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.
based (NHB). These trip purposes were coded based on a lookup table of the 517 possible combinations of pro- duction place, production activity, attraction place, and attraction activity. The data were then coded into a tour format. Three codes were developed to support the most common approaches to tour- based modeling: tour code, with trips in the same tour given a common tour identi- fication number; tour mode, which designated the pri- mary travel mode for each tour; and primary destination, which designated one of the stops on each tour as the primary destination. ⢠A program was developed to group trips into tours. A tour is a sequence of trips starting and ending at home, defining a single round- trip. A subtour is a sequence of trips starting and ending at work defining a single round- trip. To code the tours, the program makes a forward pass through each trip, incrementing the tour identification whenever the traveler departs home. The program tracks when the traveler last departed home and last departed work for each trip. The program then makes a reverse pass through trips, identifying trips where the traveler departed work more recently than he or she departed home. These subtours are noted and the model makes one more forward pass through the trips, incrementing the identification of the subtours and of all subsequent tours. ⢠For each tour, one place is designated as the pri- mary destination. In traditional tour- based models, the primary destination is important because the model structure assumes that the activity at this destination controls the behavior of the tour and that other stops are scheduled around this activity. The model has 16 differ- ent activity types that are ranked according to duration. ⢠The primary mode of each tour is identified by assigning a priority to the mode of each trip in the fol- lowing order: school bus, kiss- and- ride, park- and- ride, walk to transit, drive alone, share ride 2, share ride 3+, bicycle, and walk. As an example, if any trip on a tour is made by a school bus, that is the primary mode for the tour. Not all trips in a tour have to use the same mode. ⢠The results from the household surveys were used in a trip- based format during the refresh phase, which included a partial reestimation and a full recalibration of the DRCOG trip- based model. The results are also being used in the update phase of developing the tour- based travel model. The household survey results show a high level of trip chaining. The data also show differences when coded in trip- based and tour- based formats. For example, 17% of trips are HBW trips, while 33% of the tours are work tours. This difference indicates that work is a key reason for travel even though the number of HBW trips is relatively small due to trip chaining. A com- parison of trip purpose to the primary purpose of the tour for each trip record indicates that only half of all trips on work tours were coded with a HBW purpose, while the other half were HBNW or NHB trips. ⢠There is also a difference in mode share between trips and tours because of the method used to define the primary mode of the tour. Drive alone is a higher prior- ity than shared rides so that any trip on a driving tour is a drive- alone trip, with the primary mode recorded as drive alone. Transit is a high priority in defining the pri- mary tour mode, resulting in a 50% higher transit mode share for tours than for trips. ⢠A number of conclusions can be drawn from the experience with the different surveys in Denver. First, the experience with the home interview survey in the devel- opment of the tour codes suggests that traditional sur- veys are sufficient for tour- based modeling and that advanced activity- based surveys are not necessary to develop reasonable tour codes. Second, the experience in Denver suggests that a robust onboard transit survey is still needed. Using the brief onboard survey to recruit transit riders to participate in the home interview survey provided useful data. The sample size was too small to provide a comprehensive picture of transit use, however. The results from the Denver survey present a reasonable picture of travel behavior and a realistic first step. The results point out the significant degree of trip chaining and highlight the differences in trip purpose and mode. VALIDATION OF THE ATLANTA REGIONAL COMMISSION POPULATION SYNTHESIZER John Bowman and Guy Rousseau John Bowman described the development and the valida- tion of the population synthesizer included as part of the new activity- based travel demand model being developed at the Atlanta Regional Commission (ARC). He described the development of the population synthesizer, the valida- tion process, and the preliminary results of the validation process. Volume 2 includes a paper on this topic.2 The fol- lowing points were covered in his presentation. ⢠ARC is developing a new activity- based travel demand model. It is anticipated that the model will even- tually be used for travel forecasts and policy analysis. The population synthesizer is the first component of the model to be completed. A population synthesizer acts as a con- duit of land use information in a travel demand model. It uses information from the census and the land use model and creates a detailed synthetic population consistent with the land use forecasts. A population synthesizer includes a record for each household in a region and a record for each person in that household. A base- year and back- cast validation was conducted on the ARC population synthe- 21DATA AND SYNTHETIC POPULATIONS 2 See Bowman, J. L., and G. Rousseau. Validation of Atlanta, Georgia, Regional Commission Population Synthesizer. Volume 2, pp. 54â62.
