Once the resident, external worker, and overnight, day, and thru-visitor models have run, then all of the person tours are transformed into zone-to-zone trip tables that are assigned to the Tahoe highway network. The assignment, performed for each skim period, is done using TransCAD transportation software. The skim periods are defined in the table below.
Skim Period | Start Time | End Time | Duration |
---|---|---|---|
AM Peak (AM) | 7:00 AM | 10:00 AM | 3 hours |
Midday (MD) | 10:00 AM | 4:00 PM | 6 hours |
PM Peak (PM) | 4:00 PM | 7:00 PM | 3 hours |
Late Night (LN) | 7:00 PM | 7:00 AM | 12 hours |
After all of the tour models have finished, trips must be synthesized from the individual tours. Each leg of a particular tour counts as one trip from the origin to the destination, unless there was a stop on that leg, in which case there is a trip from the origin to the stop, and a trip from the stop to the destination. The trip information is sorted according to skim period and mode and is stored in separate TransCAD trip tables. When performing the traffic assignment, only the drive-alone and shared auto modes are assigned. The drive-to-transit, walk-to-transit, transit and the non-motorized trips are not assigned to the network because there are so few of them that their impact on traffic totals are negligible.
The street network is built from a TransCAD line and node layer. The node layer is a representation of the street intersections and TAZ nodes, while the line layer is a representation of the streets. The streets are broken into the following functional classifications:
Principle Arterials
Minor Arterials
Collectors
Centroids
For each of these functional classifications, the following capacity per lane per hour values are assigned:
Functional Class | Capacity per Lane per Hour |
---|---|
Principle Arterial | 1100 |
Minor Arterial | 800 |
Collector | 500 |
Centroid | 9999 |
In addition to this, the number of lanes and the speed limit of each link are recorded in the network. From the latter, a free flow travel time is determined for each link by taking the time it takes to traverse the link at its speed limit times a “travel-time multiplier.” The travel-time multiplier is initially set to one, but was changed for some links during validation to help match observed counts. Also associated with each link is a volume-delay function (VDF), which is used in the actual traffic assignment procedure. The VDF provides a travel-time for each link given its free-flow travel time, capacity, and assigned volume.
A different value for \(\alpha\) and \(\beta\) are used for each link depending on its speed limit, number of lanes, and urban-type location (urban type definitions follows). These values are summarized in the following table:
Area Type | Speed Limit | Lanes | \(\alpha\) | \(\beta\) |
---|---|---|---|---|
Rural | 60 | 2+ | 0.09 | 6 |
Rural | 55 | 2+ | 0.08 | 6 |
Rural | 50 | 2+ | 0.07 | 6 |
Rural | 45 | 2+ | 0.07 | 6 |
Rural | 35 | 2+ | 0.92 | 5 |
Rural | 25 | 2+ | 1.10 | 5 |
Rural | Any | <2 | 0.34 | 4 |
Suburban | 55 | Any | 0.33 | 4 |
Suburban | 50 | Any | 0.34 | 4 |
Suburban | 45 | Any | 0.42 | 5 |
Suburban | 40 | Any | 0.38 | 5 |
Suburban | 35 | Any | 0.96 | 5 |
Suburban | 30 | Any | 1.11 | 5 |
Suburban | 25 | Any | 1.20 | 5 |
Suburban | 20 | Any | 1.25 | 5 |
Suburban | 15 | Any | 1.30 | 5 |
Urban | 50 | Any | 0.74 | 5 |
Urban | 45 | Any | 0.72 | 5 |
Urban | 40 | Any | 0.70 | 5 |
Urban | 35 | Any | 1.00 | 5 |
Urban | 30 | Any | 1.20 | 5 |
Urban | 25 | Any | 1.30 | 5 |
In order to quantify geographic differences between the various zones, an urban type model was developed. This model maps each zone into one of four urban type categories:
The latter category is assigned if a zone is an external zone. The other three categories are designated by using the following model structure:
WP is the sum of the employment and population densities, where the area is measured in square miles.
A traffic assignment is performed for each skim period, using the trip tables generated during trip synthesis. The assignment type is capacity restraint, meaning that all of the trips between two zones are iteratively assigned to the shortest calculated path by time (all-or-nothing assignment). For each iteration, link volumes are updated by the following (MSA) method:
From these link volumes, new link travel times are computed, and the traffic is re-assigned. This procedure is continued until the average change in link volumes becomes very small. This assignment method is guaranteed to converge. During model development, a convergence criterion of 0.0001 was used, and it was found that a maximum of 50 iterations was needed to reach convergence.
Once trips have been assigned to the street network, skims can be produced. Skims are matrices of values giving travel time and cost components between zones for a set of modes. For the Tahoe model, three main sets of skims are produced: highway (street) skims, transit skims, and walk skims. These skims are used throughout the model for such things as calculating accessibilities and destination, time-of-day, and mode choice determinations. The walk skims are simply a zone-to-zone shortest distance matrix using 3 miles/hour as the walk speed. These do not require the traffic assignment results. For highway skims, the shortest path by time as determined during the traffic assignment results is stored.
Transit skims are somewhat more complicated. For full details, see Travel Demand Modeling with TransCAD 4.8. Briefly, the transit skimming procedure determines the shortest path between two zones using transit modes. There is a separate skim for walk-to-transit and drive-to-transit modes. For the former, the access to transit is by walking from the origin zone to the (usually closest) transit stop; for the latter, the access to transit is determined by driving to a qualifying transit stop. A qualifying stop is often a park and ride lot but for the Tahoe basin, because of the transit and street layout, all transit stops were considered qualified. In the transit skim, transit transfers are allowed, and access from the final transit station to the destination zone is always made by walking.
For each transit skim, a number of matrices are produced:
Transit access time (time from origin to transit stop)
Transit in-vehicle time (time in transit vehicle)
Transit egress time (time from final transit stop to destination)
Transit headway wait times
Transit transfer times
Transit fare costs
The fare matrix can be translated into time units using an appropriate value of time (which varies across model segments).
Systems Analysis Group, WSP USA 2018