New Features of Applied Activity-based Models in US

New Features of Applied Activity-based Models in US


P Vovsha, Parsons Brinckerhoff, US


This paper describes the structure, implementation, and application experience of seven different regional Activity-Based Models (ABMs) that share the Coordinated Travel - Regional Activity Modeling Platform (CT-RAMP) design and software platform.


This paper describes the structure, implementation, and application experience of seven different regional Activity-Based Models (ABMs) that share the Coordinated Travel - Regional Activity Modeling Platform (CT-RAMP) design and software platform. The CT-RAMP models are characterized by a number of advanced features, including a full simulation of travel decisions for discrete households and persons; explicit tracking of time in half-hourly increments and use of time constraints on the generation of travel; and explicit intra-household interactions across a range of activity and travel dimensions.
The following main modules are included in the basic CT-RAMP design: 1=Population synthesis, 2= Long-term choices, 3=Coordinated Daily Activity-Travel Pattern (CDAP) for all household members, 4=Tour-level choices, 5=Trip-level choices. Modules 3-5 are interlinked through time-space constraints.
The first ABM of the CT-RAMP family was developed in 2004 for the Mid-Ohio Regional Planning Commission (MORPC) located in Columbus, OH. The Columbus core model structure was adapted for the Tahoe Regional Planning Agency (TRPA) in 2006. The Tahoe ABM included special components to account for the seasonal variability in the Lake Tahoe Area?s population and travel moving to/from/through the model?s boundary.
The third and forth ABMs of the CT-RAMP family have been developed in parallel for the Atlanta Regional Commission (ARC) and the San Francisco Bay Area?s Metropolitan Transportation Commission (MTC). These ABMs were completed in 2009 and included many refinements:
? Simultaneous CDAP for all household members that accounts for interactions between them.
? Modeling of multiple intermediate stops on each half-tour (up to four).
? Implementation of distributed values of time.
Three new members of the CT-RAMP family were added in 2008 and 2009, including: San Diego, CA, for the San Diego Association of Governments (SANDAG); Phoenix, AZ, for the Maricopa Association of Governments (MAG); and Jerusalem, Israel, for the Jerusalem Transportation Master Plan Team (JTMT).
The San Diego ABM development started in late 2008 with the following new features incorporated:
? Improvement of the structure and segmentation of long-term models for workplace and school location using detailed information provided by the Land-Use model.
? Fine spatial resolution for all location choices at a level of 33,000 Master Geography Reference Areas. Transit procedures and mode choice are among the primary beneficiaries.
? Improved ABM system integrity by inclusion of a wide set of accessibility measures in upper-level models for car ownership, CDAP and tour generation. ?Flat? area-type dummy variables (e.g., CBD, urban, suburban) were completely avoided.
? Population synthesizer that incorporates both household-level controls (size, income group, dwelling type) and person-level controls (population distribution by age brackets).
The Phoenix ABM development started in mid 2009 and included the following new features:
? Addressing seasonal fluctuations in travel demand explicitly. The model system will have a switch that allows for representation specific to summer, winter, fall or spring.
? Accounting for non-residents explicitly in the population synthesis and subsequent chain of travel models.
? Sub-models for university-related travel addressing differences between students living in the households vs. student living in shared non-family households and group quarters. Arizona State University is the largest higher-education learning center in the United States, with more than 62,000 students.
? Special Events integrated with the core travel model in a disaggregate fashion; participation in a special event is organically incorporated in the individual daily pattern.
? Extended sub-models for joint travel including escorting children to school and carpooling between workers.
The Jerusalem ABM development started in 2008. The following new features can be mentioned:
? Wide range of modeled mobility attributes including possession of a driver license, transit pass holders, transit ticket discounts and/or subsidies from the employer or school, employer provided transportation for commuting and/or parking, school bus availability, etc.
? A model that allocates household cars to individual tours and creates a logical linkage across mode choice decisions has been developed.
? Parking choice and constrained parking equilibrium that makes the model sensitive to constraints and policies associated with parking. The model can portray the dynamics of parking in each traffic zone during the day.


Association for European Transport