At the start of a full travel demand model run, the model estimates household auto-ownership levels, then trip generation rates for land-use data to calculate trip ends at the Transportation Analysis Zone (TAZ) level. These trip ends are paired into origins and destinations in the distribution model. In the mode-split model, a mode of travel is selected for each trip. Vehicle trips are assigned to the highway network in the assignment model.

The travel time feedback loop in the model is accomplished prior to mode choice by converting person trips to vehicle trips based on observed data.

The model is comprised of several steps with each step programmed or scripted separately. These steps include, but are not limited to the following:

Auto Ownership

Is a function of the characteristics of a household and where it is located. Auto ownership and availability is a strong predictor of trip making and mode choice behavior.

Trip Generation

Calculates the number of person trips generated within each TAZ. The trip generation model parameters are developed from travel surveys collected in 1993 and 2001. The number of trips to and from a place is a function of the amount and types of land-use activity within the zone.

Trip Distribution

Pairs the origins and destinations for each zone for each of the trip purposes. Trip generation estimates the number of trips to and from each TAZ. Trip distribution completes the trip by describing which trip origins are linked with which trip destinations. The result of this is a person-trip matrix for each trip type. Trip distribution links trip-ends of the same type based primarily on the spatial separation of different land uses and observed sensitivities to trip length. One output of trip distribution is the person-trip table for home to work that can be compared to the “Journey-to-Work” data provided by the Bureau of the Census.

Vehicle Assignment

Locates the “best” routes between each origin/destination pair and assigns vehicle trips to the highway network. Important outputs of this module include number of vehicles on each roadway segment by time period. Several other pieces of data can be extracted, including operating speeds, travel times, Vehicle Miles Traveled (VMT), Vehicle Hours Traveled (VHT), and Volume over Capacity (V/C) on roadway links. In addition, one can configure the vehicle assignment to save all the vehicle trips that use a single link in either direction (select link analysis) or all the vehicle trips that originate or are destined for a zone (select zone analysis).

Travel Time Feedback

Finds the best available travel path via each of the travel modes explicitly modeled. Several modes are explicitly modeled, including auto, transit modes (local bus, bus rapid transit, light rail, commuter rail), and non-motorized modes. Skims are reasonable approximations of the travel time and cost between all pairs of TAZs, and skims are described for each travel mode. The path-finding algorithms are calibrated based on observed travel paths and observed relationships between volumes and congested speeds.

Mode Choice

Calculates which mode each person trip is likely to take based on availability and mode-specific parameters (e.g. time, cost, transit frequency). Mode split provides a breakdown of person trips by mode, both for captive riders (people without automobiles) and for the total population. The mode split model is developed based on observed data on mode preferences and what those preferences imply about sensitivities to mode attributes.

Final Assignment

Uses the trip table from mode split and assigns the person trips using transit to the appropriate transit route. It also gives the final number of vehicles on each roadway segment by time period. This provides a means of viewing roadway volumes and transit ridership graphically and understanding the relative effectiveness of different segments of the road and transit networks.

Model Output

Is summarized automatically by the model, including regional statistics (e.g. VMT, VHT, transit shares and trip lengths), corridor and segment performance statistics (e.g. delay, volume, and ridership), district and county-level trip flow, MOVES emissions model inputs (EPA air quality model), and calibration statistics.