Center for Activity Systems Analysis

What is the activity-based approach (ABA) and how does it differ from the conventional trip-based model of travel behavior? From where has the activity approach evolved, what is its current status, and what are its potential applications in transportation forecasting and policy analysis. What have been the contributions of activity-based approaches to understanding travel behavior?

The conventional trip-based model of travel demand forecasting (see Chapters 2 and 3) has always lacked a valid representation of underlying travel behavior. This model, commonly referred to as the four-step model (FSM), was developed to evaluate the impact of capital-intensive infrastructure investment projects during a period where rapid increases in transportation supply were arguably accommodating, if not directing, the growth in population and economic activity of the post-war boom. As long as the institutional environment and available resources supported this policy, trip-based models were sufficient to assess the relative performance of transportation alternatives. It was clear from the beginning, however, that the derived nature of the demand for transportation was understood and accepted, yet not reflected in the FSM. The 1970s, however, brought fundamental changes in urban, environmental, and energy policy, and with it the first re-consideration of travel forecasting. It was during this period that the ABA was first studied in depth.

A wealth of behavioral theories, conceptual frameworks, analytical methodologies, and empirical studies of travel behavior emerged during this same period that the policy environment was evolving. These advances shared "a common philosophical perspective, whereby the conventional approach to the study of travel behavior ... is replaced by a richer, more holistic, framework in which travel is analyzed as daily or multi-day patterns of behavior, related to and derived from differences in lifestyles and activity participation among the population" (Jones et al., 1990). This common philosophy has become known as the “activity-based approach”. The motivation of the activity approach is that travel decisions are activity based, and that any understanding of travel behavior is secondary to a fundamental understanding of activity behavior. The activity approach explicitly recognizes and addresses the inability of trip-based models to reflect underlying behavior and, therefore, the inability of such models to be responsive to evolving policies oriented toward management versus expansion of transportation infrastructure and services.

In the next section, a summary and critique of the convention trip-based approach is presented, followed by an overview of ABAs, focusing on how these approaches address the various limitations of the conventional model. This is followed by a review of representative examples of activity-based approaches, including several perhaps best considered as contributions to understanding travel behavior, and several oriented toward direct application in forecasting and policy analysis. Some summary comments are then provided including an assessment of the future of both trip-based and activity-based approaches.

The history of demand modeling for person travel has been dominated by the modeling approach that has come to be referred to as the four step model (FSM) (see Chapter 2). Travel, always viewed in theory as derived from the demand for activity participation, in practice has been modeled with trip-based rather than activity-based methods (as presented in Chapter 4). Trip origin-destination (O-D) rather than activity surveys form the principle database. The influence of activity characteristics decreases, and that of trip characteristics increases, as the conventional forecasting sequence proceeds. The application of this modeling approach is near universal, as in large measure are its criticisms (these inadequacies are well documented, e.g., by McNally and Recker (1986)). The current FSM might best be viewed in two stages. In the first stage, various characteristics of the traveler and the land use - activity system (and to a varying degree, the transportation system) are "evaluated, calibrated, and validated" to produce a non-equilibrated measure of travel demand (or trip tables). In the second stage, this demand is loaded onto the transportation network in a process than amounts to formal equilibration of route choice only, not of other choice dimensions such as destination, mode, time-of-day, or whether to travel at all (feedback to prior stages has often been introduced, but not in a consistent and convergent manner). Although this approach has been moderately successful in the aggregate, it has failed to perform in most relevant policy tests, whether on the demand or supply side.

This chapter extends the material in Chapter 2 by providing a concise overview of the mechanics of the FSM, illustrated with a hypothetical case study. The discussion in this chapter, however, will focus on U.S. modeling practice. Transportation modeling developed as a component of the process of transportation analysis that came to be established in the U.S.A. during the era of post-war development and economic growth. Initial application of analytical methods began in the 1950s. The landmark study of Mitchell and Rapkin (1954) not only established the link of travel and activities (or land use) but called for a comprehensive framework and inquiries into travel behavior. The initial development of models of trip generation, distribution, and diversion in the early 1950s lead to the first comprehensive application of the four-step model system in the Chicago Area Transportation Study (see Weiner, 1997) with the model sandwiched by land use projection and economic evaluation. The focus was decidedly highway-oriented with new facilities being evaluated versus traffic engineering improvements. The 1960s brought federal legislation requiring "continuous, comprehensive, and cooperative" urban transportation planning, fully institutionalizing the FSM. Further legislation in the 1970s brought environmental concerns to planning and modeling, as well as the need for multimodal planning. It was recognized that the existing model system may not be appropriate for application to emerging policy concerns and, in what might be referred to as the "first travel model improvement program", a call for improved models led to research and the development of disaggregate travel demand forecasting and equilibrium assignment methods that integrated well with the FSM and have greatly directed modeling approaches for most of the last 30 years. The late 1970s brought "quick response" approaches to travel forecasting (Sosslau et al., 1978; Martin and McGuckin, 1998) and independently the start of what has grown to become the activity-based approach (see Chapter 4). The growing recognition of the misfit of the FSM and relevant policy questions in the 1980s led to the (second, but formal) Travel Model Improvement Program in 1991; much of the subsequent period has been directed at improving the state-of-thepractice relative to the conventional model while fostering research and development in new methodologies to further the state-of-the-art (see Chapter 4).

The FSM is best seen as a particular application of transportation systems analysis (TSA), a framework due to Manheim (1979) and Florian et al. (1988), which positions the model well to view its strengths and weaknesses. A brief presentation of this TSA framework introduces the FSM context and leads to a discussion of problem and study area definition, model application, and data requirements. The models that are perhaps most commonly utilized in the FSM are then presented in the form of a sample application.

This paper examines the nature of land use based substitution effects on travel modes, identified by Greenwald, examining the direct impact of land uses inducing trip-making behaviors. These impacts are analyzed in the context of trip chaining, defined here as consolidating two or more non-home activities in a single departure from home. The findings suggest rather than strictly promoting one type of transportation over another, the regional impact of localized urban design practices is to consolidate trip making behavior closer to the home. As such, urban design “carrots” must be complemented with policy “sticks” in order to promote true exchanges of travel modes.

With increase in traffic volume and change in travel related characteristics, vehicular emissions and energy consumption have increased significantly since two decades in India. Current models are not capable of estimating vehicular emissions accurately due to inadequate representation of real-world driving. The focus of this paper is to understand the level of Indian Driving cycle (IDC) in representing the real-world driving and to assess the impact of real-world driving on vehicular emissions.

The study has revealed that IDC does not represent the real-world driving. Irrespective of road classes, about 30% of time is spent below 20 km/hr and the speed too exceeds IDC’s maximum limit of 42 km/hr. Emissions are estimated for different driving patterns using International Vehicle Emission (IVE) model. Emission rates vary significantly from one class of road to another and the largest effect is on local streets.

Increase in traffic volumes and changes in travel-related characteristics increase vehicular emissions significantly. It is difficult, however, to accurately estimate emissions with current practice because of the reliance on travel forecasting models that are based on steady state hourly averages and, thus, are incapable of capturing the effects of traffic variations in the transportation network. This paper proposes an intermediate model component that can provide better estimates of link speeds by considering a set of Emission Specific Characteristics (ESC) for each link. The intermediate model is developed using multiple linear regression; it is then calibrated, validated, and evaluated using a microscopic traffic simulation model. The improved link speed data can then be used to provide better estimates of emissions. The evaluation results show that the proposed emission estimation method performs better than current practice and is capable of estimating time-dependent emissions if traffic sensor data are available as model input.

The dynamic process of how individuals organize their activities and travel is often termed activity scheduling. Investigation of the dynamic processes has been the interest of transportation researchers in the past decade, because of its relevance to the effectiveness of congestion management and intelligent transportation systems. To empirically examine this process, a computerized survey instrument was developed to collect household activity scheduling data. The instrument is unique in that it records the evolution of activity schedules from intentions to final outcomes for a weekly period.

This paper summarizes the investigation on the dynamic processes of activity scheduling and trip chaining based on data collected from a pilot study of the instrument. With the data, ordered logit models are applied to identify factors that are related to the scheduling horizon of activities. Results of the empirical analyses show that activities of shorter duration were more likely to be opportunistically inserted in a schedule already anchored by their longer duration counterparts. Additionally, analysis of travel patterns reveals that many trip-chains were formed opportunistically. Travel time required to reach an activity was positively related to the scheduling horizon for the activity , with more distant stops being planned earlier than closer locations.

This study uses two-day travel diary data to examine whether land use matters more for an individual's total vehicle miles traveled (VMT). More specifically, sociodemographic, land use, and street connectivity variables are used to estimate nonwork trip frequency and nonwork vehicle miles traveled via ordered probit and ordinary least-squares regression models. We compare standardized coefficients of the models and conclude that: (1) the influence of land use variables is similar in both the trip generation and VMT regressions; and (2) income is the primary determinant of both trip frequency and VMT, but that land use exerts an influence that is on par with other sociodemographic characteristics after the primary role of income is considered.

The fundamental principle of intelligent transportation systems is to match the complexity of travel demands with advanced supply-side analysis, evaluation, management, and control strategies. A fundamental limitation is the lack of basic knowledge of travel demands at the network level. Modeling and sensor technology is primarily limited to aggregrate parameters or micro-simluations based on aggregate distributions of behavior. Global Positioning Systems (GPS) are one of several available technologies which allow individual vehicle trajectories to be recorded and analyzed. Potential applications of GPS which are relevant to the ATMS Testbed are implemented in probe vehicles to deliver real-time performance data to complement loop and other sensor data and implementation in vehicles from sampled households to record route choice behavior. An Extensible GPS-based in-vehicle Data Collection Unit (EDCU)has been designed, tested, and applied in selected field tests. Each unit incorporates GPS, data logging capabilities, two-way wireless communications, and a user interface in an extensible system which eliminates driver interaction. Together with supporting software, this system is referred to as TRACER. The design and initial implementation tests Testbed are presented herein. This research is a continuation of PATH MOU 3006; selected portions of the interim report for that MOU are repeated here to provide a complete overview of the research effort

The growth of the Internet has rekindled interest in the relationship between communications and travel. New communication technologies have expanded the range, the type, and the number of transactions that can take place without travel. A number of promotions capture the new tradeoffs between communications and travel: initially, the Internet was referred to as “the information superhighway” and Microsoft ran an ad campaign dubbed “where do you want to go today?” The connection between travel and bytes has been summed up as “The Death of Distance” (Cairncross, 1997). A parallel evolution in telecommunication and transportation was envisioned more than 150 years ago with the inventions of the telegraph and telephone. The telephone was expected to “speed the movement of perishable goods,” “reduce the travels of salesmen,” and “let (itinerant) workers stay at home to be phoned for jobs” (Pool, 1983). Today, the Internet has fueled similar expectations, and many of them center on travelrelated issues. The Internet might relieve demand for new road capacity, slow down the rate of new vehicle ownership, and divert existing travel trips to less congested times. The Internet might help create more sustainable growth in transportation, by providing virtual accessibility. In this paper, we explore the transportation aspects of consumer electronic commerce (e-commerce). Shopping activities are currently automobileintensive in many countries, and increases in e-commerce could portend important changes in transportation patterns and activities.

This two-year project focused on obtaining travel behavior data that more truly reflects underlying behavior. In the first year of the project a prototype of REACT!, a web-based, self-administered survey instrument for collecting household travel/activity data was produced.REACT! documents not only the resultant behavior but also the scheduling process that produces that behavior by having each respondent record activities as they are initially planned, updated, and executed. In the second year, following a beta test of REACT! and final program modification, a formal REACT! field study was completed for 47 households who used REACT! to provide 24 hours of travel/activity data over a 7 day period. Ensuing analyses focused on the activity scheduling process.