Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
While progress is being made in many areas, it is slow. There is also wide diversity among MPOs in the use of models, available resources, and staff capabilities. A total of 11 out of 228 responding agencies reported using destination choice models. Fewer than 50% of large MPOs distribute person trips rather than vehicle trips. Approximately 75% of smaller MPOs and one- third of large MPOs do not use impedances in trip distribution that vary by time of day. These results indicate there is still a long way to go to implement what most of us regard as the state of the practice. Only about half of the responding MPOs reported using K- factors or any zone- specific adjustments to trip distribution models. The lack of independent data was frequently noted as the reason for not using K- factors or other zone- specific adjustments. A total of 22 MPOs reported being engaged in New Start transit planning, but do not have mode choice modeling capabilities. Approximately 80% of large MPOs, 40% of midsized MPOs, and a few small MPOs reported feeding back travel times from assignment into distribution. Only 40% of large MPOs reported feeding back travel times to land use or automobile ownership models. Goods movement is emerging as a very important public policy issue. The survey results indicate that approximately half of the small and medium- sized MPOs and 80% of large MPOs model truck trips. Fur- ther, 20% of the respondents reported using synthetic trip tables and 30% reported using factoring procedures. Some 50% reported using other methods, including bor- rowing coefficients from other regions. Approximately 25% of the MPOs reported using truck models that are more than 10 years old. Despite decades of discussions in the literature, only one MPO reported using an activity- based model set and two reported trying tour- based modeling but abandoned that approach. The vast majority of MPOs stated they have no interest in trying those approaches. Most of the MPOs specifically reported seeing no reason to consider changing their current practice. This result suggested there is still a wide gap between the state of the art and the state of the practice. The survey results indicated that validation is not con- ducted at all by most agencies. Where something called validation was performed, it usually consisted of com- paring model outputs across a screenline with ground counts, but often using the same data that were used to calibrate the model. Statisticians would tell us that this approach is flawed as method of validation. Fewer than 10 agencies nationally demonstrated statistically appro- priate validation procedures. The committee identified a number of pressing issues. These included error propagation through chains of models, poor representation of the price or the cost of travel, and poor representation of goods movement. Fur- ther, the committee found that point estimates provide very poor support of policy making. Existing models are difficult to apply to new policy issues, such as evacuation scenarios addressing terrorism or hurricanes. The com- mittee also found poor representation of nonresident travel, such as conventions and tourism, which are a growing percentage of the trips made in some cities. These individuals have different travel patterns, use dif- ferent modes, and stay in different areas than residents. We do a poor job of modeling the travel of these groups, which are very important to the economy of those areas. Despite these findings, most of the agencies respond- ing to the survey rated their performance as acceptable and their models as adequate to their tasks. The com- mittee even thought that the modeling community has performed reasonably well with the resources at hand. Most MPOs reported that they have too few staff and not enough staff time to consider the development of model improvements. If MPOs do not advocate for model improvements as a priority, who will? Given this situation, the committee considered approaches by which to enhance travel models and the modeling process. There was agreement that better data, better use of existing models, and new model develop- ment are all needed. As a modeling community, we may be guilty of resting on our laurels. There is a need for fed- eral leadership beyond TMIP. The possibility of MPO pooled- funding efforts is another approach. More MPOs and university partnerships represent another approach for advancing the state of the art and state of the practice. MODELING NEEDS Keith Lawton My comments focus on some of the limitations with cur- rent models and possible approaches to address these issues. There are both structural problems and problems of practice with current travel demand models. Exam- ples of structural issues include the use of aggregate trip- based models with a matrix- based approach and the application of static assignment using volume and delay functions. An example of a problem of practice is the limited use of integrated or linked transport and land use models. Trip- based aggregate travel modes do not address the concept of trip sequencing in a tour. This concept is important because it affects location choice and mode choice. The needs of certain activities in a sequence affect earlier or later mode choices. Trip- based models cannot address questions related to the impact of variable pric- ing or dynamic pricing. Aggregate matrix processing lim- its the ability to examine multiple market segmentation, such as effects of household composition, income, and 10 INNOVATIONS IN TRAVEL DEMAND MODELING, VOLUME 1
