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.
high- occupancy vehicle; 3, walk to transit; 4, drive to transit; 5, nonmotorized; or 6, school bus), and which one of five transit submodes is chosen for each half tour for walk- to- transit and drive- to- transit tours [1, local bus; 2, express bus; 3, bus rapid transit; 4, light rail tran- sit (LRT); or 5, commuter rail]; and ⢠Stop- frequency model that defines whether there is an intermediate stop at each half- tour. Because only one potential stop on each half tour is considered, the model at the tour level has only four explicitly modeled alter- natives: 1, no stops; 2, outbound stop; 3, inbound stop; and 4, stops on both half tours. Two subsequent models relate to the following trip- level choices, which are conditional upon the previously made tour- level decisions: ⢠Stop- location model that defines a location for each stop at the same level of spatial resolution as pri- mary destination (1,805 zones and three transit- access subzones for each zone). Stop location availability is strongly conditional on availability of the chosen tour mode and transit sub- mode to access the location; and ⢠Trip- mode model that defines mode and transit submode for each trip on the tour. If there is no stop on a half tour, the entire half tour is considered one trip and the chosen mode and transit submode are preserved. If there is a stop, the half tour is broken into two successive trips (to and from the stop). After processing through all tour- level and trip- level stages, trip tables are constructed for all modes and tran- sit submodes. These tables are assigned to the corre- sponding highway and transit subnetworks. Loaded networks are skimmed to produce level- of- service attri - butes necessary for the models. The model system is designed to process through several global iterations, including all (or a chosen subset of) models and network assignments until an equilibrium is reached. Furthermore, several important upward linkages of the choice models through log sums from the lower- level choices used in upper- level choices are incorporated: ⢠Entire- tour bidirectional mode choice log sums for the representative time- of- day periods (for example, a.m.âp.m. combination for work tours and a.m.âmidday combination for school tours) are used as variables in the primary tour destination choice models; the reason that only representative mode choice log sums are used in the destination choice is that this choice dimension has 1,805 3 5,415 alternatives and is extremely compu- tationally intensive. ⢠Entire- tour bidirectional mode choice log sums for all time- of- day periods are used as variables in time- of- day choice; because the time- of- day choice model is applied conditionally upon the chosen destination, it is signifi- cantly less intensive computationally than destination choice, and it is possible to explicitly consider mode choice log sums for all possible time- of- day combinations. 34 INNOVATIONS IN TRAVEL DEMAND MODELING, VOLUME 2 Daily Activity Pattern / Tour Production Primary Tour Destination Time-of-Day by Half-Tours Entire-Tour Mode / Best Transit Sub-mode Stop Frequency by Half-Tours Stop Location Trip Mode Traffic & Transit Assignment Zonal Accessibility Log-sums for all TOD periods Log-sums for representative TOD periods Log-sums Highway & best transit skims Access by best transit sub-mode Transit sub-mode skims FIGURE 1 General structure of the MORPC model system.