Skip to Main Content
Peer Review Program
HomeUnited States Department of Transportation, HomeContact UsSite Map
Travel Model Improvement Program - TMIP
About TMIPTMIP ServicesClearinghouse, selectedConferences and CoursesContactsTravel Model DiscussionsTRANSIMSLinks

BROWSE

SEARCH

ORDER DOCUMENTS

NEWSLETTER

Clearinghouse

TMIP Newsletter Archive

TMIPConnection
The Travel Model Improvement Program Newsletter

Issue 7, November 1997

The Travel Model Improvement Program is sponsored by:
U.S. Department of Transportation
Federal Highway Administration
Federal Transit Administration

Table of Contents
TMIP Seminars
FHWA Directory of MPOs
Report Announcements
TRANSIMS Case Study – Executive Summary
TMIP Reports

TMIP Seminars

The availability of Advanced Travel Demand Forecasting seminars were announced in the previous newsletter. These seminars are intended to disseminate new techniques and help establish or strengthen professional contacts among travel demand modelers. We held our first seminar in Columbus, Ohio in July 1997. The seminar topics were land use modeling and the allocation of regional land use forecast to TAZs. The attendees were from state DOTs and MPOs in Ohio, Kentucky, Indiana, and Michigan. Paul Waddell, of the University of Washington, presented an overview of the topics, answered general questions and offered advice on specific problems confronted by the attendees.

TMIP staff is ready to offer similar seminars in the coming year, on land use or other topics of concern in your region. Please contact Gordon Shunk at (817) 277-5503 to discuss holding a seminar in your region.

Introduction to Travel Demand Forecasting Course

The TMIP program conducted a course on Introduction to Travel Demand Forecasting in Raleigh, North Carolina and Albany, New York in the summer and fall of 1997. This course covers the traditional four step planning processes: trip generation, trip distribution, mode choice, and traffic assignment. The course is intended for federal, state, and local planners new to planning that want to gain a better understanding of the principles and techniques of travel demand forecasting. Anyone interested in sponsoring a course in their region should contact Mike Culp at (202) 366-9229.

Advanced Travel Demand Forecasting Course

This course is designed for experienced travel demand modelers, generally those who have been practicing travel demand modeling for several years at metropolitan planning organizations, state Departments of Transportation, in consulting or academia, and in other environments. While it is recognized that auto travel dominates in many urban areas in the U.S., many procedures discussed in this course deal with non-auto modes such as transit or walking or with mode choice issues. However, substantial information in the course can be applied in any U.S. urban area or state. Anyone interested in sponsoring a course should contact Patrick DeCorla-Souza at (202) 366-4076 or Kim Fisher at (202) 366-4054.

Table of Contents

The Directory of Metropolitan Planning Organizations (MPOs) has Gone Electronic

The Metropolitan Planning Division of Federal Highway Administration has traditionally maintained a listing of all MPOs across the country. The information was available either as a paper copy or a file on a floppy upon request. Now the information is available on the web. The information may be found on the TMIP web page in the Communication Center. Any changes to the file should be reported to the web master, Lynette Engelke, whose email address is on the web site, or to Kim Fisher at (202) 366-4054.

Table of Contents

Report Announcements

Quick Response Freight Forecasting

The Intermodal Surface Transportation Efficiency Act of 1991 requires that States and Metropolitan Planning Organizations consider urban freight in their long-range plans, transportation improvement programs, and annual work elements. There are several impediments to effective freight planning which includes: very little experience in freight planning, limited data, and the models used historically are very complex. This manual provides background information, identifies data sources, describes simple techniques and transferable parameters to develop commercial vehicle trip tables, and techniques for site planning.

Time-of-Day Modeling

In recent years, there has been increasing interest in the ability of travel demand models to estimate travel not only for the average weekday, but for different periods within the day. Travel demand models are increasingly required to be analysis tools for a broad range of issues on transportation policy and project alternatives. These issues often require detailed analysis, not only spatially, but temporally as well. This report provides documentation on methods used in U.S. urban areas to handle the issue of time of day in their travel demand models. Commonly used practices and emerging approaches are described, and the most innovative methods used by metropolitan planning organizations and states are documented.

Urban Design, Telecommuting and Travel Forecasting Conference Report

This conference was sponsored by the Travel Model Improvement Program. The two principal goals of the conference were: to improve understanding of the influence on travel behavior of urban development patterns specifically designed to reduce motor vehicle travel and to assess the potential for telecommunications, particularly telecommuting, to reduce motor vehicle travel. The conference was charged with identifying what is already known and unknown about these effects, what knowledge can be applied today for use by Metropolitan Planning Organizations (MPO) and state Department of Transportation (DOT) planners, and what research and development on these subjects is needed to improve today's urban and transportation planning practices. Deliberations at the conference were organized in three subject tracks each of which addressed several specific questions related to its subject: Principles of Urban Design, Effects of Urban Design on Travel Behavior, and Effects of Telecommunication on Travel Behavior.

Activity Analysis Conference

http://tmip.fhwa.dot.gov/clearinghouse/docs/abtf/

The principal goal of this conference was to promote the use of activity-based approaches for travel forecasting. Corollary purposes were to identify activity-based forecasting techniques that can be used now and to recommend actions to advance the state-of-the-art. This report includes papers that document the keynote address and five other presentations in the plenary session. Following those papers are summaries of the discussions and recommendations in the three workshops:

The workshops began by considering techniques that are currently available for activity based travel forecasting. Gaps in the availability and work ability of those techniques were identified, and research and development were recommended to overcome those deficiencies.

Transfer Penalties in Urban Mode Choice Modeling

This is a report on recent research done on the subject of transfer penalties in urban mode choice decision-making. The research was undertaken to decide whether such penalties, thought by many observers to exist, could be quantified. If so, then the Federal Transit Administration has a firmer basis on which to gauge the merits of travel model sets and the forecasts of transit ridership that they produce.

Using Boston area household travel survey data and hand-coded transit impedance data, mode choice models containing transfer related variables were estimated by the Central Transportation Planning Staff. Transfer penalties for work trips were, in fact, found and quantified. In models based on an extremely carefully constructed data base, a transfer dummy variable representing whether a transfer is required to complete a trip, emerged as a modestly significant variable. It was equivalent to 12 to 15 minutes of transit in-vehicle time, depending on the particular model specification. Other interesting and useful information emerged from this research as well. It was found that hand-coding transit impedances and establishing liberal definitions of whether transit is an available mode had a significant impact on model estimation and resulted in more accurate parameter coefficients.

Global Positioning Systems for Personal Travel Surveys — Lexington Area Travel Data Collection Test

http://www.fhwa.dot.gov/ohim/lextrav.pdf PDF files require the use of a PDF reader. You may download a free copy of Adobe Acrobat Reader to view these files. If you already have the Adobe Acrobat Reader installed on your computer it should automatically open to view the files when you click on the link to the document.

Personal travel and how it changes is of continuing concern to transportation planners and policy makers. Transportation professionals and other users of the collected data surmise that people omit very short trips, like stopping at the post office or video store, using self-reported methods. Also, some data of interest such as route choice and travel by highway functional class are not accessible using traditional travel survey methods. This report describes a project to test an automatic data collection device (Global Positioning System - GPS) and self-reported information for the collection of personal travel data in Lexington, Kentucky.

Table of Contents

Summary of a Case Study of the TRANSIMS Microsimulation Module

Part of the TRansportation ANalysis SIMulation System in Dallas, Texas — November 1997

What is TRANSIMS?

TRANSIMS is a new system of travel forecasting models being developed by the Los Alamos National Laboratory for the Travel Model Improvement Program. The TRANSIMS models are a wholly new approach to travel forecasting, specifically designed to meet the needs of today's transportation decision makers for more accurate information on traffic congestion, differential impacts of transportation and motor vehicle emissions. TRANSIMS is composed of four basic modules.

four TRANSIMS modules

Activity-Based Travel Demand estimates the number, characteristics and locations of activities in which individuals will participate during the forecast period. Activities are work, shopping, recreation, etc. These activity estimates are based on characteristics of individuals, their households and vehicles determined by a synthetic population generator.

Intermodal Trip Planning computes combined route and mode trip plans to accomplish individuals' desired activities. Intermediate activities such as shopping may occur during the routing of a principal trip such as work. TRANSIMS maintains the identities and characteristics of individual drivers, vehicles and other travelers throughout their trips. Trips are identified by specific geographic points of origin and destination.

Traffic Microsimulation computes the movement of persons, goods and vehicles on the simulated transportation network second by second during the forecast period. The microsimulation continuously computes the operating status of all vehicles and engines throughout the trips, including locations, speeds, acceleration or deceleration. Every motor vehicle in the study area is monitored in this way, thereby indicating areas and times of traffic congestion and emission concentrations.

Air Quality Analyses identify the kinds of emissions and calculate the effects of emissions on the atmosphere in the study area. The air quality module estimates the nature, amount and conditions of emissions by each motor vehicle. The output of the emissions modules are consistent with airshed models.

TRANSIMS is a considerable departure from traditional, four-step travel forecasting procedures. The new technical approaches in TRANSIMS permit equity analyses of transportation alternatives, service reliability and forecast uncertainty. Moreover, all of the traditional analyses conducted by the best current four-step models can be conducted with TRANSIMS.

Executive Summary

This report summarizes the procedure, results and conclusions of a case study application of the TRANSIMS microsimulation procedure in Dallas, Texas. The microsimulation is part of the larger TRANSIMS transportation simulation system. The purpose of this case study was to test the capabilities of the TRANSIMS microsimulation to track the movement of simulated individual vehicles through a simulated representation of a transportation network. The case study successfully demonstrated those capabilities using real travel data to address actual planning issues. An additional case study now under- way in Portland, Oregon, will test other components of TRANSIMS.

The case study examined the degree to which two different kinds of roadway improvements would relieve traffic congestion in an intensely developed suburban area in north Dallas. The case study area is served by north-south and east-west freeways that intersect near office buildings, hotels and the largest shopping mall in Dallas. The roadway network changes considered were (1) adding a lane in each direction to one of the freeways and (2) modifying arterial street operations, intersections and capacity. The microsimulation results described the effects on traffic flow of the two improvement options. The network performance measures compared for the two improvement options and the base existing condition were median travel time, network reliability and traveler uncertainty. These performance measures were prepared for three population subgroups: travelers with destinations at the shopping mall, travelers to or from other locations in the study area and travelers passing through but not stopping in the study area.

The simulation study results indicated that both improvement options reduced the median travel times from those in the base existing condition for all persons except those traveling non-stop through the study area. The reductions in travel times were about equal for the two improvement options, but the effects of the arterial improvements were observed about half an hour earlier than the effects of the freeway improvements. Both improvement options reduced travel times for trips that did not stop at the shopping mall, and those reductions were nearly equal. The arterial street improvements effected greater travel time reductions than the freeway improvements for persons with destinations at the shopping mall. Travel times resulting from the arterial street improvements were more reliable than for the freeway option. These results must be balanced in consideration of the magnitude of the arterial and freeway system investments implied by the two improvements.

Four important conclusions were demonstrated by the microsimulation applications in this case study:

  1. the simulations successfully reproduced observed traffic patterns in the study area;
  2. effects of both infrastructure and operational improvements were demonstrated;
  3. estimates of travel time reliability were developed; and
  4. impacts of transportation system changes on population sub-groups were available.

The remainder of this report describes the Dallas case study and results in more detail. A major technical report including the detailed results of the simulations and tests on the models will be available later this year.

Description of the Case Study

The case study examined traffic in a 25 square mile area centered on the Galleria shopping mall, the largest mall in Dallas. Other development in the study area includes Valley View shopping mall and several high-rise hotels and office buildings. The case study simulated approximately 200,000 vehicle trips into, out of and through the heavily congested study area during four hours of the morning peak with special attention to trips to and from the Galleria.

Microsimulations were run for two different roadway network improvement options and compared to a base case microsimulation of the existing situation. Figure 1 displays the principal roadways in the study area.

Roadways in the Case Study Area

Figure 1 - Roadways in the Case Study Area

Interim Methodology

This case study tested the TRANSIMS microsimulation module. The data used by the microsimulation would normally be produced by the Trip Planner, but the full Planner was not operational at the time of this test. Therefore a vehicle trip table developed by NCTCOG1 was used here by the microsimulation as a surrogate for the activity based trip plans. Future applications of TRANSIMS will use person trips generated by the activity generator and the Trip Planner.

The trips produced by the Planner were described by their geographic points of origin and destination. For this case study the regional transportation zones of trip origin and destination were disaggregated, and trips were assigned to specific geographic locations in their zones based on statistical analysis of actual activity locations. The trips were also assigned second-by-second starting times based on a trip starting time distribution provided by NCTCOG.

For this case study a trip router identified the minimum time paths for all trips, initially based on free traffic flow conditions, and assigned trips to those minimum time paths. Then the simulation moved individual vehicles along those paths (considering posted speed limits and intersection traffic controls). The traffic speeds observed from the first simulation were fed back to adjust the original speeds on network links to reflect congestion effects. A portion of the trips were then assigned to new minimum time paths based on the adjusted trip speeds. This procedure was repeated until the overall iterated solution between the router and simulation produced reasonable traffic patterns.

Calibration and Validation

In TRANSIMS, calibration and validation are somewhat different concepts than usually understood by transportation planners and are accomplished using different techniques. Calibration adjusts simulated traffic dynamics resulting from the simulation to observed or standard traffic dynamics reported in references such as the Highway Capacity Manual. The microsimulation calibration for this case study was accomplished by assigning both simulated and observed traffic to simplified calibration networks that represented selected portions of the base case and alternative networks. The characteristics of simulated traffic, such as flow-density relationships and the number of left turns across oncoming traffic, were compared to characteristics observed in the study area and reported characteristics. The microsimulation in this case study matched the standards satisfactorily.

Validation compares simulated traffic volumes with actual traffic counts on the study network. This case study showed reasonable agreement between the simulated counts and the estimated volumes for all roadway classes except collectors, for which NCTCOG considers its estimates may be low.

Results

Traditional Measures

Figure 2 shows the median travel times at different points during the four-hour period simulated. The travel times are the medians of all trips initiated in the five-minute period that started at the time indicated. Both roadway improvement options produced better median travel times than the base case from before 6:00 a.m. until after 8:30 a.m. The most pronounced differences occurred between approximately 7:30 a.m. and 8:30 a.m. Travel times for the freeway improvement option were better than for the arterial improvement option until approximately 8:05 a.m.; after that the arterial option dominated.

Figure 2 shows the median travel times at different points during the four-hour period simulated

Figure 3a shows little difference in vehicle miles traveled except for being slightly greater for the freeway improvements after approximately 7:30 a.m. There was a more pronounced difference in vehicle hours traveled (VHT) shown in Figure 3b for both improvement options. The effect of the arterial improvements was especially great.

Figure 3a shows little difference in vehicle miles traveled except for being slightly greater for the freeway improvements after approximately 7:30 a.m. There was a more pronounced difference in vehicle hours traveled (VHT) shown in Figure 3b for both improvement options. The effect of the arterial improvements was especially great.

Figure 4b shows the effects of the roadway improvement options on travel times for persons traveling to or from the Galleria, and Figure 4a shows the effects on other persons traveling into, from, or through the study area. The effects of the two improvement options on non-Galleria travelers were much greater than for the Galleria travelers, but both were better than the base case situation. Figure 4a and 4b show another interesting situation. The "non-Galleria-only" line in Figure 4a shows that both of the improvement options yielded better travel times than the base case situation if only non-Galleria travelers were on the network. On the other hand, the "Galleria only" line in Figure 4b shows that the travel times for Galleria travelers were much better for the base case than for either improvement if the Galleria travelers were the only ones on the network. This kind of analysis would be difficult if not impossible to accomplish with conventional four-step travel forecasting models but is readily possible with microsimulation.

Figure 4b shows the effects of the roadway improvement options on travel times for persons traveling to or from the Galleria, and Figure 4a shows the effects on other persons traveling into, from, or through the study area. The effects of the two improvement options on non-Galleria travelers were much greater than for the Galleria travelers, but both were better than the base case situation. Figure 4a and 4b show another interesting situation. The non-Galleria-only line in Figure 4a shows that both of the improvement options yielded better travel times than the base case situation if only non-Galleria travelers were on the network. On the other hand, the Galleria only line in Figure 4b shows that the travel times for Galleria travelers were much better for the base case than for either improvement if the Galleria travelers were the only ones on the network. This kind of analysis would be difficult if not impossible to accomplish with conventional four-step travel forecasting models but is readily possible with microsimulation.

Statistical Measures

The TRANSIMS microsimulation produces statistical measures not commonly available from the four-step travel forecasting process. One of these, "variability," describes the amount of variation or spread expected between different estimates of a variable calculated under the same conditions. (This is identical to the term "variance" used in statistics.) Uncertainty is another measure produced by TRANSIMS and is the amount of variability a traveler should expect in any individual estimate such as travel time. A third measure is reliability, which is the amount of variation in travel time the individual traveler should expect to encounter in his or her trip. The application and results of these measures in the Dallas case study are described below.

Variability Analyses

Common travel forecasting methods depend on averages and expected values of such key variables as link travel volumes, travel times and speeds. Those averages are calculated over all vehicles using the links during the time period of interest. Estimates of those averages are based on simple relationships between average speed and link traffic volume and capacity. In contrast, TRANSIMS simulates actions of individual vehicles (updated every second) to generate distributions of instantaneous vehicle speeds that can be averaged over any user-defined time period to calculate travel time. Those distributions are sensitive to local traffic conditions, signal timing and other traffic controls. Because TRANSIMS tracks individuals' trips throughout the roadway network, distributions of speeds and other traffic operating characteristics can be obtained for any operational situation. Therefore TRANSIMS can produce statistical variability measures to address questions such as:

Reliability Analyses

The basic notion of reliability is whether travelers' day-to-day trip times will meet their expectations. For example, if travelers experience variation in the time it takes to make the same trip, using the same route, under the same conditions from one day to the next, they will likely question the network dependability.

That is, those travelers are not very confident that they can accurately predict, day-to-day, how long their trips will take. Transportation planners must decide what percentage of the population who view their travel time as unreliable is acceptable and how large a variation in travel time is satisfactory.

TRANSIMS readily computes reliability measures, which are relevant and frequently useful in real traffic systems–especially those operating at or near capacity. The ability to be able to compare proposed roadway improvement alternatives using measures such as reliability is an attractive attribute of TRANSIMS and other systems that track individual travelers. These measures will allow planners to focus better on segments of the population likely to be dissatisfied about differences experienced in travel times on repetitive trips such as home-to-work. If the percentage of travelers in this category is large enough and the variability in travel time is great enough, improving reliability may be more important than improving average travel time for the overall population.

Figure 5 shows the difference between expected and actual travel times for the five percent of the population experiencing the largest such differences. This measure is a surrogate for congestion. Travelers in this group are the least confident that their expected travel times for given trips will be realized. In this case study, between 7:00 a.m. and 9:00 a.m. the arterial improvement option demonstrated much less difference in expected travel times than either the freeway improvement or the base case. Hence, the arterial street improvements yielded the more reliable network. The freeway option was better than the base case until shortly after 8:00 a.m. Also of interest is the greater than expected difference in travel time from the freeway option as late as 9:00 a.m., after traffic flow on the other two roadway options had stabilized. This difference is an indication of continued congestion in the study area even if the freeway itself is improved.

Figure 5 shows the difference between expected and actual travel times for the five percent of the population experiencing the largest such differences.

Uncertainty Analyses

Like most situations, TRANSIMS only approximates reality and will produce different approximations each time it is run, even if input data and assumptions do not change. Accordingly it is important to understand how much of the variation in simulation results is attributable to the structure of the microsimulation module. This variation is termed uncertainty. Knowing the uncertainty expected from a particular simulation is important for testing whether differences between results of simulations where inputs changes are statistically significant. For example, if a particular algorithm represents a random process, we would like to know how changing a random number seed with affect the simulation results. If the resulting effects are large compared to effects of changing input data or other conditions, one would be much less confident that changing input data actually caused the difference.

This case study included considerable effort to identify and quantify the degree of uncertainty in the simulation results. Figure 6 shows median travel times for three simulations of the base case situations and for the two improvement options. (The plots for the two improvement options in Figure 6 are identical to those in Figure 2.) The uncertainty in travel times was estimated by randomly varying trip day, start time and routing for the base case. Uncertainty due to trip start times was estimated by randomly selecting trip start times according to a trip time distribution observed in an NCTCOG travel survey. Uncertainty in routing was estimated by randomly selecting from a set of alternative trip plans. Uncertainty in day-to-day variability was estimated by selecting different random number seeds.

The plots in Figure 6 show the possible size of uncertainty in travel times for the base case. Those results are due to intrinsic computational properties of the TRANSIMS microsimulation module. Because the median travel times of both roadway improvement options tested were less than the uncertainty of the base situation, both improvements can be considered effective for improving travel times to a degree greater than the uncertainty of the underlying simulation system. Where this is not the case, there would be less confidence that the impact of the simulation roadway improvements is significant.

The plots in Figure 6 show the possible size of uncertainty in travel times for the base case.

Tracking Individuals

TRANSIMS has the capability to track and accumulate information on individual travelers as they proceed through the transportation network. The information for individuals can then be grouped for analysis based on similar characteristics of the groups. The groups for which comparisons may be of interest could have similar travel characteristics such as origins or destinations, length of trips or traveler demographics. As TRANSIMS is extended in the future to include environmental and other kinds of impact analyses, it will become possible to measure effects such as pollution, energy utilization, etc., on such groups. While traditional equity analyses usually discriminate such groups based on characteristics of their databases (e.g., zone of origin, type of trip, etc.), TRANSIMS has the capability to discriminate groups based on characteristics of individual travelers. For example, it could be asked if there is any correlation between the groups of travelers experiencing a large variability in travel times day-to-day, where they live, what routes they have in common, etc.

The case study examined the effects of the alternative roadway improvements on two different groups: travelers destined to the Galleria area and all other travelers in the study area. As shown here previously (Figure 4), non-Galleria and Galleria travelers benefitted from both roadway improvement options, but non-Galleria travelers benefitted more from both improvements than did Galleria travelers. Galleria-bound travelers benefitted more from the arterial option than from the freeway option.

Conclusions

The results reported here demonstrate that TRANSIMS can produce the same information as traditional four-step travel forecasting models but with several advantages. The TRANSIMS microsimulation produces measures of variability in traveler decisions under similar conditions. The reliability of travel time can be estimated from the microsimulation results. The uncertainty of travel time can be obtained from these results as well. Finally, the ability to track individuals and their characteristics provides the opportunity to understand the effects and effectiveness of transportation improvements.

This case study is an important first step in the evolution of the full TRANSIMS travel forecasting system. It demonstrates conclusively that the capabilities of the microsimulation module can produce relevant and useful information about travel forecasts. It further demonstrates important additional capabilities that will permit expanding the kind and amount of information available from travel forecasts for understanding and predicting traffic congestion and air quality effects of transportation improvements. The microsimulation now provides the essential platform for the remainder of the TRANSIMS forecasting modules. Obviously additional work will be performed, including sensitivity analysis of microsimulation elements, impacts of network detail on microsimulation results and refinements of performance measures used in TRANSIMS.

The computer configuration on which this case study was processed consisted of five Sun Sparc 5 and 20 workstations. These were connected by and ethernet LAN. The transportation network for the study area was divided among the work stations based on the amount of available computing capacity. Computing capability has already progressed beyond the environment used for this case study and is expected to be even greater by the time the full TRANSIMS model system is operational and available to users.

Next Steps

The next steps in further developing the TRANSIMS microsimulation procedure are to review and digest the results of the case study and begin additional testing by groups outside the Los Alamos team. NCTCOG will review the results of the case study in detail and perform sensitivity tests on those results. They will also produce additional simulations to test other features of the procedure, including development of transportation performance measures. It is also anticipated that several universities will want to analyze the simulation procedure and incorporate it into their research programs and teaching curricula.

The next phase of the TRANSIMS project has begun. That work will develop the trip forecasting capabilities that will feed the microsimulation and will test the air quality module. The data from travel surveys in Portland, Oregon, will be used to define the travel behavior components of TRANSIMS. It is expected that the trip planner will be based on activity patterns identified in the Portland surveys. The characteristics of Portland's transit system will be used to develop the multimodal components of TRANSIMS.

The Los Alamos team has been packaging the interim microsimulation capability used for the Dallas case study for release to a limited number of sites for testing and research feedback. This package includes an updated microsimulation with improved driving logic and enhanced portability. In addition, the package includes a plan preprocessor to select individual plans from a plan set, a plan viewer and a microsimulation output viewer.

The package for release will include the networks used in the Dallas case study and the final trip plan sets. The package will also include calibration networks and plan sets for microsimulation calibration against four typical situations that limit network capacity.

The package is designed to run on either multi-processor Sun workstations or a local area network of Sun workstations. Use of a third party software license is limited in the package, and package users will be required to sign a license agreement.

Information about the Travel Model Improvement Program (TMIP), the TRANSIMS forecasting system and the TRANSIMS microsimulation procedures can be obtained by visiting the TMIP web site or by contacting Lisa Day at (817) 277-5503 or Kim Fisher at (202) 366-4054.

Endnotes:
1 NCTCOG: North Central Texas Council of Governments, Metropolitan Planning Organization for the Dallas/Fort Worth metropolitan area.

Table of Contents

TMIP Reports

Reports without DOT numbers can be ordered from:

TMIP Report Requests
400 7th Street, SW
HEP-22, Room 3232
Washington, DC 20590
E-mail: kim.fisher@fhwa.dot.gov

Reports with DOT order numbers should be ordered from:

U.S. Department of Transportation
Ardmore East Business Center
3341 Q 75th Avenue
Landover, MD 20785

Table of Contents

To subscribe to this free newsletter send an e-mail to TMIP@tamu.edu or contact Gary Thomas at (ph.) (979) 458-3263, (fax) (979) 845-6001, (mail) Gilchrist, Room 112, Texas Transportation Institute, Texas A&M University System, 3135 TAMU, College Station, TX 77843-3135