Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 40
THE SAN FRANCISCO MODEL IN PRACTICE 29
the neighborhood. The model suggested two interesting but rather in the road and transit skim-building and
findings. The first was that anyone who makes trips to assignment procedures. The desire to decrease random
or from the Tenderloin, for any reason, chose walking, microsimulation variation (by running multiple itera-
transit, and bicycling at greater rates than the average tions), combined with the highly granular nature of the
San Franciscan. Trips to and from the Tenderloin were skim-building and path-building steps, made obvious the
about half as likely to be made by car as the average San need for a parallel structure instead of the existing top-
Francisco trip. down model process.
The other interesting finding was that when non- SFCTA devised a job control system to allow a model
Tenderloin residents' trips were included in the totals, car job to be submitted as a transaction, which would then
use increased. This indicated that non-Tenderloin resi- be processed by all available machines as quickly as pos-
dents who make a trip to the Tenderloin--for work, sible, in parallel. The most difficult aspect of this process
social activities, or any reason--were one-third more was analyzing the dependency tree of model steps to
likely to use a car than a Tenderloin resident. This finding determine which ones could be made parallel and which
suggested that about one-third of the cars destined for the could not; some steps obviously required that earlier
Tenderloin were from outside the neighborhood. The actions be complete before the steps could be made par-
greater use of cars by non-Tenderloin residents was even allel. Job files were rewritten to unlink the pieces that did
greater when only work trips were analyzed. Employees not depend on each other. The revised job files were
who work in the Tenderloin, but live elsewhere, were passed to a new dispatcher utility program that could
more likely to drive into the Tenderloin for work. The allocate each step to available computers and keep track
auto mode share for all San Francisco residents with ori- of the model run progress.
gins in the Tenderloin (35.4% for work trips) was double The extraordinary time saving of this method was
the automobile mode share of trips made only by Tender- limited only by the amount of hardware available and
loin residents (17.7%). This difference can be explained the granularity of the model steps. In practice, full runs
by a large number of Tenderloin workers who commuted shortened from 36 to 9 h. The goal of an overnight run
from outlying neighborhoods by private automobile. The thus attained, staff added five additional core iterations
specific characteristics of residents versus nonresidents to reduce error due to microsimulation variability. The
making trips in the neighborhood were easy to analyze model now runs in just under 12 h.
because of the disaggregate nature of the SF-CHAMP
outputs, which thus provided a new way of using model
results to support planning project work. REFERENCES
1. Cambridge Systematics, Inc. San Francisco Travel Model
New Starts Development Executive Summary. San Francisco County
Transportation Authority, San Francisco, Calif., June 30,
SFCTA developed an application of the San Francisco 2001.
model to the proposed New Central Subway project in 2. Cambridge Systematics, Inc. San Francisco Travel Model
downtown San Francisco (5). This is the first application Development Report on Validation of 1998 Models. San
of a tour-based travel demand model in the United States Francisco County Transportation Authority, San Francisco,
to a major infrastructure project in support of a submis- Calif., May 7, 2001.
sion to the FTA for project funding through the New Starts 3. Cambridge Systematics, Inc. San Francisco Travel Model
program. To enable the submittal of a New Starts request, Development Report on MTC Consistency. San Francisco
software was developed to collapse the microsimulation County Transportation Authority, San Francisco, Calif.,
output of the models for tour and trip mode choice into a May 15, 2001.
format compatible with the FTA SUMMIT program. 4. Castiglione, J., R. Hiatt, T. Chang, and B. Charlton. Appli-
SUMMIT was then successfully used to summarize and cation of Travel Demand Microsimulation Model for
analyze user benefits accruing to the project and to pre- Equity Analysis. In Transportation Research Record: Jour-
pare an acceptable New Starts submittal. nal of the Transportation Research Board, No. 1977, Trans-
portation Research Board of the National Academies,
Washington, D.C., 2006, pp. 3542.
Parallel Processing 5. Freedman, J., J. Castiglione, and B. Charlton. Analysis of New
Starts Project by Using Tour-Based Model of San Francisco,
The initial implementation of the SF-CHAMP model California, In Transportation Research Record: Journal of the
took 36 h to run, which became a major impediment to Transportation Research Board, No. 1981, Transportation
both further model development and application. The Research Board of the National Academies, Washington,
bulk of this time was not in core microsimulation steps D.C., 2006, pp. 2433.
