 |  |  |  |
Dynamic Activity-Based Travel Forecasting System
Principal Investigator:
John Gliebe, Resource Systems Group, INC
Summary:
Metro, the regional transportation planning agency for Portland, Oregon, has long been a leader in the advanced travel demand modeling practices in the U.S.. In recent years, the questions being asked of modelers at Metro have sometimes exceeded the output capabilities of their current trip-based systems. Past research projects at Metro have involved more academically-oriented model development ...
Metro, the regional transportation planning agency for Portland, Oregon, has long been a leader in the advanced travel demand modeling practices in the U.S.. In recent years, the questions being asked of modelers at Metro have sometimes exceeded the output capabilities of their current trip-based systems. Past research projects at Metro have involved more academically-oriented model development efforts, but have yet to result in the development of a practical tool for every day use. In response to current needs, Metro has identified at least three critical objectives for near-term model development:
• to consider trips as components of larger tours;
• improved sensitivity to stimuli that affect time-of-day decisions; and
• better model response to pricing signals.
The proposed research project has as its primary goal the development of a dynamic activity-based demand model system for Metro that will be capable of meeting these objectives through explicit consideration of time of day and accumulated activity times in the propensities of individuals to construct tours. The success of this project should make it possible to achieve true integration between the proposed demand model system and dynamic network assignment methods, and possibly even microscopic simulations.
Although activity-based travel demand models have been developed or are currently under development in several cities in the U.S. and elsewhere, sensitivity to time-dependent path information seems to be lacking in these efforts. Specifically, extant models tend to treat activity episode generation, duration, location, starting time, and travel mode choices as essentially independent, which they are able to do because they ultimately produce trip tables for static network assignment methods. In recent work by the PI and Metro on the FHWA-sponsored TRANSIMS demonstration project, it was discovered that internal consistency between these interrelated choices is essential to realistic outcomes when applied in a microsimulation framework that provides time-dependent travel cost information to the demand model.
The approach to be followed in this research is based on recent work by the PI on a microsimulation model of commercial activity-travel, using dynamic discrete choice methods, as well as on previous experience with activity-based model research and development. Model components to be developed under this project include:
• Activity pattern choice
• Daily starting time choice
• Tour generation
o Tour mode choice
o Next stop purpose
o Next stop location
o Next stop mode
o Next stop timing
• System simulation event tracker
This project will be a collaborative effort between the Metro modeling staff and the PSU research team. PSU will work directly with the Metro modeling staff throughout the design, development, application and testing process. Metro staff will receive hands-on training in advance modeling methods—specification, estimation, calibration—and at the same time will lend their practical knowledge to the PSU researchers, providing valuable experience for graduate student researchers.
It is the intent of this project to present the research-worthy results of this work at conferences and in papers submitted to refereed journals, and to facilitate the transfer of modeling technology to other potential users.
See More
Project Details
| Project Type: | Research |
|---|---|
| Project Status: | In Progress |
| End Date: | February 16, 2009 |
| UTC Funding: | $35,616 |