Article Image

Travel Demand Forecasting Model

Travel analysis incorporates a wide spectrum of topics as part of regional transportation planning activities. In general, travel analysis is performed to assist decision makers in making informed transportation planning decisions. The strength of modern travel demand forecasting is the ability to ask critical “what if” questions about proposed plans and policies.

To do this, we use a travel demand forecasting model - a computer model used to estimate travel behavior and travel demand for a specific future time frame, based on a number of assumptions.

How does a travel demand model work? Traditionally, an approach known as the “four-step process” has been used for regional transportation planning analysis. As its name implies, this process has four basic phases:

  1. Trip generation (the number of trips to be made);
  2. Trip distribution (where those trips go);
  3. Mode choice (how the trips will be divided among the available modes of travel); and
  4. Trip assignment (predicting the route trips will take).

By looking at these four areas, we can answer the following questions:

How many jobs and people?

Looking at land use characteristics (i.e. how we use land in our region) is the main way to figure out trip generation rates. This is because factors like the number and size of households, automobile ownership, types of activities (residential, commercial industrial, etc.), and density of development all drive how much travel flows from or to a specific area within the region.

For simplicity, a geographic unit called a transportation analysis zone is used to create trip generation rates for the region. Specifically, a number of existing or projected characteristics within the TAZ are used for this. The Baltimore region has 1,151 TAZs, each with its own travel characteristics.

Currently, all trips, no matter what mode of travel, are generated; however, only motorized trips (via automobile and transit) are carried through the process to assignment.

Where will jobs and people locate?

Once the model generates a certain number of trips from each TAZ, it needs to determine to which zone each trip goes. This is called trip distribution and the analysis involves a sophisticated process for weighting the “attractiveness” of each TAZ based on the number of attractions it has and the travel time from other TAZs. This step leads to a picture of origin and destination points within the region and how many trips are going between each pair of TAZs – 1151 x 1151 matrix!

How will people travel?

Mode choice shows which mode people use to travel between their origins and destinations. That is, whether people take transit, their own car, or a carpool to and from work or another destination.

A complex sub-model has been created for determining the modal choice, and is based both on certain assumptions about transit capacity, schedules, and fare levels and on real-world observations of how, when, and where people use transit.

What routes will people take?

Trip assignment determines the routes people will take from start (origin) to finish (destination). Generally, the computer assumes everyone will take the quickest route to their destination. To compute route selection requires all kinds of information regarding actual or predicted congestion levels, road conditions, transit schedules and fares, traffic signal systems, etc.

How do we know the model predicts reasonable trips?

Once the four steps are completed, the model provides planners with a picture of existing travel patterns. The results are then given a reality check. Planners consult with local jurisdictions to make sure the numbers make sense, and cross-check how well the model predicts current “observed” data, such as park-and-ride utilization and highway vehicle counts. This “checking” is called a validation.

A validation to the year 2000 has been prepared. Information related to the 2000 Census was used to verify model output. To forecast travel demand in a future year, population, land use, and transportation improvements are incorporated into the process.

What can the model tell us?

The results from the model vary depending on the ideas and information used. For example, whether someone takes the bus for some trips depends on a number of factors, including gas and parking costs, transit fares and travel times. If we assume gas and parking costs skyrocket, forecasts of bus ridership will go up. Assume the opposite and the model will show more people driving. These results can assist decision makers in making informed transportation planning decisions.

What resources are available to update the model?

The latest updates to model development programs are being fostered by the federal government’s Travel Model Improvement Program (external web site). This is a cooperative effort by which travel demand model professionals and users are attempting to bring more accuracy and robustness to regional models.

What committees are responsible for these updates?

The Technical Committee is responsible for overseeing the functions in the work program. 

For more information:
Charles Baber, or 410-732-0500 x1056.

Last Updated on Monday, 27 February 2017 20:06