Argonne National Laboratory Transportation Research and Analysis Computing Center DOE Logo

TRANSIMS: Applications and Development Workshop

April 8–9, 2010

Building An Integrated Activity-Based and Dynamic Network Assignment Model For Sacramento Area Council of Governments (SACOG)

Brian Grady (RSG) and/or Joe Castiglione (RSG)

List of Authors
================
Joe Castiglione, Resource Systems Group, Inc.
29 Belmont Street, Somerville, MA 02143
jcastiglione@rsginc.com

Brian Grady, Resource Systems Group, Inc.
55 Railroad Row, White River Junction, VT 05001
802-295-4999
bgrady@rsginc.com

John Bowman, Transportation Systems and Decision Sciences
29 Beals Street, Brookline, MA 02446
John_L_Bowman@alum.mit.edu

Mark Bradley, Mark Bradley Research and Consulting
524 Arroyo Avenue, Santa Barbara, CA 93109
Mark_bradley@cox.net

Stephen Lawe, Resource Systems Group, Inc.
55 Railroad Row, White River Junction, VT 05001
slawe@rsginc.com

Abstract
=========
This paper describes the integration of DaySim, an activity-based travel demand forecast model developed for the Sacramento region, with the TRANSIMS Router, a disaggregate dynamic network assignment tool. The objective is a fully disaggregate and spatially and temporally detailed model system, where individual travel behavior is simulated from longterm travel choices such as usual work location through the selection of specific paths through the regional network for each individual trip segment. Benefits of this disaggregation include the ability to use individuals' characteristics to explain travel behavior, more accurate measures of level-of-service especially for short trips of all modes, and the ability to reflect time-sensitive travel conditions and pricing policies.

The development of activity-based demand models and TRANSIMS have largely followed separate paths for the last 13 years. Activity-based models have been unable to achieve their full potential because they have been integrated with traditional equilibrium assignment models, which discard the behavioral, spatial and temporal detail provided by activity-based models such as DaySim. TRANSIMS has been unable to achieve its full potential because implementation efforts have encountered difficulties developing and integrating a behaviorally based activity and travel demand component. This model integration effort sought to overcome both weaknesses by integrating the existing Sacramento Area Council of Governments (SACOG) DaySim activity-based model with TRANSIMS.

DaySim synthesizes a population for the entire Sacramento region and simulates a detailed itinerary for each person in that population. DaySim is comprised of a series of submodels that simulate long-term choices such as usual work and school location and household auto availability, as well as short-term choices such as the number and purpose of tours and stops, the parcel-level destinations of these tours and stops, the travel modes used to access these destinations, and the timing of travel. The model system is implemented using Monte Carlo simulation.

The Router determining paths for individual vehicle trips using volume-delay functions, and the Microsimulator simulating the interactions of these vehicles on the transportation network. These tools can be used to estimate roadway performance measures at finer spatial and temporal resolution than is possible with the traditional equilibrium assignment. Because using disaggregate demand models in conjunction with dynamic network models is still largely uncharted territory, for this effort we focused on the integration of DaySim with the TRANSIMS Router, and did not address Microsimulator performance.

Biography
==========
Joe Castiglione (RSG) is a senior project consultant in RSG's Travel and Land Use Forecasting practice, with expertise in advanced travel demand forecast model development and application and transportation planning. His experience includes a particular focus on the development and application of advanced activity-based models, as well as extensive practice in model calibration, refinement and integration. With RSG, Mr. Castiglione is participating in activity- based model development, enhancement, and implementation efforts in the Seattle, Jacksonville, and Sacramento areas. He has also been actively engaged in linking activity-based travel demand models with detailed regional traffic simulation models. Before joining RSG, Mr. Castiglione led model development and application efforts in both the public and private sectors in the San Francisco and Boston regions.

Stephen Lawe (RSG) is the Managing Director of RSG's Travel Demand Modeling practice and a modeler with over 20 years experience. Over the years, Mr. Lawe has worked on several advanced modeling practices ranging back to integrated GIS-based land-use models when GIS was first introduced. Currently, he leads a research effort at the University of Vermont integrating UrbanSim with DaySim (an activity based model), TRANSIMS (a supply side microsimulator), and MOVES. He is also developing advanced techniques for modeling carbon emissions, and is managing the SACOG DaySim activity based model-TRANSIMS integration project. Prior to the FHWA-funded project for SACOG, Stephen served on the TMIP peer review panel for the SACOG Activity Based Model implementation.

Brian Grady (RSG) develops, calibrates, and applies regional land-use and transportation models. Prior to joining RSG, Mr. Grady worked at Smart Mobility, a firm that specializes in sustainable transportation planning and project development that is responsive to environmental concerns, land-use goals, and current transportation policies and programs. Mr. Grady has worked on scenario planning and visioning projects that addressed numerous concerns, including sprawl, land consumption, equitable economic opportunity, and development of new transportation infrastructure. Mr. Grady has also worked on several advanced modeling projects that utilized activity-based demand models and dynamic traffic assignment supply models. He also has special expertise in the review and critique of travel demand models, particularly as they relate to the National Environmental Policy Act (NEPA) process and air quality conformity.

Mark Bradley has been developing and applying tour-based travel demand models for more than 20 years, the first ten years in Europe (beginning with the Dutch National Model in 1984), and the last ten years in the US. He is well known for bridging the gaps between theory and practice by providing practical and efficient implementation of state-of-the-art modeling techniques.

John Bowman earned MST and PhD degrees at MIT, with concentrations in advanced econometrics and demand modeling, studying under faculty advisor Moshe Ben-Akiva. In his academic research at MIT, Bowman developed the activity schedule approach that is used in many of the activity-based model implementations in the United States. Since 1998 he has been actively involved in the enhancement and practical deployment of this approach, and he continues to teach graduate level demand modeling occasionally at MIT.

 
U.S. Department of Energy Office of Science | UChicago Argonne LLC
U.S. Department of Transportation
Privacy & Security Notice | Contact Us | Site Map