Activity-based Travel Model Research Articles

Modeling Exposure to Mobility-Related Pollution: Review and Key Challenges

The health and equity issues related to urban pollution, coupled with the need to assess the consequences of implementing transport policies or new mobility solutions, make it imperative to provide decision-makers with a relevant modeling framework. The coupling of activity-based travel models with environmental models presents a promising approach to thoroughly investigate the interconnected issues related to the impact of road traffic on individual exposure to noise and air pollution. Integrated modeling chains provide a robust framework to comprehensively assess traffic scenarios on a city scale and address issues related to environmental inequalities. This article reviews the recent work in this field to identify both the main contributions and key challenges still to be addressed. It illustrates how the use of a dynamic approach, which considers the intraday dynamics of individual mobilities and pollution concentrations, can contribute to more accurate estimates of the exposure of individuals to pollutants. This modeling approach also has the potential to investigate how the level of exposure depends on the characteristics of the individuals (socioeconomic factors, commuting patterns, location of residence, and main activities). Nonetheless, there are still important challenges to overcome to further enhance the accuracy and applicability of these models: taking into account all types of traffic and transport modes; refining the modeling of exposure to pollution during travel; taking better consideration of microenvironments; and so forth.

Effectiveness of travel behavior and infrastructure change to mitigate heat exposure

Urban heat exposure is an increasing health risk among urban dwellers. Many cities are considering accommodating active mobility, especially walking and biking, to reduce greenhouse gas emissions. However, promoting active mobility without proper planning and transportation infrastructure to combat extreme heat exposure may cause more heat-related morbidity and mortality, particularly in future with projected climate change. This study estimated the effectiveness of active trip heat exposure mitigation under built environment and travel behavior change. Simulations of the Phoenix metro region's 624,987 active trips were conducted using the activity-based travel model (ABM), mean radiant temperature (TMRT, net human radiation exposure), transportation network, and local climate zones. Two scenarios were designed to reduce traveler exposure: one that focuses on built environment change (making neighborhoods cooler) and the other on travel behavior (switching from shorter travel time but higher exposure routes to longer travel time but cooler routes) change. Travelers experienced TMRT heat exposure ranging from 29°C to 76°C (84°F to 168°F) without environmental or behavioral change. Active trip TMRT exposures were reduced by an average of 1.2–3.7°C when the built environment was changed from a hotter to cooler design. Behavioral changes cooled up to 10 times more trips than changes in built environment changes. The marginal benefit of cooling decreased as the number of cooled corridors transformed increased. When the most traveled 10 km of corridors were cooled, the marginal benefit affected over 1,000 trips/km. However, cooling all corridors results in marginal benefits as low as 1 trip/km. The results reveal that heavily traveled corridors should be prioritized with limited resources, and the best cooling results come from environment and travel behavior change together. The results show how to surgically invest in travel behavior and built environment change to most effectively protect active travelers.

Advancing the practice of regional transportation equity analysis: a San Francisco bay area case study

As the transportation industry continues to evolve, it is urgent that we develop and implement methods for clearly evaluating the range of transportation engineering, planning, and policy impacts experienced by various population segments. While theories of transportation equity have advanced over the past decade, such advancements outpace existing methods for evaluating the fairness of large-scale transportation investments for disadvantaged communities. In this study, a regional activity-based travel model for the Bay Area, California is used to perform an equity analysis of two of the region’s transportation and land-use planning scenarios. Equity outcomes are tested relative to three equity standards: Equality, Proportionality, and Rawlsian justice. The primary objective is to demonstrate the usefulness of a full-scale activity-based travel model for regional transportation equity analysis. We demonstrate that fine-grained distributional measures play an important role in examining the individual and household-level impacts of regional transportation scenarios, and can complement existing Environmental Justice assessments and equity analyses by helping to explain underlying reasons for average group impacts. Distributional measures can further reveal harmful cases when disadvantaged groups are most likely to experience the disbenefits of the transportation scenarios. Yet, each type of measures in isolation does not tell the complete story of which planning scenario is likely to deliver more equitable outcomes. Finally, we demonstrate the significance of applying equity standards for ranking planning scenarios, and we find that the ranking of scenarios will vary according to the equity standard, as well as how associated evaluation criteria are defined.

A systematic review of the agent-based modelling/simulation paradigm in mobility transition

Simulations and agent-based modelling (ABM) have gained momentum as techniques in the transport, energy, and technology diffusion literature to analyse the mobility transition as a complex emergent phenomenon. This study systematically reviews the application of the ABM paradigm in 86 mobility transition studies. The study reveals several research gaps and proposes avenues for future research. Our review highlights that (i) the field has considerably matured in studying the diffusion of electric vehicles, (ii) both price-based and preference-based scenarios for mobility transition should be considered in future research, (iii) most of the empirical model calibrations have been confined to Western countries. Not only will the literature benefit from similar research in other Western regions, but also from non-Western nations with their unique mobility transition pathways, (iv) the conceptual modelling framework of studies can be divided into the two categories of theory-driven and heuristic models. The theory-driven models, which include psychological and non-psychological (e.g., activity-based travel) models, tend to use well-established behavioural rules, (v) most of the models have used the random utility maximization concept, social psychological models, and other simple assumptions/thresholds for the decision-making process of agents, (vi) half of the studies did not validate their models, and (vii) two-thirds of studies omitted to discuss interaction topology among agents. Major remaining challenges and gaps are identified in the review.

An activity-based travel and charging behavior model for simulating battery electric vehicle charging demand

The expansion of the battery electric vehicle (BEV) market requires considerable changes in the supply of electricity to fulfill the charging demand. To this end, understanding the spatio-temporal distribution of BEV charging demand at a micro level is crucial for optimal electric vehicle supply equipment (EVSE) planning and electricity load management. This research proposes an integrated activity-based BEV charging demand simulation model, which considers both realistic travel and charging behaviors and provides high-resolution spatio-temporal demand in real-world applications. Moreover, a novel charging choice model is proposed which provides more realistic demand modeling by allowing critical non-linearities in random utility to better describe observed charging behaviors. The results of a case study for the Atlanta metropolitan area imply that work/public charging has a substantial potential market, which can comprise up to 64.5% of the total demand. Out of multiple charging modes, demand for direct-current fast charging (DCFC) is prominent at work/public locations, and it makes up the largest portion of the non-residential demand in all simulation scenarios. Moreover, charging behaviors have significant impacts on the demand distribution. Peak power demand for use of level-2 chargers is 49% to 91% higher among high-risk-sensitive users than among risk-neutral users. Users’ preferences for fast charging rates can change the DCFC demand from 36.4% of the total demand to 53.7% of the total. This study helps to qualitatively analyze the factors that figure in charging demand and their impacts on the demand distribution. The results can be directly used in EVSE planning and electricity load prediction.

The case for microsimulation frameworks for integrated urban models

The primary objective of this paper is to “make the case” for adoption of microsimulation frameworks for development of integrated urban models. Similar to the case of activity-based travel models, microsimulation in integrated urban models enables such models to deal better with: heterogeneity and non-linearity in behavior; identification of the detailed spatial and socioeconomic distribution of impacts, benefits and costs; tracing complex interactions across agents and over time; providing support for modelling memory, learning and adaptation among agents; computational efficiency; and emergent behavior. The paper discusses strengths, weaknesses and challenges in microsimulating urban regions, including the extent to which microsimulation models are still subject to Lee’s famous “seven sins of large-scale modelling,” as well as the extent to which they may help alleviate or reduce these sins in operational models. The paper concludes with a very brief discussion of future prospects for such models.

The Pros and Cons of Using the Change in Destination Choice Logsums as a Practical Measure of User Benefits

The Metropolitan Transportation Commission carries out an extensive project performance assessment in support of regional planning efforts. For the 2017 regional transportation plan, various types of potential transportation investments were evaluated against qualitative and quantitative criteria. The quantitative metric was a benefit–cost ratio, in which a project’s expected annual benefits were compared with the project’s expected annualized cost. To quantify user benefits (one component of the total benefit estimate), agency staff used the difference in Project/No Project destination choice logsums generated by the region’s activity-based travel model. The destination choice models include, as covariates, mode choice logsums, which means the change in destination choice logsums estimates the change in consumer surplus across all destinations and all travel modes. This approach has long been advocated for in academia and was applied by the Federal Transit Administration for evaluating transit investments in the 2000s. This is the first time that it has been fully implemented in a large-scale, multi-modal application that informs investment decisions. This paper discusses the pros and cons of using this approach. The pros include consistency between simulated behavioral changes in the travel model and quantification of benefits, which cannot be achieved using typical measures of benefits such as travel time savings, as well as the ability to map changes in access across the region. The cons include issues related to the form and creation of alternatives in the mode choice model and the difficulty in explaining changes in logsums to non-technical audiences.

Coordinated Daily Activity Pattern Type in an Applied Activity-Based Travel Model System: Specifics of Modeling Large Households and Modality Preferences

This paper describes a modified approach to modeling an individual daily activity-travel pattern (DAP) coordinated at the household level. The model was primarily introduced to handle large households that are typical for the city of Jerusalem. However, the developed method proved useful in adding more behavioral aspects to the model. The study introduced daily modality and added the emphasis on modeling it for all household members simultaneously. It is of special practical value for Jerusalem since such population sectors as ultra-Orthodox Jewish and Arabs are characterized by a large share of persons who have an entire day of travel implemented in a non-motorized fashion. In addition to daily modality, this paper presents a simple but useful approach to understand at-home time-use. A binary choice model was formulated for the main reason of being inactive (working-at-home or other). The paper also discusses possible extensions of the current approach as well as alternative approaches to frame DAP type choice in the context of a complete activity-based travel model (ABM). In particular, this model can be easily extended in order to better address the joint nature of the choices using the Gibbs sampler. The paper discusses how DAP type choice can be framed in many ways and with different levels of detail subject to the ultimate model system design.

Understanding activity engagement across weekdays and weekend days: A multivariate multiple discrete-continuous modeling approach

This paper is motivated by the increasing recognition that modeling activity-travel demand for a single day of the week, as is done in virtually all travel forecasting models, may be inadequate in capturing underlying processes that govern activity-travel scheduling behavior. The considerable variability in daily travel suggests that there are important complementary relationships and competing tradeoffs involved in scheduling and allocating time to various activities across days of the week. Both limited survey data availability and methodological challenges in modeling week-long activity-travel schedules have precluded the development of multi-day activity-travel demand models. With passive and technology-based data collection methods increasingly in vogue, the collection of multi-day travel data may become increasingly commonplace in the years ahead. This paper addresses the methodological challenge associated with modeling multi-day activity-travel demand by formulating a multivariate multiple discrete-continuous probit (MDCP) model system. The comprehensive framework ties together two MDCP model components, one corresponding to weekday time allocation and the other to weekend activity-time allocation. By tying the two MDCP components together, the model system also captures relationships in activity-time allocation between weekdays on the one hand and weekend days on the other. Model estimation on a week-long travel diary data set from the United Kingdom shows that there are significant inter-relationships between weekdays and weekend days in activity-travel scheduling behavior. The model system presented in this paper may serve as a higher-level multi-day activity scheduler in conjunction with existing daily activity-based travel models.

Viewpoint: Integrated urban modeling: Past, present, and future

Integrated urban models (IUMs) (aka, integrated transport/land-use models) have been developed and (sometimes) applied for more than 50 years, dating back to the early 1960s. IUMs have been criticized over this same period on both practical and theoretical grounds. At the same time, continuing and very significant technological developments have made possible the development, implementation and use of such models in operational planning settings in various countries worldwide. A major review of the IUM state of the art and recommendations for evolution of this state were prepared by the author and colleagues 20 years ago. This paper presents an update of the 1998 report in terms of a summary of progress over the past 20 years, a critical assessment of the current IUM state of the art and practice, and needs and prospects for future development. This paper argues that the current modeling state is in “the doldrums,” similar to concerns raised by Pas in the seminal 1990 critique of activity-based travel models. It then outlines research and development needs to exploit current and emerging data, computing, and methodological developments that hold promise for the development of a much more powerful and useful “next generation” of IUMs.

Pricing and Reliability Enhancements in the San Diego, California, Activity-Based Travel Model

The estimation of demand for priced highway lanes is becoming increasingly important to agencies seeking to improve mobility and find alternative revenue sources for the provision of transportation infrastructure. However, many modeling tools fall short of what is required for robust estimates of demand with respect to toll and managed lanes in two key areas: the value of time is often aggregate and not consistently defined throughout the model system, and the reliability of transport infrastructure is rarely considered. This paper describes an effort that implemented recommendations of the Strategic Highway Research Program on pricing and reliability within a regional activity-based modeling system for the San Diego, California, region. The implemented recommendations included distributed travel time sensitivities across the synthetic population and special travel markets, continuous cost sensitivity on the basis of income, and multiple value of time bins in highway skimming and assignment. The work also included innovative research related to the analysis of travel time variability on the basis of a temporally disaggregate (1-min interval) data set of automobile travel speeds for most automobile links in the San Diego network for the month of October 2012. Regression equations that related the travel time reliability to link characteristics, incorporated reliability in automobile travel skims, incorporated those skims in the travel demand model system, and calculated toll elasticity on toll roads in San Diego County were estimated. The enhanced model matched observed toll demand better than the original model. Resulting elasticity values were generally found to be in the ranges reported in the literature.

Analysis of the Long-Distance and Overnight Travel Tour Planning Process

For many households, long-distance and overnight travel (LDOT) requires considerably more planning and financial investment than everyday travel activities do. As LDOT occupies a larger portion of Americans’ total travel than it did in the past, it is necessary to better understand how planning for LDOT tours contributes to activity-based travel modeling. Existing studies suggest that factors associated with households, persons, and the tours themselves affect the complexity of planning for LDOT, but few studies have ever used data from a data set on the temporal process of planning for a tour over a long period of time. This study used data from a unique survey, the Longitudinal Survey of Overnight Travel, which collected information on planned and completed tours from 1,220 respondents over 1 year. This paper presents descriptive statistics and three statistical models that identify the factors that contribute to the complexity of LDOT tour planning, including the number of legs or stops on the tour, the maximum distance from home, the presence of adult friends or children on the tour, tour duration, and whether a personal vehicle was available on the tour. Other findings highlight the need for improved data collection on LDOT tour making and tour planning, including further description of the external constraints that lead to the cancellation of tour plans.

Choice of Workplace Location

This paper analyzes the spatial patterns of commuting in two-worker households and explores the associated interdependencies of work location decisions for the two workers. A joint multinomial location choice framework is used to model the work location choices of both workers in a multiworker household simultaneously. The model incorporates a rich set of explanatory variables for each worker; the variables have been substantiated in previous research. In addition, the model includes several interaction variables that describe the possible impacts on each worker of the other worker's workplace choices. The corresponding statistical analysis and model estimation are implemented in parallel for four major Ohio cities—Columbus, Cleveland, Cincinnati, and Dayton—on the basis of available household travel surveys. The developed and estimated models have been incorporated as a component in the activity-based travel model being developed for the regions. The model structure and segmentation are discussed in detail; a comparison of the most critical impacts across the regions is included. Contrary to expectations, the results show that workers in two-worker households do not try to trade off a longer commute distance for one worker with a shorter commute distance for the second worker. However, several other interaction effects—captured by the distance between the workplaces and by the angle formed at the home location between the two workplaces— are significant. In particular, the angle is smaller (more acute) for workers who reside in areas with better accessibility to jobs, as well as for full-time workers and low-income workers.

Using an Activity-Based Model to Explore the Potential Impacts of Automated Vehicles

Automated vehicles (AVs) may enter the consumer market with various stages of automation in 10 years or even sooner. Meanwhile, regional planning agencies are envisioning plans for time horizons out to 2040 and beyond. To help decision makers understand the effect of AV technology on regional plans, modeling tools should anticipate its impact on transportation networks and traveler choices. This research uses the Seattle, Washington, region's activity-based travel model to test a range of travel behavior impacts from AV technology development. The existing model was not originally designed with AVs in mind, so some modifications to the model assumptions are described in areas of roadway capacity, user values of time, and parking costs. Larger structural model changes were not yet considered. Results of four scenario tests show that improvements in roadway capacity and in the quality of the driving trip may lead to large increases in vehicle miles traveled, while a shift to per mile usage charges may counteract that trend. Travel models will need to have major improvements in the coming years, especially with regard to shared ride, taxi modes, and the effect of multitasking opportunities, to better anticipate the arrival of this technology.

Achieving reductions in greenhouse gases in the US road transportation sector

It is well established that GHG emissions must be reduced 50 to 80% by 2050 in order to limit global temperature increase to 2°C. Achieving reductions of this magnitude in the transportation sector is a challenge and requires a multitude of policies and technology options. The research presented here analyzes three scenarios: changes in the perceived price of travel, land use intensification, and increases in transit. Elasticity estimates are derived using an activity-based travel model for the state of California and broadly representative of the US. The VISION model is used to forecast changes in technology and fuel options that are currently forecast to occur in the US for the period 2000–2040, providing a life-cycle GHG forecast for the road transportation sector. Results suggest that aggressive policy action is required, especially pricing policies, but also more on the technology side, especially increases in the carbon efficiency of medium and heavy-duty vehicles.

Development of Vehicle Fleet Composition Model System for Implementation in Activity-Based Travel Model

The development of a vehicle fleet composition and utilization model system that may be incorporated into a larger activity-based travel demand model is described. It is of interest and importance to model household vehicle fleet composition and utilization behavior because the energy and environmental impacts of personal travel are dependent not only on the number of vehicles but also on the mix of vehicles that a household owns and the extent to which different vehicles are used. A vehicle composition (fleet mix) and utilization model system was developed for integration into the activity-based travel demand model that was being developed for the greater Phoenix metropolitan area in Arizona. At the heart of the vehicle fleet mix model system is a multiple discrete continuous extreme value model capable of simulating vehicle ownership and use patterns of households. Vehicle choices are defined by a combination of vehicle body type and age category and the model system is capable of predicting vehicle composition and utilization patterns at the household level. The model system is described and results are presented of a validation and policy sensitivity analysis exercise demonstrating the efficacy of the model.

Rural Communities and Transportation Equity in California's San Joaquin Valley

Smart growth policy and planning have tended to emphasize urban centers and regions, yet rural communities can also be important sites of innovation. Recent work demonstrated that these communities had surprisingly high levels of current and potential nonmotorized travel. Legislation in California mandates reductions in greenhouse gas emissions across all of the state's metropolitan planning organization (MPO) regions, including the heavily rural San Joaquin Valley. Advocates for rural communities are finding common cause with more traditional environmental organizations around the vision of investing in and enhancing extant rural places as an alternative to leapfrog patterns of urban and suburban sprawl. Because of existing patterns of extreme disparity and legion underserved unincorporated communities, analyses that can help integrate social equity within regional planning are needed to serve and empower rural residents. This paper presents the results of several new analyses of the social equity dimensions of regional transportation plans in the San Joaquin Valley. Activity-based travel model data were used to analyze equity, with a particular focus placed on eight disadvantaged unincorporated communities identified by community advocates to be important demonstration sites. The investigators showed how improvements to traditional equity analysis could enhance the consideration of equity in the planning process and compared the results developed by innovative techniques with those obtained by use of their traditional counterparts. The methods outlined here can make substantial contributions to reduce disparities in rural communities, which would likely be overlooked in typical regional equity analyses because of their small size, and offer lessons for MPOs serving rural areas across the country.

The location selection problem for the household activity pattern problem

In this paper, an integrated destination choice model based on routing and scheduling considerations of daily activities is proposed. Extending the Household Activity Pattern Problem (HAPP), the Location Selection Problem (LSP–HAPP) demonstrates how location choice is made as a simultaneous decision from interactions both with activities having predetermined locations and those with many candidate locations. A dynamic programming algorithm, developed for PDPTW, is adapted to handle a potentially sizable number of candidate locations. It is shown to be efficient for HAPP and LSP–HAPP applications. The algorithm is extended to keep arrival times as functions for mathematical programming formulations of activity-based travel models that often have time variables in the objective.

Incorporating crowding into the San Francisco activity-based travel model

Information produced by travel demand models plays a large role decision making in many metropolitan areas, and San Francisco is no exception. Being a transit first city, one of the most common uses for San Francisco’s travel model SF-CHAMP is to analyze transit demand under various circumstances. SF-CHAMP v 4.1 (Harold) is able to capture the effects of several aspects of transit provision including headways, stop placement, and travel time. However, unlike how auto level of service in a user equilibrium traffic assignment is responsive to roadway capacity, SF-CHAMP Harold is unable to capture any benefit related to capacity expansion, crowding’s effect on travel time nor or any of the real-life true capacity limitations. The failure to represent these elements of transit travel has led to significant discrepancies between model estimates and actual ridership. Additionally it does not allow decision-makers to test the effects of policies or investments that increase the capacity of a given transit service. This paper presents the framework adopted into a more recent version of SF-CHAMP (Fury) to represent transit capacity and crowding within the constraints of our current modeling software.

Activity-Based Travel Models and Transportation Equity Analysis

The current state of practice for transportation equity analysis is moving toward the use of activity-based travel models for scenario analysis. However, little has been done to validate the use of these models for equity analysis. The first objective of this paper is to present a research framework for the equity analysis of long-range transportation plans, for the purpose of critiquing the current state of practice and pointing to key research needs. This research framework is used to identify four research areas for furthering transportation equity analysis: (a) identifying transportation priorities for different groups of interest, (b) identifying appropriate modeling tools for measuring equity outcomes, (c) identifying and analyzing equity indicators, and (d) linking the equity indicators and the transportation priorities. The second objective is to explore how well an activity-based travel demand model represents heterogeneity of travel behavior among different income classes, for the purpose of transportation equity analysis. Graphical and statistical tools are used to access how well a sample of the model data compares with a sample of travel survey data for the San Francisco Bay Area in California. The results show that although the tests of distribution equality fail, the general shape of the distributions, the central tendencies, and the relative difference of the low- versus high-income samples perform well.