Prediction of travel time variability for cost-benefit analysis

Unreliable travel times cause substantial costs to travelers. Nevertheless, they are often not taken into account in cost-benefit analyses (CBA), or only in very rough ways. This paper aims at providing simple rules to predict variability, based on travel time data from Dutch highways. Two different concepts of travel time variability are used, which differ in their assumptions on information availability to drivers. The first measure is based on the assumption that, for a given road link and given time of day, the expected travel time is constant across all working days (rough information: RI). In the second case, expected travel times are assumed to reflect day-specific factors such as weather conditions or weekdays (fine information: FI). For both definitions of variability, we find that the mean travel time is a good predictor. On average, longer delays are associated with higher variability. However, the derivative of variability with respect to delays is decreasing in delays. It can be shown that this result relates to differences in the relative shares of observed traffic ‘regimes’ (free-flow, congested, hyper-congested) in the mean delay. For most CBAs, no information on the relative shares of the traffic regimes is available. A non-linear model based on mean travel times can then be used as an approximation.     

Additive measures of travel time variability

This paper derives a measure of travel time variability for travellers equipped with scheduling preferences defined in terms of time-varying utility rates, and who choose departure time optimally. The corresponding value of travel time variability is a constant that depends only on preference parameters. The measure is unique in being additive with respect to independent parts of a trip. It has the variance of travel time as a special case. Extension is provided to the case of travellers who use a scheduled service with fixed headway.     

On the cost of misperceived travel time variability

Because individuals may misperceive travel time distributions, using the implied reduced form of the scheduling model might fall short of capturing all costs of travel time variability. We reformulate a general scheduling model employing rank-dependent utility theory and derive two special cases as econometric specifications to study these uncaptured costs. It is found that reduced-form expected cost functions still have a mean–variance form when misperception is considered, but the value of travel time variability is higher. We estimate these two models with stated-preference data and calculate the empirical cost of misperception. We find that: (i) travelers are mostly pessimistic and thus tend to choose departure times too early to achieve a minimum cost, (ii) scheduling preferences elicited using a stated-choice method can be relatively biased if probability weighting is not considered, and (iii) the extra cost of misperceiving the travel time distribution might be nontrivial when time is valued differently over the time of day and is substantial for some people.     

The influence of travel time variability on the value of time

Current benefits from travel time savings have only been related to the benefits from reducing mean travel time. Some previous attempts of including variability in the generalised cost function have mainly assumed commuters with fixed arrival time. This paper presents a comprehensive framework for valuing travel time variability that allows for any journey purpose and arrival time constraint. The proposed model is based on the expected utility approach and the mean-standard deviation approach. Stated Preference methods are considered the best technique for providing the data for calibrating the models. The values of time derived from the models are highly influenced by the value of travel time variability and it strongly depends on the probability distribution function travellers are faced with.     

Review of optimal sensor location models for travel time estimation

The problem of optimally locating fixed sensors on a traffic network infrastructure has been object of growing interest in the past few years. Sensor location decisions models differ from each other according to the type of sensors that are to be located and the objective that one would like to optimize. This paper surveys the existing contributions in the literature related to the problem of locating fixed sensors on the network to estimate travel times. The review consists of two parts: the first part reviews the methodological approaches for the optimal location of counting sensors on a freeway for travel time estimation; the second part focuses on the results related to the optimal location of Automatic Vehicle Identification (AVI) readers on the links of a network to get travel time information.     

Distributions of travel time variability on urban roads

Reliability is an important factor in route, mode and also departure time choice analysis and is a key performance indicator for transport systems. However, the current metrics used to measure travel time variability may be not sufficient to fully represent reliability. Better understanding of the distributions of travel times is needed for the development of improved metrics for reliability. A comprehensive data analysis involving the assessment of longitudinal travel time data for two urban arterial road corridors in Adelaide, Australia, demonstrates that the observed distributions are more complex than previously assumed. The data sets demonstrate strong positive skew, very long upper tails, and sometimes bimodality. This paper proposes the use of alternative statistical distributions for travel time variability, with the Burr Type XII distribution emerging as an appropriate model for both links and routes. This statistical distribution has some attractive properties that make it suitable for explicit definition of many travel time reliability metrics. Copyright © 2011 John Wiley & Sons, Ltd.     

Investigating empirical implications of hysteresis in day-to-day travel time variability

Day-to-day travel time variability plays a significant role in travel time reliability. Nowadays, travelers not only seek to minimize their travel time on average, but also value its variation. The variation in the mean and the variance of travel time (across days, for the same departure time) has not been thoroughly investigated. A temporary decrease in capacity (e.g. congestion caused by an active bottleneck) leads to a quite significant difference in the variance of travel time for congestion onset and offset periods. This phenomenon results in hysteresis loops where the departure time periods in congestion offset exhibit a higher travel time variance than the ones in congestion onset with the same mean travel time. The aim of this paper is to identify empirical implications that yield to the hysteresis phenomenon in day-to-day travel times. First, empirical hysteresis loop observations are provided from two different freeway sites. Second, we investigate the potential link with the hysteresis observed in traffic networks on macroscopic fundamental diagram (MFD). Third, we build a piecewise linear function that models the evolution of travel time within the day. This allows us to decompose the problem into its components, e.g. start time of congestion, peak travel time, etc. These components, along with their probability distribution functions, are employed in a Monte Carlo simulation model to investigate their partial effects on the existence of hysteresis. Correlation among critical variables is the most influential factor in this phenomenon, which should be further investigated regarding traffic flow and traffic equilibrium principles.     

Assessing the effects of stochastic perception error under travel time variability

Perceived mean-excess travel time is a new risk-averse route choice criterion recently proposed to simultaneously consider both stochastic perception error and travel time variability when making route choice decisions under uncertainty. The stochastic perception error is conditionally dependent on the actual travel time distribution, which is different from the deterministic perception error used in the traditional logit model. In this paper, we investigate the effects of stochastic perception error at three levels: (1) individual perceived travel time distribution and its connection to the classification by types of travelers and trip purposes, (2) route choice decisions (in terms of equilibrium flows and perceived mean-excess travel times), and (3) network performance measure (in terms of the total travel time distribution and its statistics). In all three levels, a curve fitting method is adopted to estimate the whole distribution of interest. Numerical examples are also provided to illustrate and visualize the above analyses. The graphical illustrations allow for intuitive interpretation of the effects of stochastic perception error at different levels. The analysis results could enhance the understanding of route choice behaviors under both (subjective) stochastic perception error and (objective) travel time uncertainty. Some suggestions are also provided for behavior data collection and behavioral modeling.     

Behavioural implications of preferences, risk attitudes and beliefs in modelling risky travel choice with travel time variability

The appropriate interpretation of a behavioural outcome requires allowing for risk attitude and belief of an individual, in addition to identification of preferences. This paper develops an Attribute-Specific Extended Rank-Dependent Utility Theory model to better understand choice behaviour in the presence of travel time variability, in which these three important components of choice are empirically addressed. This framework is more behaviourally appealing for travel time and travel time variability research than the traditional approach in which risk attitude and belief are overlooked. This model also reveals significant unobserved between-individual heterogeneity in preferences, risk attitudes and beliefs.     

Electric and hydrogen buses: Shifting from conventionally fuelled cars in the UK

For the UK to meet their national target of net zero emissions as part of the central Paris Agreement target, further emphasis needs to be placed on decarbonizing public transport and moving away from personal transport (conventionally fuelled vehicles (CFVs) and electric vehicles (EVs)). Electric buses (EBs) and hydrogen buses (HBs) have the potential to fulfil requirements if powered from low carbon renewable energy sources.A comparison of carbon dioxide (CO₂) emissions produced from conventionally fuelled buses (CFB), EBs and HBs between 2017 and 2050 under four National Grid electricity scenarios was conducted. In addition, emissions per person at different vehicle capacity levels (100%, 75%, 50% and 25%) were projected for CFBs, HBs, EBs and personal transport assuming a maximum of 80 passengers per bus and four per personal vehicle.Results indicated that CFVs produced 30 gCO₂ km⁻¹ per person compared to 16.3 gCO₂ km⁻¹ per person by CFBs by 2050. At 100% capacity, under the two-degree scenario, CFB emissions were 36 times higher than EBs, 9 times higher than HBs and 12 times higher than EVs in 2050. Cumulative emissions under all electricity scenarios remained lower for EBs and HBs.Policy makers need to focus on encouraging a modal shift from personal transport towards sustainable public transport, primarily EBs as the lowest level emitting vehicle type. Simple electrification of personal vehicles will not meet the required targets. Simultaneously, CFBs need to be replaced with EBs and HBs if the UK is going to meet emission targets.     

Quantifying the benefit of responsive pricing and travel information in the stochastic congestion pricing problem

This paper is concerned with roadway pricing amidst the uncertainty which characterizes long-term transportation planning. Uncertainty is considered both on the supply-side (e.g., the effect of incidents on habitual route choice behavior) and on the demand-side (e.g., due to prediction errors in demand forecasting). The framework developed in this paper also allows the benefits of real-time travel information to be compared directly against the benefits of responsive pricing, allowing planning agencies to identify the value of these policy options or contract terms in publicly-operated toll roads. Specifically, six scenarios reflect different combinations of policy options, and correspond to different solution methods for optimal tolls. Demonstrations are provided on both the Sioux falls and Anaheim networks. Results indicate that providing information to drivers implemented alongside responsive tolling may reduce expected total system travel time by over 9%, though more than 8% of the improvement is due to providing information, with the remaining 1% improvement gained from responsive tolling.     

The cost of travel time variability: Three measures with properties

This paper explores the relationships between three types of measures of the cost of travel time variability: measures based on scheduling preferences and implicit departure time choice, Bernoulli type measures based on a univariate function of travel time, and mean-dispersion measures. We characterise measures that are both scheduling measures and mean-dispersion measures and measures that are both Bernoulli and mean-dispersion. There are no measures that are both scheduling and Bernoulli. We consider the impact of requiring that measures are additive or homogeneous, proving also a new strong result on the utility rates in an additive scheduling measure. These insights are useful for selecting cost measures to use in applications.     

Identifying optimal data aggregation interval sizes for link and corridor travel time estimation and forecasting

With the recent increase in the deployment of ITS technologies in urban areas throughout the world, traffic management centers have the ability to obtain and archive large amounts of data on the traffic system. These data can be used to estimate current conditions and predict future conditions on the roadway network. A general solution methodology for identifying the optimal aggregation interval sizes for four scenarios is proposed in this article: (1) link travel time estimation, (2) corridor/route travel time estimation, (3) link travel time forecasting, and (4) corridor/route travel time forecasting. The methodology explicitly considers traffic dynamics and frequency of observations. A formulation based on mean square error (MSE) is developed for each of the scenarios and interpreted from a traffic flow perspective. The methodology for estimating the optimal aggregation size is based on (1) the tradeoff between the estimated mean square error of prediction and the variance of the predictor, (2) the differences between estimation and forecasting, and (3) the direct consideration of the correlation between link travel time for corridor/route estimation and forecasting. The proposed methods are demonstrated using travel time data from Houston, Texas, that were collected as part of the automatic vehicle identification (AVI) system of the Houston Transtar system. It was found that the optimal aggregation size is a function of the application and traffic condition.

Estimation of the value of travel time and of travel time reliability for heterogeneous drivers in a road network

Transportation - Travel time reliability has been recognized as an important factor in cost–benefit analysis in a transportation network. To estimate the benefit and cost of travel time...     

Systematic variability in repetitious travel

The focus of this paper is the degree to which day-to-day variability in the individual's travel pattern has a systematic, or nonrandom, component. We first review the different sources of variability in travel, emphasizing the difference between between-individual and within-individual variation and the implications of this difference for travel analysis. After discussing the impact of measurement (i.e. the way in which travel behavior is measured) on the study of repetition and variability, we use the Uppsala data to examine the level of systematic variability in an individual's longitudinal travel record. The analysis focuses on two questions:

- How well does observation over one week capture longer-term (five-week) travel behavior; in other words, is behavior highly repetitive from week to week?

- How systematic is within-individual variability; in other words, are certain stops distributed over the five-week record in a nonrandom, that is either regular or clustered, fashion?

Using measures of travel that include more than one stop attribute (e.g. activity, mode, time of day, and location), we found that:

- A seven-day record of travel does not capture most of the separate behaviors exhibited by the individual over a five-week period, but it does capture, for most people, a good sampling of the person's different typical daily travel patterns.

- Whereas a considerable portion of intraindividual variability is systematic (nonrandom), clustering is a more important source of nonrandom variation than is regularity.

The results suggest that behavior does not follow a weekly cycle closely enough for a one-week travel record to measure the longer-term frequency with which the individual makes certain stops or to assess the level of day-to-day variation present in the individual's record. Because these results are likely to reflect the particular measures of behavior we used, one conclusion of this study is the need for other studies that replicate the aims of this one but use a variety of other travel measures. Only through such additional work can we truly assess the sensitivity of our findings to measurement techniques.     

Embedding risk attitude and decision weights in non-linear logit to accommodate time variability in the value of expected travel time savings

In recent years we have seen important extensions of logit models in behavioural research such as incorporation of preference and scale heterogeneity, attribute processing heuristics, and estimation of willingness to pay (WTP) in WTP space. With rare exception, however, a non-linear treatment of the parameter set to allow for behavioural reality, such as embedded risk attitude and perceptual conditioning of occurrence probabilities attached to specific attributes, is absent. This is especially relevant to the recent focus in travel behaviour research on identifying the willingness to pay for reduced travel time variability, which is the source of estimates of the value of trip reliability that has been shown to take on an increasingly important role in project appraisal. This paper incorporates, in a generalised non-linear (in parameters) logit model, alternative functional forms for perceptual conditioning (known as probability weighting) and risk attitude in the utility function to account for travel time variability, and then derives an empirical estimate of the willingness to pay for trip time variability-embedded travel time savings as an alternative to separate estimates of time savings and trip time reliability. We illustrate the richness of the approach using a stated choice data set for commuter choice between unlabelled attribute packages. Statistically significant risk attitude parameters and parameters underlying decision weights are estimated for multinomial logit and mixed multinomial logit models, along with values of expected travel time savings.     

Investigating the effect of travel time variability on drivers' route choice decisions in Shanghai,China

Abstract

This paper investigates the effect of travel time variability on drivers' route choice behavior in the context of Shanghai, China. A stated preference survey is conducted to collect drivers' hypothetical choice between two alternative routes with designated unequal travel time and travel time variability. A binary choice model is developed to quantify trade-offs between travel time and travel time variability across various types of drivers. In the model, travel time and travel time variability are, respectively, measured by expectation and standard deviation of random travel time. The model shows that travel time and travel time variability on a route exert similarly negative effects on drivers' route choice behavior. In particular, it is found that middle-age drivers are more sensitive to travel time variability and less likely to choose a route with travel time uncertainty than younger and elder drivers. In addition, it is shown that taxi drivers are more sensitive to travel time and more inclined to choose a route with less travel time. Drivers with rich driving experience are less likely to choose a route with travel time uncertainty.     

Inter-temporal variation in the travel time and travel cost parameters of transport models

The parameters for travel time and travel cost are central in travel demand forecasting models. Since valuation of infrastructure investments requires prediction of travel demand for future evaluation years, inter-temporal variation of the travel time and travel cost parameters is a key issue in forecasting. Using two identical stated choice experiments conducted among Swedish drivers with an interval of 13 years, 1994 and 2007, this paper estimates the inter-temporal variation in travel time and cost parameters (under the assumption that the variance of the error components of the indirect utility function is equal across the two datasets). It is found that the travel time parameter has remained constant over time but that the travel cost parameter has declined in real terms. The trend decline in the cost parameter can be entirely explained by higher average income level in the 2007 sample compared to the 1994 sample. The results support the recommendation to keep the travel time parameter constant over time in forecast models, but to deflate the travel cost parameter with the forecasted income increase among travellers and the relevant income elasticity of the cost parameter. Evidence from this study further suggests that the inter-temporal and the cross-sectional income elasticities of the cost parameter are equal. The average elasticity is found to be ?0.8 to ?0.9 in the present sample of drivers, and the elasticity is found to increase with the real income level, both in the cross-section and over time.     

The significance and measurement of variability in travel behaviour

It is argued that an understanding of variability is central to the modelling of travel behaviour and the assessment of policy impacts, and is not the peripheral issue that it has often been considered. Drawing on recent studies in the UK and Australia, in conjunction with a review of the literature, the paper first examines the policy and analytical rationale for using multi-day data, then illustrates different ways of measuring variability, and finally discusses issues relating to the collection of suitable data for such analyses. In a policy context, there is a growing need for multi-day data to examine issues that affect general rather than one-day behaviour (e.g. to assess the distribution of user charges for road pricing, or patterns of public transport usage); while analytically, multi-day data is needed to improve our ability to identify the mechanisms behind travel behaviour and to derive better empirical relationships. Three measures of variability are presented: a graphical form showing daily differences in behaviour at the individual level; an aggregate, similarity index; and a hybrid graphical/numerical measure, which provides new insights into variability in daily patterns of behaviour. The paper raises a number of issues for debate, probably the most crucial of which is: variability in what? The way in which behaviour is measured crucially affects our conception of stability and variability.