Topics

• The Big Picture
  • Types of Models
    • Sketch Planning
    • Quick Response
    • Three Step
    • Four Step
    • Activity Based
    • Freight
    • Land Use
    • Air Quality
    • Integrated
  • Decisions on Approach
  • Transition Paths
  • End Users
• Model Input
• Model Development
• Model Application

Types of Models

Travel demand modeling was first developed in the late 1950's as a means to do highway planning. As the need to look at other problems and issues arose, the modeling process has been modified to add additional techniques to deal with these problems. The objective of travel demand forecasting is to predict changes in travel behavior and transportation conditions as a result of proposed transportation projects, policies, and future changes in socioeconomic and land use patterns.

Models are simulations of the "real world" that can be used to show the impact of changes in a particular area on the transportation system (such as adding a new road or transit line, or increases in population or employment). Travel demand modeling was first developed in the late 1950’s as a means to do highway planning.[1]  As the type and extent of transportation related problems and issues multiply, additional travel modeling techniques have been and continue to be developed. Travel models may be used to test the travel impacts of changes in land use, economic development, fuel and parking cost, and new highway or transit system capacity.

A variety of forecasting methods has been developed to predict changes in travel behavior. Forecasting methods are generally founded on theoretical models and then verified by empirical studies, which describe how people change their behavior in response to changes in the major factors which influence this behavior.

Travel behavior may be studied or modeled from two perspectives:

• The aggregate perspective. Aggregate studies look at travel from an areawide perspective. They attempt to relate characteristics of an area (e.g., population, employment, or average income) to travel characteristics of that area (e.g., average number of trips per household, or the number or percent of trips made by foot or bicycle). In the context of non-motorized travel, these studies may also look at characteristics of specific facilities (e.g., roadway and sidewalk width or type) in conjunction with characteristics of the surrounding area (e.g., population density, or number of students) to predict the number of people using the facility.
• The disaggregate or individual perspective. Disaggregate studies look at travel decisions from the perspective of the individual. The individual's personal characteristics (e.g., age, gender, attitudes, beliefs) interact with the travel options available to them (e.g., time, cost, comfort of competing modes). To predict overall demand, models of individual behavior are applied across a population with known characteristics.

Each approach has its advantages and disadvantages. Aggregate-level methods tend to be relatively easy to apply, with readily available data sources and computational methods, and can be useful for sketch-planning purposes. Disaggregate-level methods are more complicated to develop but can be much more effective at predicting behavior changes. This is because they explain individual choices rather than making generalizations based on overall population characteristics. [2]

Three important ingredients are part of any model used for transportation analysis:

• Key base, or current-year characteristics of travelers and the transportation system, described in terms of quantifiable variables (e.g., the number of highway travel lanes, transit service headways, household size and income, number of vehicles per household, employment patterns by type and job classification, etc.).
• The relationship between these variables and the travel behavior of individuals (e.g., the more automobiles per household, the greater the number of automobile trips per household). This relationship is most often expressed in mathematical terms.
• Future-year forecasts of key traveler and transportation system characteristics. This relationship is the same for all individuals and is constant over time.

Challenges to the validity of travel models often focus on one of these three assumptions.[3]

There are a wide variety of models available to allow transportation planners to respond to the extensive list of issues and decisions surrounding transportation today. They vary in:

• size (number of persons, jobs, links, nodes, zones, square miles, etc.)
• complexity (traditional 4 step trip based, activity based, micro, meso and macro scale, integrated, etc.)
• scope (intersection, city, region, state, multi-state,etc.)
• run-time (minutes, hours, days)
• computing requirements (desk top, multi-tiered processors, megabytes  versus gigabytes of storage)
• observed data requirements (for calibration, validation of models; from household surveys, workplace surveys, census reports, traffic counts, transit patronage studies, travel time studies, speed studies, GPS tracking devices for trip patterns, economic studies, demographic estimates, parking cost revenues, truck and freight studies, etc.)
• purpose (traffic signal analysis, pricing/tollroad studies, fixed guideway transit analysis, new roadway capacity planning, local thoroughfare planning, freight corridor analysis, land use/transport impact evaluation, air quality evaluation, etc.)

The list continues to grow as practitioners and researchers develop new ways to respond to the ever growing range of issues decision makers face.

Comments