A model of complex travel behavior: Part I—Theoretical development

This paper presents a policy sensitive approach to modeling travel behavior based on activity pattern analysis. A theoretical model of complex travel behavior is formulated on a recognition of a wide range of interdependencies associated with an individual's travel decisions in a constrained environment. Travel is viewed as input to a more basic process involving activity decisions. A fundamental tenet of this approach is that travel decisions are driven by the collection of activities that form an agenda for participation; the utility of any specific travel decision can be determined only within the context of the entire agenda. Based on the theoretical model of complex travel behavior, an operational system of models, STARCHILD (Simulation of Travel/Activity Responses to Complex Household Interactive Logistic Decisions), has been developed to examine the formation of household travel/activity patterns, and is presented in a companion paper (Recker et al., 1986).     

A model of complex travel behavior: Part II—An operational model

Based on the theoretical model of complex travel behavior developed in a companion paper (Recker et al., 1986), an operational system of models, STARCHILD (Simulation of Travel/Activity Responses to Complex Household Interactive Logistic Decisions), has been developed to examine the formation of household travel/activity patterns. The system employs a simulation approach in combination with techniques of pattern recognition, multiobjective optimization and disaggregate choice models. Initial empirical verification of the system of models is presented based on results obtained from a sample data set.     

Where do cyclists ride? A route choice model developed with revealed preference GPS data

To better understand bicyclists’ preferences for facility types, GPS units were used to observe the behavior of 164 cyclists in Portland, Oregon, USA for several days each. Trip purpose and several other trip-level variables recorded by the cyclists, and the resulting trips were coded to a highly detailed bicycle network. The authors used the 1449 non-exercise, utilitarian trips to estimate a bicycle route choice model. The model used a choice set generation algorithm based on multiple permutations of path attributes and was formulated to account for overlapping route alternatives. The findings suggest that cyclists are sensitive to the effects of distance, turn frequency, slope, intersection control (e.g. presence or absence of traffic signals), and traffic volumes. In addition, cyclists appear to place relatively high value on off-street bike paths, enhanced neighborhood bikeways with traffic calming features (aka “bicycle boulevards”), and bridge facilities. Bike lanes more or less exactly offset the negative effects of adjacent traffic, but were no more or less attractive than a basic low traffic volume street. Finally, route preferences differ between commute and other utilitarian trips; cyclists were more sensitive to distance and less sensitive to other infrastructure characteristics for commute trips.     

Where Is My Bus? Impact of mobile real-time information on the perceived and actual wait time of transit riders

In order to attract more choice riders, transit service must not only have a high level of service in terms of frequency and travel time but also must be reliable. Although transit agencies continuously work to improve on-time performance, such efforts often come at a substantial cost. One inexpensive way to combat the perception of unreliability from the user perspective is real-time transit information. The OneBusAway transit traveler information system provides real-time next bus countdown information for riders of King County Metro via website, telephone, text-messaging, and smart phone applications. Although previous studies have looked at traveler response to real-time information, few have addressed real-time information via devices other than public display signs. For this study, researchers observed riders arriving at Seattle-area bus stops to measure their wait time while asking a series of questions, including how long they perceived that they had waited.The study found that for riders without real-time information, perceived wait time is greater than measured wait time. However, riders using real-time information do not perceive their wait time to be longer than their measured wait time. This is substantiated by the typical wait times that riders report. Real-time information users say that their average wait time is 7.5 min versus 9.9 min for those using traditional arrival information, a difference of about 30%. A model to predict the perceived wait time of bus riders was developed, with significant variables that include the measured wait time, an indicator variable for real-time information, an indicator variable for PM peak period, the bus frequency in buses per hour, and a self-reported typical aggravation level. The addition of real-time information decreases the perceived wait time by 0.7 min (about 13%).A critical finding of the study is that mobile real-time information reduces not only the perceived wait time, but also the actual wait time experienced by customers. Real-time information users in the study wait almost 2 min less than those arriving using traditional schedule information. Mobile real-time information has the ability to improve the experience of transit riders by making the information available to them before they reach the stop.     

How do built-environment factors affect travel behavior? A spatial analysis at different geographic scales

Much of the literature shows that a compact city with well-mixed land use tends to produce lower vehicle miles traveled (VMT), and consequently lower energy consumption and less emissions. However, a significant portion of the literature indicates that the built environment only generates some minor—if any—influence on travel behavior. Through the literature review, we identify four major methodological problems that may have resulted in these conflicting conclusions: self-selection, spatial autocorrelation, inter-trip dependency, and geographic scale. Various approaches have been developed to resolve each of these issues separately, but few efforts have been made to reexamine the built environment-travel behavior relationship by considering these methodological issues simultaneously. The objective of this paper is twofold: (1) to better understand the existing methodological gaps, and (2) to reexamine the effects of built-environment factors on transportation by employing a framework that incorporates recently developed methodological approaches. Using the Seattle metropolitan region as our study area, the 2006 Household Activity Survey and the 2005 parcel and building data are used in our analysis. The research employs Bayesian hierarchical models with built-environment factors measured at different geographic scales. Spatial random effects based on a conditional autoregressive specification are incorporated in the hierarchical model framework to account for spatial contiguity among Traffic Analysis Zones. Our findings indicate that land use factors have highly significant effects on VMT even after controlling for travel attitude and spatial autocorrelation. In addition, our analyses suggest that some of these effects may translate into different empirical results depending on geographic scales and tour types.     

How do people respond to congestion mitigation policies? A multivariate probit model of the individual consideration of three travel-related strategy bundles

This study explores the relationships between adoption and consideration of three travel-related strategy bundles (travel maintaining/increasing, travel reducing, and major location/lifestyle change), linking them to a variety of explanatory variables. The data for this study are the responses to a fourteen-page survey returned by nearly 1,300 commuting workers living in three distinct San Francisco Bay area neighborhoods in May 1998. We first identified patterns of adoption and consideration among the bundles, using pairwise correlation tests. The test results indicate that those who have adopted coping strategies continue to seek for improvements across the spectrum of generalized cost, but perhaps most often repeating the consideration of a previously-adopted bundle. Furthermore, we developed a multivariate probit model for individuals’ simultaneous consideration of the three bundles. It is found that in addition to the previous adoption of the bundles, qualitative and quantitative Mobility-related variables, Travel Attitudes, Personality, Lifestyle, Travel Liking, and Sociodemographics significantly affect individual consideration of the strategy bundles. Overall, the results of this study give policy makers and planners insight into understanding the dynamic nature of individuals’ responses to travel-related strategies, as well as differences between the responses to congestion that are assumed by policy makers and those that are actually adopted by individuals.

How do local actions affect VMT? A critical review of the empirical evidence

In this paper, we present a discussion of the challenges for research on the topic of vehicle miles traveled. We then summarize and critique evidence from the US on the association between 14 distinct factors and vehicle miles traveled. Our results quantify how much vehicle miles traveled can be expected to change in response to changes in policy or land use factors, including residential density and land use mix, as well as specific transport policies and programs such as transit improvements, road pricing, and programs aimed at changing people’s travel choices. Overall, though individual studies differ as to exact effect sizes, it is clear that local-level policymakers can take actions that are likely to affect vehicle miles traveled. However, we highlight gaps in the knowledge base at a time when decision makers at the local level are being increasingly called upon to take action to reduce vehicle miles traveled. Variation in effect size based on local context or interaction with related policies and programs has been left largely unexplored. In addition, experimental research designs that can identify causal direction are rare, and appropriate data that quantifies vehicle miles traveled are often lacking.     

A partial differential equation formulation of Vickrey’s bottleneck model,part I: Methodology and theoretical analysis

This paper is concerned with the continuous-time Vickrey model, which was first introduced in Vickrey (1969). This model can be described by an ordinary differential equation (ODE) with a right-hand side which is discontinuous in the unknown variable. Such a formulation induces difficulties with both theoretical analysis and numerical computation. Moreover it is widely suspected that an explicit solution to this ODE does not exist. In this paper, we advance the knowledge and understanding of the continuous-time Vickrey model by reformulating it as a partial differential equation (PDE) and by applying a variational method to obtain an explicit solution representation. Such an explicit solution is then shown to be the strong solution to the ODE in full mathematical rigor. Our methodology also leads to the notion of generalized Vickrey model (GVM), which allows the flow to be a distribution, instead of an integrable function. As explained by Han et al. (in press), this feature of traffic modeling is desirable in the context of analytical dynamic traffic assignment (DTA). The proposed PDE formulation provides new insights into the physics of The Vickrey model, which leads to a number of modeling extensions as well as connection with first-order traffic models such as the Lighthill–Whitham–Richards (LWR) model. The explicit solution representation also leads to a new computational method, which will be discussed in an accompanying paper, Han et al. (in press).     

Why do you care what other people think? A qualitative investigation of social influence and telecommuting

The effect of social interactions on decision-making is a topic of current interest in the travel behavior literature. These interactions have been investigated primarily from an intra-household perspective, but increasingly too in other types of social settings. In the case of interactions within a workplace, it has been suggested that the decision to telecommute may have some important social components. Previous research has concentrated on social isolation, and the effect on job satisfaction of qualitatively different (i.e., telecommunications-mediated) relationships with managers and colleagues. A topic that remains unexplored is the way social norms, in effect the influence of other people’s behavior, may influence the decision to adopt telecommuting. In this paper we set to investigate, within a qualitative framework, the role of social contact in the process of acquiring information on, and making decisions about, telecommuting. The results indicate that social contact does play a subtle but non-trivial role in the adoption and continuation process, and offer some insights about the importance of the social dimension, institutional set-up, and how they interact to influence the decision to telecommute.     

How do passengers use travel time? A case study of Shanghai–Nanjing high speed rail

Traditional travel behavior theory regards travel time as a waste. Recent studies suggest that it carries a positive utility, among other reasons for the benefit of the activities conducted while traveling. However, most studies of travel time use have focused on conventional trains in developed countries. Few have systematically examined the permeation of information and communication technology (ICT) into travel time use and the correlates of activity participation in developing countries, particularly on high speed rail (HSR). Using a survey conducted on the Shanghai–Nanjing corridor (N = 901), this study examines how HSR passengers use their travel time and explores the correlates of the different types of activities of business and non-business travelers, respectively, through multivariate probit models. We found that 96% of the respondents use ICT during their HSR journey and that most passengers spend some of their travel time on work-related activities. Moreover, items carried and advance planning as well as work-related travel attributes contribute significantly to activity participation. However, the factors affecting time use of business and non-business travelers differ. HSR service design should facilitate passenger engagement in various activities and improvement of their travel experience. A stable internet connection, adequate power sockets, and a noise-free environment will promote both work and leisure activities on the HSR.     

Can rail/sail compete with air travel to Cyprus? A comparison of emissions

Rail and sea voyage journeys to Cyprus from a variety of origins are constructed to derive the travel emissions and travel time per person to compare popular aviation routes. The hypothetical ‘slow travel’ routes are approximately eight to ten times longer than flying. Emissions are lower from certain origins by about 100 kg CO₂ per person per round trip under reasonably high occupancy conditions when compared to current direct air services. Emissions from the sea voyages are derived from a sample of 162 marine vessels using the energy efficiency design index for European ships running at 20 knots.     

What will autonomous trucking do to U.S. trade flows? Application of the random-utility-based multi-regional input–output model

Transportation - This study anticipates changes in U.S. highway and rail trade patterns following widespread availability of self-driving or autonomous trucks (Atrucks). It uses a...     

A week in the life of baby boomers: how do they see the spatial–temporal organization of their activities and travel?

In this paper, we take an initial look at the spatial and temporal flexibility in the activity patterns of the so-called “baby-boomer” cohort (born 1947–1966) in comparison with younger and older adults. Using a unique longitudinal survey carried in Quebec City from 2002 to 2005, we explore activity patterns and trip rates over a seven-day observation period during the first wave, and take a first look at some aspects of their evolution over two subsequent waves at about one-year intervals. We model the propensity to undertake activities within selected conventional non-work classifications such as “shopping” and “leisure”, and also according to respondents’ own perceptions of the spatial and temporal flexibility of each out-of-home activity that they had executed. While we cannot strictly separate cohort effects from age-related effects, after controlling for gender and household structure, we infer that age and related lifestyle effects dominate in explaining these propensities. However, the boomers were the only age stratum to increase their total out-of-home activity participation over the course of the panel, an intriguing starting point for the future study of this cohort.

CO2 emissions: Are land-use changes enough for California to reduce VMT? Specification of a two-part model with instrumental variables

With vehicle miles of travel increasing at a faster pace than population, one strategy being actively pursued by both state and local governments is compact development. California recently passed legislation that aggressively promotes sustainability by endorsing and rewarding compact development. Likewise, the California Air Resources Board has set a statewide reduction target of 5MMT of greenhouse gas reductions from land use, based largely on achieving compact development patterns. In this paper, we use a multivariate two-part model with instrumental variables, which corrects for residential location self-selection bias. We use a much larger and more geographically representative travel survey on household travel patterns and socio-economic characteristics than represented in previous California studies; this allows us to robustly consider other influences on travel. Our results indicate that, all else equal, a 10% in residential density would reduce VMT by 1.9%. This elasticity is larger than the reported in previous econometric studies for the US, and specifically for California. However, as we show, the magnitude of this impact is still low considering reasonable ranges for policies aimed to increase residential density.     

Can agents measure up? A comparative study of an agent-based and on-line optimization approach for a drayage problem with uncertainty

Experiments studying the behavior of agent-based methods over varying levels of uncertainty in comparison to traditional optimization methods are generally absent from the literature. In this paper we apply two structurally distinct solution approaches, an on-line optimization and an agent-based approach, to a drayage problem with time windows under two types of uncertainty. Both solution approaches are able to respond to dynamic events. The on-line optimization approach utilizes a mixed integer program to obtain a feasible route at 30-s intervals. The second solution approach deploys agents that engage in auctions to satisfy their own objectives based on the information they perceive and maintain locally. Our results reveal that the agent-based system can outperform the on-line optimization when service time duration is highly uncertain. The on-line optimization approach, on the other hand, performs competitively with the agent-based system under conditions of job-arrival uncertainty. When both moderate service time and job-arrival uncertainties are combined, the agent system outperforms the on-line optimization; however, in the case of extremely high combined uncertainty, the on-line optimization outperforms the agent-based approach.