Company (TNC) - Uber, Lyft, etc.
17: School Bus - Specialized service for students

Utility Specification Framework ¶

Level-of-Service Variables ¶

Time Components:

In_Vehicle_Time = Auto: driving time
                 Transit: time on transit vehicles
                 Walk/Bike: total travel time

Access_Time = Transit: walk/drive time to first boarding
             Auto: walk time from parking to destination

Wait_Time = Transit: initial wait + transfer wait times

Transfer_Time = Transit: walk time between transit vehicles

Cost Components:

Auto_Operating_Cost = distance * cost_per_mile + tolls

Parking_Cost = hourly_rate * duration (for short trips)
              daily_rate (for long trips)

Transit_Fare = base_fare + distance_fare + transfer_penalties

TNC_Cost = base_fare + time_rate * total_time + 
           distance_rate * distance + surge_pricing

Value of Time Calculations

Cost components in tour mode choice are evaluated using income-stratified value of time calculations. The system differentiates between individual and joint tours when applying time values. See Value of Time Analysis for detailed specifications of the heterogeneous time value assignment system.

Reliability and Comfort:

Reliability = Travel time standard deviation
             Service frequency (transit)
             Weather exposure (walk/bike)

Comfort = Crowding levels (transit)
         Bike infrastructure quality
         Walking environment safety

Demographic and Household Variables ¶

Personal Characteristics:

Age_Effects = Different mode preferences by age group
             Senior discounts and accessibility needs
             Youth transit pass availability

Gender_Effects = Safety perceptions (especially transit, walk)
                Mode-specific preferences and constraints

Income_Effects = Cost sensitivity variations
                Vehicle quality and comfort preferences
                Transit subsidy eligibility

Household Context:

Vehicle_Availability = Autos per licensed driver ratio
                      Vehicle type and condition effects
                      Parking availability at residence

Household_Structure = Joint travel coordination effects
                     Escort responsibilities
                     Activity participation patterns

Employment_Status = Parking subsidies and transit passes
                   Flexible scheduling options
                   Expense reimbursement policies

Built Environment Variables ¶

Density and Mix:

Population_Density = Pedestrian activity levels
                    Transit service justification
                    Parking scarcity effects

Employment_Density = Activity clustering benefits
                    Competition for parking
                    Transit ridership support

Mixed_Use = Reduced travel needs
           Pedestrian-friendly environment
           Shorter trip distances

Transportation Infrastructure:

Transit_Service = Route frequency and coverage
                 Stop/station accessibility and amenities
                 Service reliability and speed

Walking_Environment = Sidewalk coverage and condition
                     Intersection density and safety
                     Weather protection and lighting

Cycling_Infrastructure = Protected bike lane availability
                        Bike parking facilities
                        Network connectivity and safety

Model Structure and Parameters ¶

Nested Logit Hierarchy ¶

Upper Level: Motorized vs. Non-Motorized

U_Motorized = ASC_Motorized + 
              lambda_1 * Logsum(Auto_Modes + Transit_Modes + TNC_Modes) +
              demographics_motorized +
              built_environment_motorized

U_NonMotorized = ASC_NonMotorized +
                lambda_2 * Logsum(Walk + Bike) +
                demographics_non_motorized +
                built_environment_non_motorized

Lower Level: Specific Mode Utilities

Auto Mode Utilities include drive alone, shared ride 2-person, and shared ride 3+ person alternatives with parking and route choice variations.

Transit Mode Utilities incorporate walk access, park-and-ride, and kiss-and-ride with full level-of-service representation.

Non-Motorized Utilities reflect infrastructure quality, safety, and environmental factors affecting walk and bike mode choice.

Market Segmentation ¶

Purpose-Based Segmentation:

Work Tours: Emphasis on reliability, parking costs, employer incentives
School Tours: Safety, specialized services (school bus), parent preferences
Shopping Tours: Convenience, cargo capacity, trip chaining efficiency
Social/Recreation Tours: Flexibility, group travel, cost sharing
Personal Business: Time efficiency, parking availability, accessibility

Demographic Segmentation:

By Income: Low (high cost sensitivity), Middle (balanced trade-offs), High (time emphasis)
By Age: Youth (transit familiarity), Adult (peak auto ownership), Senior (mobility limitations)
By Life Stage: Single (individual preferences), Family (safety priorities), Empty Nester (comfort preferences)

Data Requirements and Sources ¶

Level-of-Service Data ¶

Highway Skims: Time, distance, cost, tolls by vehicle class and time period

Transit Skims: In-vehicle time, access/egress, wait times, transfers, fares, frequency

Non-Motorized Skims: Walk/bike time and distance, safety indices, infrastructure quality

Demographic Data ¶

Person Attributes: Age, gender, income, employment status, disabilities

Household Context: Vehicle ownership, residential location, household structure

Revealed Preferences: Past travel behavior, mode choice patterns

Built Environment ¶

Zone Characteristics: Density, land use mix, walkability, parking supply

Transportation Infrastructure: Transit service, pedestrian facilities, bike infrastructure

Model Estimation and Calibration ¶

Estimation Data Sources ¶

Travel Surveys:

California Household Travel Survey (CHTS)
Regional travel behavior surveys
Transit on-board surveys
Workplace travel surveys

Revealed Preference Data:

Transit ridership counts by route and time period
Traffic counts by vehicle occupancy
Parking occupancy and pricing data
Emerging mobility usage patterns

Stated Preference Surveys:

Mode choice trade-off experiments
Technology adoption preferences
Policy response scenarios

Calibration Targets by Purpose ¶

Work Tours: Auto ~85%, Transit ~12%, Walk/Bike ~3%

School Tours: Auto ~65%, Transit ~15%, Walk/Bike ~15%, School Bus ~5%

Shopping Tours: Auto ~90%, Transit ~5%, Walk/Bike ~5%

Social/Recreation: Auto ~80%, Transit ~10%, Walk/Bike ~10%

Geographic Variation Targets ¶

Urban Core: Higher transit and non-motorized shares

Suburban Areas: Auto dominance with moderate transit usage

Rural Areas: Nearly complete auto dependence

Model Outputs and Integration ¶

Direct Model Outputs ¶

Tour Mode Assignments: Primary mode for each generated tour by person and purpose

Mode Share Statistics: Aggregate usage patterns by market segment and geography

Accessibility Measures: Mode-specific accessibility logsums for land use integration

Policy Sensitivity Indicators: Elasticity measures for infrastructure and pricing policies

Network Model Integration ¶

Auto Assignment: Vehicle trips by occupancy class for traffic assignment

Transit Assignment: Passenger loads by route, access mode, and time period

Active Transportation: Pedestrian and bicycle flows for infrastructure planning

Policy Analysis Applications ¶

Transit Investment Evaluation: Ridership response to service improvements

Pricing Policy Assessment: Behavioral response to tolls, parking fees, fare changes

Land Use Policy Integration: Mode choice sensitivity to development patterns

Technology Impact Analysis: Adoption and usage of new mobility services

Model Validation and Performance ¶

Statistical Performance Measures ¶

Goodness of Fit: Rho-squared statistics indicating model explanatory power

Parameter Significance: t-statistics and confidence intervals for coefficient estimates

Market Share Accuracy: Comparison between predicted and observed mode shares

Elasticity Validation: Realistic response to level-of-service and policy changes

Behavioral Validation Criteria ¶

Demographic Reasonableness: Realistic variations across population segments

Geographic Consistency: Appropriate urban/suburban/rural differences

Temporal Stability: Consistent behavior patterns across time periods

Policy Sensitivity: Reasonable response to transportation policies

Implementation Considerations ¶

Computational Efficiency: Execution time for large regional populations

Numerical Stability: Robust handling of extreme values and edge cases

Convergence Properties: Stability in iterative model feedback loops

Model Maintainability: Clear structure for parameter updates and enhancements

Integration with Other CT-RAMP Components ¶

Upstream Dependencies ¶

Mandatory Tour Frequency: Determines work and school tours requiring mode choice

Auto Ownership Model: Vehicle availability constraints on auto mode alternatives

Coordinated Daily Activity Pattern: Household coordination effects on mode choice

Downstream Integration ¶

Tour Time-of-Day Model: Mode choice logsums influence departure time preferences

Stop Frequency Model: Mode characteristics affect intermediate stop generation

Trip Mode Choice Model: Tour mode provides constraints and preferences for trip-level decisions

Feedback Mechanisms ¶

Network Congestion: Highway and transit level-of-service updates based on usage

Parking Availability: Dynamic parking costs and availability based on demand

Mode-Specific Accessibility: Land use model integration through accessibility measures

This comprehensive Tour Mode Choice Model serves as the central behavioral engine for transportation policy analysis in CT-RAMP, providing detailed representation of how travelers choose among available alternatives based on their circumstances and the transportation system characteristics.