Pavement management and rehabilitation systems require reliable detection and assessment of road pavement distresses. Traditional methods that rely on visual inspections and manual surveys are time-consuming and inconsistent. Several computer vision techniques for distress detection automation have been developed. Most techniques are based on fully supervised labeling, which is especially challenging with respect to locating the exact position of a distress in an image. This paper introduces an innovative approach combining machine learning and image analysis techniques for the automated classification and localization of multiple pavement distresses. In contrast to previous approaches that detected a single distress per image, the proposed approach can both classify and locate multiple distress types per image without prior knowledge of the distress' location. The system classifies multiple distress types within a single image, using a multilabel convolutional neural network (CNN). Distress localization is enhanced through an interpretability method, namely, integrated gradients, followed by postprocessing denoising techniques such as dilatation and connected components. The proposed method is applied in a large-scale dataset collected as part of the pavement management system, which contains information about the number and type of distresses, but not their location in the image. Results indicate that the multilabel CNN model for classifying distress types reached a precision of 84.41%, and the proposed localization method reached 77% overall accuracy, 63.3% for images with multiple distress types, and 82.9% for images with a single distress. The method localizes line distress types more precisely than area distress types. The selected postprocessing parameters critically affect the localization precision.

This paper investigates the mode-shift potentials of a road-space reallocation policy to reduce car use, improve liveability, and reduce transport-related emissions in urban areas. We show the mode-choice effects of repurposing 46 % of Zürich road surface area from cars to bicycles, e-bikes (25 km/h) and s-pedelecs (45 km/h) as well as internalizing external transport costs. The preferences for the different modes are analysed and discussed through the estimation of an Integrated Choice and Latent Variable mode-choice model. The model results show how s-pedelecs and e-bikes have a substantially higher demand potential than conventional bicycles among car drivers. At the same time, latent preferences are not only defined through car ownership in a binary way, namely owning a car or not, but also vary substantially when comparing owners of different car types. Besides considering individual-specific cycling travel times for the three different bicycle types, we also advance mode-choice modelling for cycling by introducing a new cycling infrastructure interaction parameter that includes route-specific cycling infrastructure information. Findings show that reallocating road space is more effective in promoting sustainable mobility than pricing mechanisms alone, highlighting the need for integrated policy measures to facilitate behavioral change.

This paper proposes an integrated equilibrium model to characterize the complex interactions between electrified logistics systems and electric power delivery systems. The model consists of two major players: an electrified logistics operator (ELO) and a power system operator (PSO). The ELO aims to maximize its profit by strategically scheduling and routing its electric delivery vehicles (e-trucks) for deliveries and charging, in response to the locational marginal price (LMP) set by the PSO. The routing, delivery, and charging behaviors of e-trucks are modeled by a perturbed utility Markov decision process (PU-MDP) while their collective operations are optimized to achieve the ELO's objective by designing rewards in the PU-MDP. On the other hand, PSO optimizes the energy price by considering both the spatiotemporal e-truck charging demand and the base electricity load. The equilibrium of the integrated system is formulated as a fixed point, proved to exist under mild assumptions, and solved for a case study on the Hawaii network via Anderson's fixed-point acceleration algorithm. Along with these numerical results, this paper provides both theoretical insights and practical guidelines to achieve sustainable and efficient operations in modern electrified logistics and power systems.

One of the most challenging tasks in pavement management and rehabilitation is to detect and classify different distress types from images collected during field surveys. In this paper, a multilabel convolutional neural network (CNN) model for classifying asphalt distress is proposed. Unlike typical CNN models that classify a single object per image, the proposed model can detect and classify multiple distress types per image, without prior knowledge of the distress location. The model can classify the distress types into four categories: alligator cracking, block cracking, longitudinal/transverse cracking, and pothole. The proposed model was trained and tested on a real data set comprising 42,520 images using different pretrained architectures with various hyperparameter combinations. The results demonstrate the robustness of the proposed model and its potential for crack detection and localization using weakly supervised machine learning methods that can cope with partially labeled data sets.

This paper presents a new ridesharing simulation model that accounts for dynamic driver supply and passenger demand, and complex interactions between drivers and passengers. The proposed simulation model explicitly considers driver and passenger acceptance/rejection on the matching options, and cancelation before/after being matched. New simulation events, procedures and modules have been developed to handle these realistic interactions. Ridesharing pricing bounds that result in high matching option accept rate are derived. The capabilities of the simulation model are illustrated using numerical experiments. The experiments confirm the importance of considering supply and demand interactions and provide new insights to ridesharing operations. Results show that higher prices are needed to attract drivers with short trip durations to participate in ridesharing, and larger matching window could have negative impacts on overall ridesharing success rate. Comparison results further illustrate that the proposed simulation model is able to replicate the predefined “true” success rate, in the cases that driver and passenger interactions occur.

This paper proposes a general equilibrium model for multi-passenger ridesharing systems, in which interactions between ridesharing drivers, passengers, platforms, and transportation networks are endogenously captured. Stable matching is modeled as an equilibrium problem in which no ridesharing driver or passenger can reduce his/her ridesharing disutility by unilaterally switching to another matching sequence. This paper is one of the first studies that explicitly integrates the ridesharing platform's multi-passenger matching problem into the model. By integrating matching sequence with hyper-network, ridesharing-passenger transfers are avoided in a multi-passenger ridesharing system. Moreover, the matching stability between the ridesharing drivers and passengers is extended to address the multi-OD multi-passenger case in terms of matching sequence. The paper provides a proof for the existence of the proposed general equilibrium. A sequence-bush algorithm is developed for solving the multi-passenger ridesharing equilibrium problem. This algorithm is capable to handle complex ridesharing constraints implicitly. Results illustrate that the proposed sequence-bush algorithm outperforms general-purpose solver, and provides insights into the equilibrium of the joint stable matching and route choice problem. Numerical experiments indicate that ridesharing trips are typically longer than average trip lengths. Sensitivity analysis suggests that a properly designed ridesharing unit price is necessary to achieve network benefits, and travelers with relatively lower values of time are more likely to participate in ridesharing.

Among the discrete choice contexts discussed in the transportation literature, route choice is particularly challenging, for which several model structures with different assumptions were developed. Recently, a group of closed-form multinomial models with asymmetric choice probability functions have been proposed to resolve the class imbalance problem in mode choice. However, these asymmetric models can hardly be estimated when the choice sets vary with observations. This paper fills the gap between the asymmetric models and the existing literature, and also adapt the models to route choice by proposing three approaches to further parameterize the shape parameter. This paper also tests the asymmetric models by two independent GPS datasets: taxi data collected in Guangzhou as well as bicycle data collected in Tel Aviv. Comparative analysis is conducted between taxi drivers and cyclists on their route choice behavior. The results indicate the existence of class imbalance in both cases. Moreover, the asymmetric models have case-dependent performance, and different parameterizations correspond to different interpretations of the asymmetry in route choice.

Disaster management systems apply different evacuation models to provide transportation responses to disaster situations. The main goal of the evacuation models is to reduce evacuation times to save lives. The significant increase in the use of location-based social networks (LBSN) as a major media factor can provide data for inferring the Origin-Destination matrix that represents the demands for trips during routine times. To achieve the goal of reducing overall evacuation travel times, an innovative methodology was developed that supports LBSN data for both demand prediction and decentralized personal destination recommendation. A modular multi-dimensional tool for evacuation scenario simulation (MMDT-ESS) was developed for supporting this methodology. This tool can handle multiple scenarios and compute performance evaluation values of VhT (Vehicle-Hours Traveled). MMDT-ESS can be applied using both conventional evacuation models (i.e., evacuation to predefined shelters) or decentralized personalized evacuation models and can provide different recommendations: either self-evacuation to nearest family or best friend that is situated outside the affected area or to a predefined shelter area. A test case using data from a U.S. metropolitan agency showed significant reduction in VhT when choosing evacuation scenarios with LBSN data for decentralized evacuation. Sensitivity analysis on main model parameters illustrate the robustness of the results.

Short-term traffic forecasting is a key element in proactive traffic management, e.g., mitigating the negative effect of impending congestion through appropriate capacity allocation at signalized intersections. In this study, we develop a data-driven methodology for reliably and robustly predicting impending stable congestion. By incorporating feature engineering techniques into an iterative machine learning process, we develop a prediction model that can be intuitively understood by traffic experts and is amenable to diagnostics during implementation. Our iterative machine learning process combines the embedded and filter approaches for feature selection with the use of expert knowledge to create aggregative input variables. The embedded approach is represented by application of a decision tree algorithm, while the filter approach is reflected in use of the mean decrease in accuracy output of a random forest algorithm for identifying expressive variables. We tested the methodology by applying it to field data from a sub-network in Tel Aviv. We demonstrated a reduction in the number of decision tree input variables from 66 raw variables to the five most effective aggregative ones, while achieving statistically significant improvement in all performance indicators. The identification rate of stable congestion increased from 65% to 74% while the robustness of the results was enhanced: the standard deviations of the identification and false alarm rates fell from 8% to 3%, respectively, to 5% and 2%.

Background: For many decades, car-following (CF) and congestion models have assumed a basic invariance: drivers’ default driving strategy is to keep the safety distance. The present study questions that Driving to keep Distance (DD) is a traffic invariance and, therefore, that the difference between the time required to accelerate versus decelerate must necessarily determine the observed patterns of traffic oscillations. Previous studies have shown that drivers can adopt alternative CF strategies like Driving to keep Inertia (DI) by following basic instructions. The present work aims to test the effectiveness of a DI course that integrates 4 tutorials and 4 practice sessions in a standard PC computer designed to learn more adaptive driving behaviors in dense traffic. Methods. Sixty-eight drivers were invited to follow a leading car that varied its speed on a driving simulator, then they took a DI course on a PC computer, and finally they followed a fluctuating leader again on the driving simulator. The study adopted a pretest-intervention-posttest design with a control group. The experimental group took the full DI course (tutorials and then simulator practice). The control group had access to the DI simulator but not to the tutorials. Results. All participating drivers adopted DD as the default CF mode on the pre-test, yielding very similar results. But after taking the full DI course, the experimental group showed significantly less accelerations, decelerations, and speed variability than the control group, and required greater CF distance, that was dynamically adjusted, spending less fuel in the post-test. A group of 8 virtual cars adopting DD required less space on the road to follow the drivers that took the DI course.

Choice set generation is a challenging task, since the consideration set is generally unknown to the modelers, and the full choice set could be too large to be enumerated. The proposed variational autoencoder approach (VAE) is motivated by the idea that the chosen alternatives must belong to the consideration set. The proposed VAE method explicitly considers maximizing the likelihood of including the chosen alternatives in the choice set and inferring the underlying generation process. This paper derives the generalized extreme value (GEV) model with implicit availability/perception (IAP) of alternatives, for bridging the VAE with choice modeling. Specifically, the cross-nested logit (CNL) model with IAP is derived as an example of IAP-GEV models. The IAP approach assumes each alternative is associated with an implicit degree of availability/perception (likelihood in the context of VAE) to be included in the choice set. The VAE approach for route choice set generation is illustrated in a toy network. Simulation experiments show that the proposed method could reproduce the pre-defined true values. We further exemplify the VAE approach using a real dataset. The IAP-CNL model estimated has the best performance in terms of goodness-of-fit and prediction performance, compared to multinomial logit models and conventional choice set generation methods.

Pavement condition index (PCI) is commonly used in pavement management systems (PMS) for indicating the extent of the distresses on the pavement surface. PCI values are a function of distress type, severity, and density. Artificial neural network (ANN) techniques have successfully modeled the performance of in-service pavements, due to their efficiency in predicting and solving nonlinear relationships and dealing with uncertain large amounts of data. Aiming to investigate and examine the efficiency and reliability of ANN models, this paper develops and trains a deep ANN (DNN) model to predict the PCI values and compares the DNN model performance against conventional prediction methods, such as linear and nonlinear regression. Several models with different hyperparameters and architecture were developed and trained using 536,848 samples and tested on 134,212 samples. The root mean square error (RSME) of the tested DNN model is significantly superior to the best fitted linear and nonlinear regression models. In line with the literature, the most influencing variables for PCI prediction are distresses related to alligator cracking, swelling, rutting, and potholes. These findings suggest that DNN models could be incorporated into the PMS for PCI determination.

This paper presents the concept of a modular multi-dimensional tool (MMDT) for evacuation planning models. The goal of MMDT is to propose alternative route and destination locations that can be evaluated and compared to one another. The proposed tool can represent a very large number of scenarios and its strength is in its modularity and efficiency. The MMDT can be applied using both conventional evacuation models and decentralised personalised evacuation models based on Location-Based Social Networks (LBSN) to reduce overall evacuation times. Large-scale test cases using anonymous LBSN data illustrate the MMDT on several scenarios. Results indicate a significant reduction in evacuation times when using decentralised personal evacuation.

On-demand peer-to-peer ridesharing services provide flexible mobility options and are expected to alleviate congestion by sharing empty car seats. An efficient matching algorithm is essential to the success of a ridesharing system. The matching problem is related to the well-known dial-a-ride problem, which also tries to find the optimal pickup and delivery sequence for a given set of passengers. In this paper, we propose an efficient dynamic tree algorithm to solve the on-demand peer-to-peer ridesharing matching problem. The dynamic tree algorithm benefits from given ridesharing driver schedules and provides satisfactory runtime performances. In addition, an efficient pre-processing procedure to select candidate passenger requests is proposed, which further improves the algorithm performance. Numerical experiments conducted in a small network show that the dynamic tree algorithm reaches the same objective function values of the exact algorithm, but with shorter runtimes. Furthermore, the proposed method is applied to a larger size problem. Results show that the spatial distribution of ridesharing participants influences the algorithm performance. Sensitivity analysis confirms that the most critical ridesharing matching constraints are the excess travel times. The network analysis suggests that small vehicle capacities do not guarantee overall vehicle-kilometer travel savings.

Residential location choice is a fundamental process determining real-estate dynamics, regional development, and spatiotemporal traffic flows. Interest in residential choice modeling spans over half a century of extensive research in regional science, urban economics, and transportation. Following the theoretical foundation of the bid-rent curve, practical needs of simulating urban dynamics have led to household-based utilitarian specification on the basis of microeconomics. Today's residential location choice modeling capacity includes the representation of multiple-earner households, spatial correlation between alternatives, extensively large choice sets, population heterogeneity, joint decisions, and extensively large choice sets. This entry outlines the theoretical foundation of residential choice modeling, surveys current modeling practices, discusses implementation in agent-based models, and offers new research directions. This article discusses the definition of the residential choice bundle, choice set and the decision-maker unit, compensatory and noncompensatory decision rules, joint decisions, population strata, and data sources. New research directions refer to online data sources, shared economy, soft amenities and lifestyle, attitudes, and norms.

This paper proposes a novel combination of machine learning techniques and discrete choice models for route choice modeling. The data-driven choice set generation method identifies routes characteristics by clustering, and implicitly generates the choice set by sampling route characteristic attributes from the clusters. Important features are selected by random forests for route choice model development. With the selected features, the methodological-iterative approach is applied to specify the utility functions and to find significant explanatory variables automatically. Results show that the proposed data-driven method produces a discrete route choice model not only with strong explanatory power, but also with high prediction accuracy compared to models estimated with conventional choice set generation methods.

Automated driving technology along with electric propulsion are widely expected to fundamentally change our transport systems. They may not only allow a more productive use of travel time, but will likely trigger completely new business models in the mobility market. A key determinant of the future prospects of both existing and new mobility services will be their production costs. Hence, in this research the production costs of various transport modes both today and in an automated-electric future are analyzed. To account for different local contexts, the study is conducted for 17 cities across the globe. The results indicate that high-income countries will benefit the most from vehicle automation, while only smaller changes can be expected in lower-income countries. This is due to the different relative contribution of labor cost to the total cost of current taxi and bus operations. In a likely final state, transportation costs will be largely decoupled from a country's income level, which will favor productivity in higher-income locations. While this research provides valuable first insights into potential future developments, the underlying assumptions will need to be updated as better information becomes available.

This paper proposes a frequency based transit assignment model that accounts for online information and strict capacity constraints. A heuristic is proposed to solve the problem, which first applies an unconstrained transit assignment procedure and then handles only the over-loaded transit line segments, re-assigning the surplus passengers. The developed procedure is efficient, requiring very short running times compared to existing capacity constrained transit assignment models. The model assumes passengers receive online information of both predicted arrival time and occupancy condition. Two cases of occupancy information are considered: (1) passengers are informed of the vehicle occupancy, and may change their route selection accordingly, (2) passengers have no occupancy information and in case their boarding is denied, they are enforced to choose a later departing alternative. The model is applied for the Winnipeg network, and as expected the inclusion of capacity constraints increases the average travel time compared to the unconstrained model. However, prior knowledge of the occupancy condition was found to reduce the additional travel time. This result emphasizes the potential benefits of providing occupancy information to passengers.

Active modes take up an increasingly important place on the global policy-making agenda. In the Netherlands, a country that is well-known for its high shares of walking and cycling, the government aims at achieving a modal shift among 200,000 commuting car drivers towards using the bicycle. To this end, policy measures need to be introduced. When the aim is to achieve a modal switch over an enduring period of time, it is more relevant to know the likelihood of including or excluding a mode in the mode choice set, compared to choosing a mode for a single trip. Therefore, we investigate the formation of the experienced choice set (set of modes used over a long period of time), where the aim is to identify determinants that influence the inclusion or exclusion of a mode in this set. We estimate discrete choice models, based on survey data from the Netherlands Mobility Panel (MPN) and a complementary survey, where individuals were asked to report the frequency of using certain modes of transport for commuting trips over the course of half a year. This study shows that the experienced choice set for commuting is unimodal for the majority of the individuals, and remains constant over time for most individuals. Reimbursement by the employer for using a certain mode is the most important determinant influencing the experienced choice set, followed by ownership characteristics and urban density. We show that the mode choice set formation depends on more determinants than previously assumed.

This paper develops an efficient heuristic for the transit network design problem, formulated as an integer programming problem. The model includes a preliminary step of route set generation, followed by an iterative procedure that simultaneously selects the best routes and corresponding headways. The iterative procedure is performed by an infeasible start algorithm that first assigns all candidate routes with the maximal frequency, and then iteratively eliminates routes and decreases frequencies of the less attractive ones. Routes are evaluated through a frequency-based transit assignment model that considers online information. The proposed model is applied to the Winnipeg transit network. The transit network found by the suggested model comprised fewer, faster and more frequent lines that serve high volume of passengers, compared to the given transit network. The running time of the algorithm is very short compared to existing methods, and its simplicity enables high level and detailed calibration.

This paper presents the concept of a modular personalized recommendation system, based on existing social networks. The system offers the users in the affected area to evacuate to their nearest best friend or family member who is outside the danger zone. The output of the system is a near real-time OD matrix, which constitutes an input for transportation evacuation models. A preliminary test case shows success in implementing the system on anonymous real location-based social network data.

This paper develops an optimization model for planning an artificial island composed of Very Large Floating Structures (VLFSs). The optimization model addresses the geometric shape of the island (the array of floating platforms), the land-use layout, and the transportation network. The model developed in this paper considers the specific properties related to an artificial island made of multiple floating modules. The model is formulated as a bi-level optimization problem. In the upper level problem, the decision-maker is required to decide the land-use layout and the road network structure. The user behavior pattern is obtained by solving the lower level problem, which is dependent on the decisions taken at the upper level. Given the complexity of the problem, the paper develops modifications to the genetic algorithm to find the Pareto front in a reasonable amount of time. A test case illustrates the feasibility of the proposed planning method. The transportation model execution is the most time-consuming task, which highlights the need to prune irrelevant scenarios. The Pareto front shows a substitution pattern between different land-use configurations, in accordance with the different objective functions.

This paper aims to demonstrate that advanced technique of modelling may provide insights and improve our understanding of driver behavior in risky decision-making situations. The paper introduces a Hybrid choice model in order to explain the overtaking decision on two-lane highways, which is well known as a risky decision in the safety literature. This model integrates a latent variable model and an overtaking choice model by combining their measurement and structural equations. Specifically, the paper investigates the role of four personality latent variables: Thrill and Adventure Seeking, Boredom Susceptibility, Geographic Ability, and Driving Anger. Respondents to a web-based survey ranked their likelihood to overtake on two-lane highways; two scenarios were captured via short videos: the first presenting a straight section of a road with good visibility, and the second approaching a curve with reduced visibility. Several indicators were collected via self-reported questionnaire. Results indicate that, two out of the four personality latent variables investigated, Thrill and Adventure Seeking and Geographic Ability provide significant explanation for overtaking decision. Both of them are positively correlated with higher risky overtaking behavior. The Hybrid model, by considering latent variables alongside observable variables and attributes of the decision, enhances the comprehension of overtaking behaviour, and therefore may be deployed for explaining other decisions related to risky driving behaviour.

This article examines the relationship between travel speeds and crashes on two-lane highways, accounting for traffic exposure and road infrastructure characteristics. The study's database included 179 road sections in Israel, which included free-flow travel speeds, 3-year injury crash data, traffic volumes, and road infrastructure characteristics. Preliminary analyses of in-data correlations supported the selection of appropriate speed and infrastructure indicators. Homogeneous groups of road sections were identified according to their characteristics. Negative binomial statistical models were fitted to injury crash counts for day and night hours, using speed indicators, section length, traffic volume, and the homogeneous road groups, which reflected various road design conditions. The models demonstrated a positive relation between mean speeds and crashes, while controlling for traffic and road characteristics. The expected crash change following higher travel speeds was more substantial for night hours. In line with previous research, section length, traffic volume, and worse road design were positively related to crashes.

Disasters and extreme events, both natural and man-made, can have dramatic implications in terms of loss of human lives, well-being and economic costs. Understanding the demand for travel during disaster events, in particular when evacuations take place, is critical for the efficient management of the event. The usual activity and travel patterns may be completely broken and not at all relevant during an event, when completely different considerations take priority. A large-scale wildfire took place in Haifa on November 24, 2016. On that day, starting at around 10 a.m., a series of wildfires occurred in the city. As a result, about 40,000 inhabitants (15% of Haifa's population) were evacuated. Shortly after the fire events, a web survey was developed and administered in order to collect data on the activities that residents of the affected areas undertook on that day. This paper presents analysis of this data to evaluate the choices of individuals whether to evacuate or not, the main factors that affect these decisions and related choices.

Drivers’ speed has significant implications on road users’ safety in general, and particularly so if a crash occurs. This paper explores the influence of environmental and road characteristics, situational factors, and individual characteristics on drivers’ observed speed selection in a simulator experiment. The paper presents a theoretical framework for drivers’ speed selection, and applies structural equation modeling for the various factors examined. The simulator experiments collected data of 111 drivers driving in 4 different scenarios composed of 22 segments for each scenario. The dataset was analyzed in several resolutions: Driver level, Trip level, and Segment level. The three models revealed that gender, age, and driving frequency are all significant in determining drivers’ perceptions and attitudes, which in turn influence speed selection. Situational factors such as traffic speed, enforcement, and time-saving-benefits are also related to speed selection, as well as infrastructure characteristics. These findings demonstrate that structural equations provide a flexible modeling tool able to concurrently analyze the variety of factors that relate to speed selection. As a result, Structural Equations Modeling provides more accurate and refined explanations for the combined effects of various factors on drivers’ speed selection than previous research so far. These tools can be useful in developing speed management strategies to improve road safety.

There is an increasing interest in technology-based solutions that can assist drivers in reducing their risk of involvement in road crashes. Previous studies showed that driving events produced by in-vehicle data recorders (IVDR) are applicable for identification of unsafe driving patterns, while combined examinations of driving events and road infrastructure characteristics are rare. This study explored the relationship between the IVDR-driving events, road characteristics and crashes, to examine a potential of the events for predicting crashes and identification of high-risk locations on the road network. The study database included 3500 segments of the interurban roads in Israel, for which the automatically produced IVDR events were matched with road infrastructure characteristics and crashes. Negative-binomial regression models were adjusted for the relationships between road characteristics and driving events, and subsequently, between events and crashes, given the exposure. Significant impacts were found, yet various event types showed different relations to the infrastructure characteristics and different effects on crashes, on various road types. Better road conditions were associated with a decrease in “braking” events and an increase in the “speed alert” events, where road layout constraints and junction proximity were associated with an opposite effect on events. “Braking” and total events showed better potential for predicting crashes on single-carriageway roads, with a positive link to crashes, where for other road types the “speed alert” events were stronger related to crashes, but with a negative link. The heterogeneity of findings indicates a need in further research of the above relationship, with a particular focus on definitions of driving events produced by the IVDR or other technologies.

Drivers' speed selection has been a great interest to road safety researchers. In this study, a driving simulator (STISIM) was used to explore drivers' speed choices, and how they are influenced by infrastructure, traffic, risk/benefit, and driver characteristics. Drivers also filled out a stated-preference survey, which included speed selection items and demographic characteristics. The experimental design included four scenarios of a specific risk or benefit to the driver: A daily trip, a higher speed enforcement scenario, a higher crash-risk scenario, and a scenario with high time pressure. The driver sample included 111 drivers from different ages between 20 and 65, with 44% women. The database included 9,768 observations. The largest effect on drivers' speed was under time-saving benefits with an average increase of 10 km/hr. Infrastructure effects included horizontal curves, longitudinal slope changes, and design speed. Age and gender also influenced speed selection. Thus, the most effective measures for speed reduction may be (1) enforcement, (2) design speed, and (3) horizontal curves, in contrast to (1) time-saving benefits and (2) a high average speed of close-by vehicles, which motivated drivers to increase their speed but may be reduced by counteracting policies.

This paper develops a frequency based transit assignment model considering that online information of predicted arrival times is available to passengers. The methodology is developed for two information levels: (1) full, where the arrival times are available for all intermediate stops in the candidate paths, and (2) partial, where the arrival times are available at the boarding stop only. Passengers are assumed to consider the estimated arrival times together with the expected travel time when choosing their path. The assignment procedure includes the finding of attractive paths, setting of route choice decision rules for different cases of predicted arrival times, and the probability calculation for these different cases. The developed model is illustrated by an application for the Winnipeg network. In comparison to the well-known optimal strategies method, the suggested model produced significantly different assignment results and a notable reduction in the total travel time. The results illustrate the potential impact of online information on assignment results, and emphasize the need for its consideration in planning models.

This paper analyzes the strategic transit network plan for the Tel Aviv metropolitan area, using graph theory and other recently developed transit network measures. The different transit modes included in the strategic plan are emphasized by adding weights to distinguish metro lines from light rail lines (LRT). This approach can help compare the combined metro or LRT alternatives of the new Tel Aviv plan to the metro-only alternatives as well as measure the performance relative to the metro systems in other cities around the world. The analysis of the alternative plans in Tel Aviv showed that when metro and LRT lines were treated as homogeneous modes, in which all were considered as metro, the alternatives resembled medium developed metro systems, such as in Barcelona and Washington DC. In contrast, when the distinguished weights were included, the combined metro/LRT alternatives resembled less developed systems, such as in Lyon and Lisbon, and only the metro-only alterative score remained high. The results also showed that the alternatives have regional coverage, and the alternatives with more LRT lines score lower in coverage. The network structure analysis showed that the metro-oriented networks score higher in both directness and connectivity. When using the weighted measures, the existing plan (LRT-only) scores low on both directness and connectivity. The analysis of the results emphasizes the need for more metro lines in the Tel Aviv metropolitan area. The results also suggest that the analysis of the complex mass transit networks based on graph theory should consider differences in line technology reflected in the line speed and coverage.

Smartphone-based travel surveys have attracted much attention recently, for their potential to improve data quality and response rate. One of the first such survey systems, Future Mobility Sensing (FMS), leverages sensors on smartphones, and machine learning techniques to collect detailed personal travel data. The main purpose of this research is to compare data collected by FMS and traditional methods, and study the implications of using FMS data for travel behavior modeling. Since its initial field test in Singapore, FMS has been used in several large-scale household travel surveys, including one in Tel Aviv, Israel. We present comparative analyses that make use of the rich datasets from Singapore and Tel Aviv, focusing on three main aspects: (1) richness in activity behaviors observed, (2) completeness of travel and activity data, and (3) data accuracy. Results show that FMS has clear advantages over traditional travel surveys: it has higher resolution and better accuracy of times, locations, and paths; FMS represents out-of-work and leisure activities well; and reveals large variability in day-to-day activity pattern, which is inadequately captured in a one-day snapshot in typical traditional surveys. FMS also captures travel and activities that tend to be under-reported in traditional surveys such as multiple stops in a tour and work-based sub-tours. These richer and more complete and accurate data can improve future activity-based modeling.

This paper considers bicycle route choice for commuter trips. Bicycle route preferences are analysed using a dataset from a GPS-assisted household travel survey conducted in the Tel Aviv metropolitan area. Different choice set generation methods were applied to generate alternative routes for each observation, and the matching with the actual route is discussed. Model estimation is performed for different route choice sets to test the sensitivity of the parameter estimates. The results obtained are quite consistent, and indicate an expected tendency to ride in longer routes, but with separated bike lanes. In the absence of such lanes, riders prefer to use local streets and avoid riding on busy arterial streets and highways.

The identification of critical links in a network has been studied in the literature, and different approaches were proposed. Most methods that analyzed critical links were based on static traffic assignment. The purpose of this paper is to find critical links in a transportation network, using existing methods in the literature. The novelty of the paper is by comparing two different methods to detect critical links. The first method will apply dynamic traffic assignment and the second will apply betweenness centrality. The methods were applied to a real network, and the results indicate that the BC measure can be used as a proxy for detection of critical network links.

This paper discusses the network design problem (NDP), solved for travel time minimization and road safety maximization. Similarly to other NDP models, the model is formulated as a bi-level multi-objective optimization problem. The estimation of the system time benefit and safety benefit is performed using previously developed methods. Two different approaches are applied and discussed for solving the problem: the constrained multi-objective optimization and the multi-objective genetic algorithm. Both approaches are tested on a real-size network, using a large set of candidate projects. The paper presents a discussion regarding the different optimal solutions obtained, emphasizing the relative attractiveness of the candidate projects.

This paper addresses the discrete network design problem (DNDP) with emphasis on the environmental benefits. These benefits are traditionally quantified by emission models, which in general account for vehicle speeds, traffic flows and emission coefficients. An alternative approach for approximating the environmental impact of traffic is developed. This approach finds the route that keeps the most balanced speed profile throughout the route, which contributes to fuel consumption reduction. The paper formulates an optimization problem that includes the described approach for the DNDP. The solution of the problem consists of projects that contribute the most to the generation of such “balanced speed routes”. The paper illustrates the problem and the solution for a real-size network with a medium-size set of candidate projects.

A macroscopic model is presented that simultaneously estimates route flows and trip matrices for congested road networks using data on link densities instead of link flows. The advantage of this approach is that it avoids errors that may occur in the individual links’ flow-cost relationships when congestion is heavy. Under the proposed methodology, both the flows and the matrices are estimated by the model using an image of the network such as an aerial photograph in which the number of vehicles on each link can be identified. The model itself is formulated as a maximum entropy optimization problem subject to linear constraints given by vehicle densities on the links, and is validated using analytic examples and traffic microsimulations. The results demonstrate the superiority of the link-density approach over the traditional flow-based method.

Choice situations with variable supply characteristics are found in many applications, including airline itinerary selection. This paper discusses the airline itinerary choice problem in dynamic supply settings. The paper develops a specially designed stated preference (SP) survey, which emulates an air travel website. The survey includes the option to delay the decision to choose an airline itinerary. The rich data set allows the estimation of discrete choice models of airline itinerary choice. The paper presents selected model estimation and application results for two market sectors (tourists and business travelers) and two flight types (medium-haul and long-haul flights). In addition to expected results for several level-of-service variables such as flight cost, cancellation fees, connection times and punctuality percentages, the model estimates the expected value of delaying a flight purchase.

This paper develops a route generation model for the transit network design problem. The generation of a large candidate route set is a preliminary step in the optimization of network design, which has a significant influence on its quality. Existing methods for route generation are widely based on variations of shortest path algorithms, generating fast routes between high demand centers. The idea of the proposed model is to form routes on corridor links, with high overall demand, in order to strengthen the service attractiveness. The suggested model form two types of routes: the commonly used shortest path and path that pass through high demand corridors. The model is compared to a shortest path method, and show a higher coverage of the network, producing a wider variety of routes, including links that combine multiple demands.

This paper discusses the Network Design Problem (NDP) solved for travel time minimization and road safety maximization. Similarly to other NDP models, the model is formulated as a bi-level multi-objective optimization problem. The estimation of the system time benefit and safety benefit is performed using previously developed methods, and the optimal solutions are found using the multi-objective genetic algorithm. The proposed method is demonstrated on a real-size network, using a large set of candidate projects.

This study explored the relationship between travel speeds and accidents on single-carriageway roads, accounting for traffic exposure and road infrastructure characteristics. The speed data are free-flow travel speeds collected by GPS devices inside the vehicles. The study's database included 179 sections, in Israel. Negative binomial statistical models were fitted to injury accident counts, in day- and night-hours, using speed indicators, section length, traffic volume and homogeneous road groups, where road groups reflect various road design conditions. The models demonstrated a positive relation between mean speeds and accidents, while controlling for traffic exposure and road infrastructure characteristics.

This paper develops a frequency based transit assignment model, assuming that transit arrival times are available to the passenger at the boarding stop. The model considers the estimated arrival times together with the expected travel times, when selecting a path. The assignment procedure includes the setting of decision rules for different cases of arrival times, and the calculation of probabilities for these different cases. The model is applied on a real-size network, and the results are compared to the well-known optimal strategies method. The suggested model showed significantly different results and a notable reduction in the total travel time. The results illustrate the potential impact of online information on travel behavior, and emphasize the need of its consideration in assignment models.

This paper develops a combined mode choice and traffic assignment model that incorporates ridesharing as an option in a mode choice model, attempting to quantify the ridesharing market share in an equilibrium context. The mode choice model takes into account that the waiting time for a ride is dependent on the available drivers. The traffic assignment model is a static user equilibrium that interacts with the discrete choice model through level of service variables. An iterative algorithm was implemented and applied in a simple network and a more realistic network. The results indicate that the quantity of ride sharing drivers is a key parameter to the service success, and below a critical mass of drivers, it is unlikely that passengers will find the service valuable. It is also shown that ride sharing has the ability to reduce in-vehicle times for all the users, although passenger may suffer from longer door-to-door times, having to wait for their ride.

Most air quality models use traffic-related variables as an input. Previous studies estimated nearby vehicular activity through sporadic traffic counts or via traffic assignment models. Both methods have previously produced poor or no data for nights, weekends and holidays. Emerging technologies allow the estimation of traffic through passive monitoring of location-aware devices. Examples of such devices are GPS transceivers installed in vehicles. In this work, we studied traffic volumes that were derived from such data. Additionally, we used these data for estimating ambient nitrogen dioxide concentrations, using a non-linear optimisation model that includes basic dispersion properties. The GPS-derived data show great potential for use as a proxy for pollutant emissions from motor-vehicles.

The problem of selecting the optimal set of transportation projects out of a given set of projects, known as the network design problem (NDP), has been researched for many years. Typical transportation projects are interdependent in their nature, which turns the problem into a very complex one. When a certain objective is sought, an exact solution of the problem can be derived only by enumerating each possible project combination. Therefore, when a large set of possible combinations is involved an alternative approach must be taken. Meta-heuristic methods usually used for this purpose do not make use of the special properties of the given problem. This paper proposes an alternative heuristic that simplifies significantly the solution process. The benefit of a certain combination of projects is inferred based on a subset (pairs or triplets) of projects. The proposed heuristic is tested on simple networks and applied for a real-size network. The paper also discusses the trade-offs between solution accuracy and computation time.

Road design characteristics should deliver a clear message to drivers on the appropriate speeds. This approach is known in the literature as "self-explaining roads" (SER). In Israel, new guidelines for setting speeds on the road network were introduced aiming at a balance between the target and actual travel speeds on various road types and, thus, supporting the SER concept. However, engineering tools are needed to implement the new approach. The purpose of this study was to explore the relationship between travel speeds and road infrastructure characteristics, on single-carriageway roads, aiming to identify the features influencing the travel speeds' selection by Israeli drivers. Statistical relations between road characteristics and speed indicators were explored using multivariate classification methods and regression models. The design speeds were reconstructed based on the infrastructure characteristics, and analyzed in relation to the travel speeds. Among the road characteristics most influencing the travel speeds were found: shoulder width and the recovery-zone width, where lower values of both characteristics were associated with lower speeds; junction density and road curvature, where higher presence of these characteristics had a moderating effect on speeds. It was concluded that changing shoulder width, recovery-zone width or junction density may be applied for promoting the SER concept and may affect travel speeds, yet a fine-tuning of existing design guidelines is required.

Earthquakes are sudden, cause huge damage to extensive areas, and may negatively affect the lives of millions of people. Thus, it is crucial to develop a system that can help people survive an earthquake and recover from its aftereffects. In this paper we present a vision of a smartphone app, called EAGA (earthquake alerter and guidance app), that will guide both victims and rescue workers, by leveraging probabilistic geosocial information collected before the event, during the earthquake and in its aftermath. The app has four modes of operation. In standby mode, EAGA collects data about users, their regularly visited locations and their social relations. In alert mode, EAGA warns of an impending earthquake, based on data collected from a variety of sensors and from government warning systems. It also provides initial guidance to the user during the onset of the quake. In disaster mode, EAGA assists users to cope with the tasks of the immediate aftermath, such as evacuation and rescue of victims buried underneath rubble of collapsed buildings. Finally, in recovery mode, EAGA facilitates family reunification, provides information about aid centers and sends warnings regarding the spread of diseases. We believe that EAGA can revolutionize disaster management, and complements current earthquake readiness efforts, by allowing a large degree of much needed decentralization and personalization in dealing with earthquakes.

The diversity of drivers’ speed selection in free-flow conditions has been assumed to originate from various human factors, mainly differences in driver characteristics and preferences. This study uses a stated-preference web-based survey with a sample of 290 participants to investigate the diversity of speed selection in relations to driver characteristics. The survey included newly developed scales of estimating driving risks and estimating personal difficulty of performing vehicle-related technical tasks. Also included were items on performing spatial tasks and drivers’ own self-assessments. The analysis of the survey results revealed that newly developed latent driver characteristics, such as risk awareness and technical aversion, were found to strongly affect individual drivers’ speed selection in a daily trip–daily speed selection. The perceived speed of the average driver, or average driver perception, also had a significant effect on drivers’ own speed selection. In addition, some latent characteristics were found to have stronger effects in certain demographic groups. Implications for further speed-selection researches and road safety policies are discussed.

This paper presents methods for inferring selected travel patterns using passive cell-phone technology. The paper handles two issues that are related to the cellular phone technology: ‘zigzagging’ patterns that do not represent a movement, and track recording of the closest antenna location that serves the cell-phone, which means that the information of the location of the cell-phone itself is not accurate. The home and commuter location of each user is validated at the aggregate district level. The paper shows selected results that can be used for transportation analysis, including a comparison of the results to known models from preceding years.

Most metropolitan areas suffer from traffic-related air pollution. A major reason for this phenomenon is related to the increase in vehicle-kilometers traveled, which is an outcome of urban sprawl and an increase in the motorization rate. Although significant progress in vehicle technology has greatly improved air quality in developed countries, there is still a lack of policy mechanisms to mitigate air quality impacts resulting from traffic pollution. This paper examines policy measures that can be implemented by decision-makers in order to improve urban air quality. Focusing on parking policy, this paper attempts to gain an understanding of the relationship between parking-policy enforcement in the Central Business District (CBD) and air-pollution emissions from motor vehicles. The site chosen for study is the Tel Aviv Metropolitan Area (TAMA). The methodology presented is applied to a real-world situation, using data and models from a Mass Transit project initiated by TAMA. The traffic-related emissions are estimated from a four-stage transportation model that provides traffic volumes and travel times for each link in the network. Two policy measures are analyzed: reducing parking supply and increasing parking fees. The paper discusses the suitability of parking policies to meet environmental objectives.

Latent class models (LCMs) can yield powerful improvements in understanding the travel behaviour over traditional approaches. All the LCMs studies in transportation used discrete choice models for both the choice model and the identification of segment membership. This paper introduces an innovative segmentation methodology for the segment (class) identification model. The method includes a fuzzy segmentation process, which takes into account the varying levels of influence of each attribute on the degree of association with a segment. Five mode choice models were estimated using a data set from a household survey: a multinomial logit model, a nested logit (NL) model, a traditional LCM, a LCM using new segment identification, and a mixed NL. The estimation results indicate that the new segmentation method used for LCM captures heterogeneity differently than the traditional models, with similar likelihood estimates and good prediction results.

Artificial Neural Networks (ANN), which are used in many different areas, have been applied to predict the vertical swelling percentage of expansive clays. In contrast to previous models that estimated ANN Models in a single phase, this paper proposes an alternative analysis in based on the following two-stage operation: (a) conducting an ANN analysis on the swelling-pressure test results (i.e., the ASTM 4546 Method C test results) to obtain the swelling- pressure model for any given clay characteristics, and (b) performing an additional ANN analysis on the swelling-percentage test results (i.e., the ASTM 4546 Method B test results), including the former ones, with the given independent variables of the clay characteristics. This second stage includes a defined expression containing the given surcharge pressure and the predicted value of the swelling pressure as obtained from the model of the previous stage. Two final ANN Models, each with a different arrangement of the given independent variables, were derived from this two-stage procedure. Their statistical fit was clearly found to be superior in comparison to previous models estimated with the same data set. Furthermore, one of these two models exhibited the expected geophysical behavior. As this new ANN Model yields higher predicted swelling-percentage values, it can definitely be regarded as a preferable one in the sense of enlarging the safety margin in heave calculations.

The purpose of this paper is to demonstrate that latent variables, with the focus on sensation seeking concepts, incorporated in new technique of route choice modeling, improve our analyzing of route choice behavior with pre-trip travel time information. The application of a hybrid discrete choice model framework integrates a latent variable model and a route choice model by combining their measurement and structural equations. The model is estimated based on data from a laboratory experiment and a field study of a simple network. The results show that certain sensation seeking domains (e.g., thrill and adventure seeking) alongside traditional variables (e.g., travel time information) enrich our understanding and provide more insight into route choice behavior. Furthermore, observed personal variables, such as gender and marital status, may serve as causal indicators to sensation seeking variables.

The objective of this study was to investigate the properties of the flexibility model in a network of freeways that conveys high volumes of traffic. A dynamic system flexibility model is defined, and its three major components are evaluated and discussed. The dynamic system flexibility model measures and estimates the optioas of drivers traveling between origin-destination pairs in a freeway network. The three models' components depend on the number of possible and feasible routes between a given origin-destination pair, on the common lengths of the possible routes, and on the amount of variability between the length of each route und the length of the shortest route. The lengths of the link in the model are perceived lengths and depend on the occurrence of congestion in them. The flexibility measure provides a good estimate of the number of options available to drivers in a given network. Moreover, this measure is sensitive to the amount of congestion in the freeway system, and therefore the model the is proposed Is flow dependent. Although the common lengths between the routes increase, the flexibility decreases. A pseudoparadox, similar to the Braess paradox, is shown to exist under certain network cnnditioas.

The Tel Aviv activity based model structure is similar to other activity based models described in the literature. The model run is supposed to converge to the equilibrium between generated tours and corresponding level of service (LOS) data. However, individual tour generation uses random draws for various choices (activity, time of day, destination, and mode). This introduces simulation errors, which combined with population sampling and limited precision of static traffic assignments, prevents the convergence of the model results. This paper analyses the above uncertainty sources on the basis of multiple model runs conducted for this study. Three averaging procedures are investigated and compared. Practical considerations regarding setting up the averaging procedures required for obtaining stable model results are discussed.

Regular use of artificial neural networks (ANN) analysis for predicting the vertical swelling percentage of expansive clays may lead to inappropriate results in terms of their geophysical behavior. This paper presents two new ANN Models derived from a two-stage procedure. The models were estimated using the same data set from the previous paper, and their statistical fit was clearly found to be superior in comparison to the previous models. Furthermore, one of these two models exhibited the expected geophysical behavior. As this new ANN Model yields higher predicted swelling-percentage values, it can definitely be regarded as a preferable one in the sense of enlarging the safety margin in heave calculations.

Spatially detailed estimation of exposure to air pollutants in the urban environment is needed for many air pollution epidemiological studies. To benefit studies of acute effects of air pollution such exposure maps are required at high temporal resolution. This study introduces nonlinear optimisation framework that produces high resolution spatiotemporal exposure maps. An extensive traffic model output, serving as proxy for traffic emissions, is fitted via a nonlinear model embodying basic dispersion properties, to high temporal resolution routine observations of traffic-related air pollutant. An optimisation problem is formulated and solved at each time point to recover the unknown model parameters. These parameters are then used to produce a detailed concentration map of the pollutant for the whole area covered by the traffic model. Repeating the process for multiple time points results in the spatiotemporal concentration field. The exposure at any location and for any span of time can then be computed by temporal integration of the concentration time series at selected receptor locations for the durations of desired periods. The methodology is demonstrated for NO₂ exposure using the output of a traffic model for the greater Tel Aviv area, Israel, and the half-hourly monitoring and meteorological data from the local air quality network. A leave-one-out cross-validation resulted in simulated half-hourly concentrations that are almost unbiased compared to the observations, with a mean error (ME) of 5.2ppb, normalised mean error (NME) of 32%, 78% of the simulated values are within a factor of two (FAC2) of the observations, and the coefficient of determination (R²) is 0.6. The whole study period integrated exposure estimations are also unbiased compared with their corresponding observations, with ME of 2.5ppb, NME of 18%, FAC2 of 100% and R² that equals 0.62.

Long-distance trips are generally under-reported in typical household surveys, because of relative low frequency of these trips. This paper proposes to utilize location data from cellular phone systems in order to study long-distance travel patterns. The proposed approach allows passive data collection on many travelers over a long period of time at low costs. The paper presents the results of a study that applies cellular phone technology to assess trips at the national level. The method was specifically designed to capture long distance trips, as part of the development of a national demand model conducted for the Economics and Planning Department of the Israel Ministry of Transport. The method allows the construction of origin-destination tables directly from the cellular phone positions. The paper presents selected results to illustrate the potential of the method for transportation planning and analysis.

Multi-dimensional discrete choice problems are usually estimated by assuming a single-choice hierarchical order for the entire study population or for pre-defined segments representing the behavior of an "average" person and by indicating either limited differences or a variety in choices among the study population. This study develops an integral methodological framework, termed the flexible model structure (FMS), which enhances the application of the discrete choice model by developing an optimization algorithm that segment given data and searches for the best model structure for each segment simultaneously. The approach is demonstrated here through three models that conceptualize the multi-dimensional discrete choice problem. The first two are Nested Logit models with a two-choice dimension of destination and mode; they represent the estimation of a fixed-structure model using pre-segmented data as is mostly common in multi-dimensional discrete choice model implementation. The third model, the FMS, includes a fuzzy segmentation method with weighted variables, as well as a combination of more than one model structure estimated simultaneously. The FMS model significantly improves estimation results, using fewer variables than do segmented NL models, thus supporting the hypothesis that different model structures may best describe the behavior of different groups of people in multi-dimensional choice models. The implementation of FMS involves presenting the travel behavior of an individual as a mix of travel behaviors represented by a number of segments. The choice model for each segment comprises a combination of different choice model structures. The FMS model thus breaks the consensus that an individual belongs to only one segment and that a segment can take only one structure.

The concept of flexibility is very common in many engineering aspects but less common in transportation planning. This study develops a dynamic model of system flexibility for freeway networks. When comparing different freeway networks, it is essential to have a measure defining a network's flexibility based not only on its topology but also on the amount of traffic compared to each segment's capacity. Application of the proposed flexibility model could be helpful in estimating the perceived amount of total system congestion in a given network at any given time and how this congestion may vary if a new link is added to the system. The proposed Dynamic Flexibility Model is based on three variables: the number of routes between each origin-destination (O-D) pair in the freeway network, the amount of length in common in each O-D pair, and the amount of variability among the various routes in each O-D pair relative to the shortest path in the same O-D. The traffic flow is included in the model via the perceived length: the relevant length is equal to its actual length in an uncongested state and increases during periods of congestion, since drivers then experience longer travel times, which are converted into an equivalent increase in the perceived-segment length. In this work, the onset of congestion that was adapted was at the Critical Occupancy Point (COP), which is an objective measure evaluating whether or not a freeway segment is congested. This paper shows that the proposed Dynamic Flexibility Model is able to compare different freeway networks, thus adding a new dimension for measuring drivers' flexibility in choosing other routes in case some links suffer congestion and breakdown. Furthermore, the proposed model helps to define the perceived total congestion in the system.

The general swelling model has recently been updated in Israel by applying the Excel-solver command (ESC) analysis to new local test results from 897 undisturbed specimens. In this analysis, the goodness-of-fit statistics obtained classify the category of their associated regression only as fair. Thus, it seems necessary to explore the possibility of enhancing the outputs of this regression analysis by applying the artificial neural networks (ANN) methodology to the same 897 undisturbed specimens. However, it is shown that the use of the ANN outputs should be accompanied by an additional check to ensure that they follow the expected physical swelling behavior, as characterized by the index properties of the soil. The ANN methodology applied in this paper is similar to previous studies in geotechnical engineering. Different models were tested using the same database (i.e., the same 897 undisturbed specimens). The statistical fit of the ANN models were clearly found to be superior to the ESC models. However, in the sense of the required physical behavior, as characterized by the index properties of the soil, the ANN models did not predict swelling values as well as ESC models did, in particular values ranging near (or outside) the data set boundaries. Thus, the former ESC models still remain preferable.

Network planning and traffic flow optimization require the acquisition and analysis of large quantities of data such as the network topology, its traffic flow data, vehicle fleet composition, emission measurements and so on. Data acquisition is an expensive process that involves household surveys and automatic as well as semiautomatic measurements performed all over the network. For example, in order to accurately estimate the effect of a certain network change on the total emissions produced by vehicles in the network, assessment of the vehicle fleet composition for each origin-destination pair is required. As a result, problems that optimize nonlocal merit functions become highly difficult to solve. One such problem is finding the optimal deployment of traffic monitoring units. In this article we suggest a new traffic assignment model that is based on the concept of shortest path betweenness centrality measure, borrowed from the domain of complex network analysis. We show how betweenness can be augmented in order to solve the traffic assignment problem given an arbitrary travel cost definition. The proposed traffic assignment model is evaluated using a high-resolution Israeli transportation data set derived from the analysis of cellular phones data. The group variant of the augmented betweenness centrality is then used to optimize the locations of traffic monitoring units, hence reducing the cost and increasing the effectiveness of traffic monitoring.

Among the main factors affecting road crash injuries, speed is considered as a leading cause and contributing factor. There are numerous studies linking travel speeds and road crashes. Hence, an essential part of road safety plans and interventions is devoted to speed management. However, to manage speed, actual travel speeds have to be systematically and consistently monitored and analyzed. In this study, a system for the collection and analysis of free-flow travel speeds on the road network is presented, enabling speed monitoring at thenationwide level. The paper focuses on the collection and analysis of travel speeds on different road sections. Using the information gathered through advanced technologies combined with geographical information systems, a comprehensive speed database in space and time is provided allowing visual presentation and comparison of the results. This analysis can identify the road sections with significant excesses of travel speeds relative to the speed limits. It can also serve as a baseline to evaluate the impact of various counter-measures employed to reduce speeds.

Purpose: This study examined the national road safety programs carried out by the ten world's leading countries in road safety, aiming to identify the most effective interventions that contributed to safety progress in those countries and to consider the possibilities of their application in Israel. Method: The best-performing countries were selected from European and other developed countries, using general safety indicators and the rate of road safety improvement achieved recently. The program documents and related reports published in the countries were screened aiming to sum up safety problems in the countries examined, safety interventions recommended for implementation by the countries' programs and evidence of the efficiency of those interventions. Results: Detailed classifications of safety problems and measures/interventions implemented by the countries' programs were produced. Evidence of the efficiency of safety interventions was collected in terms of associated accident reductions and the scope of measures' implementation during the program's performance. The study demonstrated a high similarity of main safety problems characteristic for the majority of leading countries, and for Israel. Thus, the summaries of safety interventions adopted by those countries were applicable for addressing similar problems in local conditions. Most safety interventions associated with the countries' safety progress over the last decade came from the well-recognized areas of infrastructure and enforcement, whereas for some common safety problems, e. g. motorcyclist injury, driving fatigue, elderly vulnerability, prominent solutions are lacking. Conclusion: The research findings can serve as a basis for developing a new national program for promoting road safety in a country.

The search for available parking is one of the most challenging consequences of global urbanization and growth in motorization. This paper presents an overall framework for parking choice and search behavior, composed of three time-space phases: (a) pretrip static decision; (b) en route, passive search; and (c) in-area search strategy adaptation. The empirical part of the paper focuses on the first phase and describes a parking choice model that is based on pooled stated and revealed preference data sources. A special, web-based survey was designed to model the choice of parking type (on-street versus off-street parking). The model estimation results showed that the choice of parking location was affected by parking cost, search time, and walk time to the destination, facility type, and decision-maker characteristics. The model was applied to a case study to illustrate its capabilities to evaluate various policy measures. Specifically, the effect of a change in the demand for on-and off-street parking was evaluated with respect to the parking pricing policy and the value of search time for various parking durations.

Head-on collisions on two-lane rural highways might result from failed passing maneuvers. This article investigates the hypothesis that drivers do not always estimate the required passing gap correctly, and their decision to overtake is made under a particular, even if small, probability that a crash will occur; this is the associated risk. The research investigates drivers' irrationality in evaluating the risks of different passing gaps and develops a risk-taking measure. This measure provides a tool for classifying drivers into different groups: risky, partially cautious, and cautious drivers. A comparison of the sociodemographic averages, driving style, and crash history parameters for these groups showed significant differences. The measure developed can be used for risk evaluation and as a measure of safety. It further provides a tool for classifying drivers into groups based on their risk-taking characteristics. This is particularly useful for safety education programs.

In the last decade, a broad array of disciplines has shown a general interest in enhancing discrete choice models by considering the incorporation of psychological factors affecting decision making. This paper provides insight into the comprehension of the determinants of route choice behavior by proposing and estimating a hybrid model that integrates latent variable and route choice models. Data contain information about latent variable indicators and chosen routes of travelers driving regularly from home to work in an urban network. Choice sets include alternative routes generated with a branch and bound algorithm. A hybrid model consists of measurement equations, which relate latent variables to measurement indicators and utilities to choice indicators, and structural equations, which link travelers' observable characteristics to latent variables and explanatory variables to utilities. Estimation results illustrate that considering latent variables (i. e., memory, habit, familiarity, spatial ability, time saving skills) alongside traditional variables (e. g., travel time, distance, congestion level) enriches the comprehension of route choice behavior.

The estimation of semi-compensatory models is gaining momentum in transport planning in recent years. However, traditional survey methodologies focus on collecting solely compensatory choice data, which leads to information loss when semi-compensatory models are estimated. The present study proposes a novel web-based survey that enables collecting data about the entire semi-compensatory choice process. The web-based environment allows seamless tracking of semi-compensatory choice protocols without interfering with the natural choice process and without introducing problems related to comprehension bias, narrative inconsistency and misinterpretation of the choice protocols. The procedure is applied to rental apartment choice by students and results shed light on semi-compensatory choice by: (1) demonstrating the importance of choice set formation; (2) unravelling the distribution of threshold selection across the population; (3) revealing the linkage between the viable choice-set and the choice.

In decisions involving many alternatives, such as residential choice, individuals conduct a two-stage decision process, consisting of eliminating non-viable alternatives and choice from the retained choice set. In light of the potential of semi-compensatory discrete choice models to mathematically represent such decisions, research is inching ahead with the aim of alleviating their high computational complexity and their severe restrictive assumptions. To date, still a major barrier for the implementation of semi-compensatory models is their underlying assumption of independently and identically distributed error terms across alternatives at the choice stage. This study relaxes the assumption by introducing nested substitution patterns and alternatively random taste heterogeneity at the choice stage, thus equating the structural flexibility of semi-compensatory models to their compensatory counterparts. The proposed model is applied to off-campus rental apartment choice by students. Results show the feasibility and importance of introducing a flexible error structure into semi-compensatory models.

Discrete choice models are widely used for travel demand analyses, and several such models were developed based on random utility theory. The estimation process in such models generally uses the entire study population and the same model structure to estimate average parameter values, but this approach limits the possiblility that the model can fully explain the behaviour of different populations. A new approach, the flexible model structure (FMS), expands existing discrete choice models with the addition of two main components: the segmentation process and model structure search. The aim of the FMS process is to guide the progress of model estimation towards a more behaviourally realistic representation. The main contribution of this paper is to suggest and illustrate a framework that simultaneously searches for the best segmentation and the best model structure for each segment. This paper presents a numerical case study that illustrates the FMS concept. The results indicate that (a) the addition of the segmentation process acts to emphasise the heterogeneity that exists among segments and demonstrates the importance of segmentation by significantly improving the estimation results; and (b) the model structure can vary along segments.

This paper develops path-based algorithms to solve the C-logit stochastic user equilibrium (SUE) problem on the basis of an adaptation of the gradient projection method. The algorithms' strategies for step size determination differ. Three strategies are investigated: (a) predetermined step size, (b) Armijo line search, and (c) self-adaptive line search. The algorithms are tested on the well-known Winnipeg (Manitoba, Canada) network. Two sets of experiments are conducted: (a) a computational comparison of different line search strategies and (b) the impact of different modeling specifications for route overlapping (a flow-independent or a flow-dependent commonality factor). The results indicate that the path-based algorithm with the self-adaptive step size strategy performs better than the other step size strategies. The paper shows that, depending on the model parameters, particularly the commonality factor parameter, the C-logit SUE flows may be quite different from the multinomial logit SUE flows.

A static stochastic user equilibrium (SUE) problem was formulated: the mode of random regret minimization (RRM) was used for route choices. The RRM approach assumes that individuals minimize anticipated regret, rather than maximize expected utility, when choosing from alternative routes. The cost function for the RRM model is not separable, and so a variational inequality approach was adopted to formulate the problem. A path-based algorithm was applied to solve the RRM-SUE problem with the method of successive averages. Implementation of the algorithm in a real-world network is illustrated, and the trade-offs and differences between the proposed model and the SUE based on random utility models is discussed.

This article considers the stochastic user equilibrium (SUE) problem with the route choice model based on the C-logit function. The C-logit model has a simple closed-form analytical probability expression and requires relatively lower calibration efforts and represents a more realistic route choice behaviour compared with the multinomial logit model. This article proposes two versions of the C-logit SUE model that captures the route similarity using different attributes in the commonality factors. The two versions differ with respect to the independence assumption between cost and flow. The corresponding stochastic traffic equilibrium models are called the length-based and congestion-based C-logit SUE models, respectively. To formulate the length-based C-logit SUE model, an equivalent mathematical programming formulation is proposed. For the congestion-based C-logit SUE model, we provide two equivalent variational inequality formulations. To solve the proposed formulations, a new self-adaptive gradient projection algorithm is developed. The proposed formulations and new solution algorithm are tested in two well-known networks. Numerical results demonstrate the validity of the formulations and solution algorithm.

The purpose of this study is to develop a new and objective measure to evaluate freeway congestion. Congestion has yet to be consistently defined because it impacts on choices regarding investments in infrastructure. A new measure, termed the Critical Occupancy Point (COP), is developed. COP evaluates the change in occupancy vs. speed and defines the point where congestion begins. COP may be developed from data which is readily available from each freeway sensor. Data analysis from a freeway in Israel shows that each segment provides different congestion characteristics and consequently a different COP. The COP measure identifies recurrent and non-recurrent congestion and the breakdown before it actually occurs. The COP and the speed at which it occurs are found to be distributed normally. Furthermore, it was found that the volume at which the COP happens increases as the distance from the previous entrance increases.

In this research the dynamic choice behavior when choosing airline tickets is investigated. 600 simulated choices over a period of 30 days were created using simulated respondents and real airline itineraries. Estimation is conducted using discrete choice models (e.g., MNL, NL and CNL) with comparison between them. It is assumed that consumers will not choose their itinerary from an initial choice set, but will postpone their decision to a later time (at least one time period) trying to maximize their utilities. In addition, the variables on-time performance (OTP) and frequent flier program (FFP) were estimated. Main finding is that passengers do postpone their choice decisions for an average of 2.5-5.6 days over a choice period of 30 days. In addition, FFP membership and OTP were found to have significant effect on consumers' choices. Although as simulated results should be taken cautiously, these results implies that dynamic behavior do take place and should be further investigated with real respondents.

Semicompensatory models represent a choice process consisting of an elimination-based choice set formation on satisfaction of criterion thresholds and a utility-based choice. Current semicompensatory models assume a purely noncompensatory choice set formation and therefore do not support multinomial criteria that involve trade-offs between attributes at the choice set formation stage. This study proposes a novel behavioral paradigm consisting of a hybrid compensatory–noncompensatory choice set formation process, followed by compensatory choice. The behavioral paradigm is represented by a mathematical model that combines multinomial-response and ordered-response thresholds with a utility-based choice. The proposed model is applied to a stated preference experiment of off-campus rental apartment choices by students. Results demonstrate the applicability and feasibility of incorporating multinomial-response thresholds into semicompensatory models.

This study intends to provide insight into pedestrian accidents by uncovering their patterns in order to design preventive measures and to allocate resources for identified problems. Kohonen neural networks are applied to a database of pedestrian fatal accidents occurred during the four-year period between 2003 and 2006. Results show the existence of five pedestrian accident patterns: (i) elderly pedestrians crossing on crosswalks mostly far from intersections in metropolitan areas; (ii) pedestrians crossing suddenly or from hidden places and colliding with two-wheel vehicles on urban road sections; (iii) male pedestrians crossing at night and being hit by four-wheel vehicles on rural road sections; (iv) young male pedestrians crossing at night wide road sections in both urban and rural areas; (v) children and teenagers crossing road sections in small rural communities. From the perspective of preventive measures, results suggest the necessity of designing education and information campaigns for road users as well as allocating resources for infrastructural interventions and law enforcement in order to address the identified major problems.

This research focused on examining the over-time developments of road safety in Israel, in order to identify the time points at which positive changes occurred, i.e. a decrease in the number of road accident fatalities, injuries or accidents, taking into account the changes observed in exposure. The time points found with a positive change - a decrease in risk level, i.e. the number of fatalities/ injuries/ accidents related to exposure, allowed to identify the safety interventions which were implemented in proximity to these time points and, therefore, probably contributed to the improvement of the level of safety.As a first stage in the research, an extensive mapping of safety interventions in Israel, for the years 1970-2008, was conducted. Among the safety interventions considered were: changes in transportation regulations regarding safety vehicle accessories, restrictions to young drivers, enhancement of the regulations for driving under the influence of alcohol, as well as significant improvements in road infrastructure, large-scale enforcement operations, vehicle improvements, improvements in the rescue services, etc.The examination of road safety developments in Israel was performed by means of fitting statistical models both for general series of the numbers of casualties and for more detailed series such as: fatalities on urban and rural roads, pedestrian fatalities, fatalities and serious casualties among young drivers, fatal and serious single vehicle accidents, fatalities and casualties in light vehicles. As exposure estimators were used: vehicle-kilometers traveled and their substitutes, e.g. the number of registered vehicles in the country, the number of private vehicles and the population size. In total, 45 casualty series with relation to exposure estimators were analyzed.Two methods were applied for fitting the models: (a) describing the over time development of the risk rate (casualties/ accidents divided by exposure), using "broken-line regression" models. Having the models fitted, the meanings of the break-points found were examined; (b) Fitting "structural time-series" models. These models are used to estimate the components of the level and slope of the linear trend local model and model's residuals, while controlling for known explanatory variables (e.g. the exposure). Using these models, forecasts from specific time-points on were produced, and differences between what would be expected to be the development of the series and the actual counts were examined. The time-points for the forecasts were determined based on former knowledge regarding the timing of the introduction of major safety interventions.For most of the positive changes in safety developments in Israel which were identified using the models, relations to the safety interventions implemented in time proximity to the changes were suggested.

This study explored the possibility of integrating an operational activity-based model and a dynamic traffic assignment framework. The Tel Aviv, Israel, activity-based model and parts of the functionality of the MATSim agent-based framework were used in an attempt to draw on the best features of both approaches: the disaggregate demand representation from the activity-based model and the disaggregate supply representation of the agent-based framework. The study used the person-activity schedule produced by the activity-based model directly and thus eliminated the need to aggregate origin-destination matrices. This paper compares results produced by this combination with those of a static assignment of the Tel Aviv model, which showed a good fit at the aggregate level. The purpose of the paper is to advance the fully disaggregate implementation of activity-based models. The paper represents a step toward this general goal.

This article provides a broad picture of fatal traffic accidents in Israel to answer an increasing need of addressing compelling problems, designing preventive measures, and targeting specific population groups with the objective of reducing the number of traffic fatalities. The analysis focuses on 1,793 fatal traffic accidents occurred during the period between 2003 and 2006 and applies Kohonen and feed-forward back-propagation neural networks with the objective of extracting from the data typical patterns and relevant factors. Kohonen neural networks reveal five compelling accident patterns: (1) single-vehicle accidents of young drivers, (2) multiple-vehicle accidents between young drivers, (3) accidents involving motorcyclists or cyclists, (4) accidents where elderly pedestrians crossed in urban areas, and (5) accidents where children and teenagers cross major roads in small urban areas. Feed-forward back-propagation neural networks indicate that sociodemographic characteristics of drivers and victims, accident location, and period of the day are extremely relevant factors. Accident patterns suggest that countermeasures are necessary for identified problems concerning mainly vulnerable road users such as pedestrians, cyclists, motorcyclists and young drivers. A "safe-system" integrating a system approach for the design of countermeasures and a monitoring process of performance indicators might address the priorities highlighted by the neural networks.

The purpose of this study is to evaluate the potential benefits from real-time travel-time information provision by gaining insights into and better understanding of the factors affecting the route-choice behaviour of drivers possessing real-time pre-trip information. Both a field study and an in-laboratory experiment were conducted to obtain revealed preferences (RP) and stated preferences (SP) on route choice. The data collected from both experiments are inputs for a combined RP-SP route-choice model. The model also includes attitudinal factors describing drivers'personal characteristics. The results show that when experience seeking increases, individuals tend to prefer a route characterised by lower average but greater variance of travel time. Other results confirmed past studies showing that information has an impact on route-choice behaviour although compliance with received information decreases with recent experience.

This paper presents the development and estimation of a novel semi-compensatory model for residential choice. The model assumes that apartment seekers engage in a two-stage process, consisting of a non-compensatory strategy to retain only alternatives that meet search-criteria thresholds, followed by a compensatory strategy to finalize the choice. The first stage is represented by hierarchically correlated ordered-response models and the second stage by a multinomial logit model. The model estimation on data retrieved from a real-estate website, designed to track two-stage choice protocols, reveals the determinants of threshold selection and apartment choice. Results show that (i) threshold selection depends on individual characteristics, (ii) the proposed model is applicable to choice contexts with many alternatives, making it unique in the literature of semi-compensatory models and especially suitable for modeling residential choice, and (iii) the proposed model outperforms the compensatory model and reveals an upward bias in the parameters of the latter.

This study proposes a two-stage method to elicit consumers' price acceptability range. The method combines a conjunctive stage to elicit price acceptability limits with a utility-based stage to choose a preferred product variation. The method is efficient in choice situations entailing many multi-attribute product variations under partial information conditions. A semi-compensatory model complements the method by jointly representing the conjunctive stage with multiple ordered-response models and the choice stage with a multinomial logit model. A case study of ceiling reservation price (CRP) elicitation for students' rental apartment choice shows (i) CRP distribution for different product variations, (ii) model estimation unraveling CRP determinants, and (iii) linkage between CRP and transaction price.

This paper analyzes the main characteristics of travel behavior by the Arab minority community in Israel and discusses two issues related to household travel surveys: data collection among minorities and under-reporting of mid-day trips. Household travel surveys are generally designed and conducted for the majority population and, therefore, lack a proper accounting of minorities and miss many of their less-frequent trips. An alternative approach to conducting household surveys is presented, with the aim of improving data quality for transportation planning. The survey was designed for and conducted in three Arab towns in Israel. The main improvement of the survey involves better interaction between interviewer and interviewee, which should materialize into a relaxed environment that allows for obtaining detailed, reliable results within a reasonable amount of time. The results of the survey employing the alternative approach were compared to a sub-sample of the same towns taken from a regional survey conducted by the regional planning agency at the same time. The paper presents simple statistics on the main variables for each survey. Significant differences are found in the two data sets, mostly regarding the frequency of less frequent, non-home-based trips. A plausible explanation for these differences relates to the more detailed and improved data collected in the new survey.

To analyse the diversion from auto modes to combined modes such as park and ride, it is common to develop mode choice models based on discrete choice theory. In most cases, park and ride is modelled as an access mode to a main transit mode. This paper proposes an approach to test similarities among modes and the appropriate model structure, providing the flexibility for various model structures. The paper explores the capability of recently developed models by specifying their structure to capture the similarities of the combined modes. The paper presents an example with real data to illustrate the methodology application. Estimation results for different model structures including the Multinomial Logit, Nested Logit, Cross-Nested Logit and the Logit Kernel with all of these previous models as kernel are presented. As expected the best estimation results are obtained for the most flexible model, the Logit Kernel with Cross Nested as Kernel.

Passing maneuver on rural two-lane highways is a complex task, which has a significant effect on capacity, level of service and safety. The maneuver is conditioned on the gap between two successive vehicles on the opposing lane. The minimum time to collision, defined as the remaining gap between the passing vehicle and the oncoming vehicle at the end of the passing process, expresses a measure of the risk involved in the passing maneuver. This paper develops a model that explains the minimum time to collision. The model formulation is based on the analysis of drivers' passing decisions on two-lane rural highways using an interactive driving simulator. The simulator enables the collection of vehicle speeds and positions for different road and traffic scenarios. In addition to the driver simulator, participants responded to a questionnaire which collected information about their socio-demographic characteristics. The composed dataset was analyzed and processed to develop a model that predicts the risk associated with the passing behavior. Tobit regression models were found to be more suitable, in comparison to ordinary least square models and Hazard-based Duration models. The explanatory variables tested represent road geometry, traffic conditions and drivers' characteristics. It was found that while the traffic related variables had the most important effect on the measure of risk chosen, factors related to the geometric design and the driver characteristics also had a significant contribution.

Moving bottlenecks in highway traffic are defined as a situation in which a slow-moving vehicle, be it a truck hauling heavy equipment or an oversized vehicle, or a long convey, disrupts the continuous flow of the general traffic. The effect of moving bottlenecks on traffic flow is an important factor in the evaluation of network performance. This effect, though, cannot be assessed properly by existing transportation tools, especially when the bottleneck travels relatively long distances in the network. This paper develops a dynamic traffic assignment (DTA) model that can evaluate the effects of moving bottlenecks on network performance in terms of both travel times and traveling paths. The model assumes that the characteristics of the moving bottleneck, such as traveling path, physical dimensions, and desired speed, are predefined and, therefore, suitable for planned conveys. The DTA model is based on a mesoscopic simulation network-loading procedure with unique features that allow assessing the special dynamic characteristics of a moving bottleneck. By permitting traffic density and speed to vary along a link, the simulation can capture the queue caused by the moving bottleneck while preserving the causality principles of traffic dynamics.

The search for the shortest path constitutes the common practice in actual traffic studies, as this simplistic route choice model enables the universal implementation of traffic assignment and simulation procedures to every network configuration. The literature illustrates the large efforts in trying to move forward from this simplistic approach, the limited attempts in modeling route choice behavior from revealed preference data, and the nonexistent endeavor in investigating the transferability of more realistic path generation techniques and route choice models. This paper introduces a test to analyze the transferability of path generation techniques that is based on a newly defined efficiency index for the evaluation of their "cost-effectiveness". Then, equality of model estimates is tested to examine the transferability of route choice models, based on a methodology normally used in the estimation of models with mixed data (typically revealed and stated preference data) and on an existing transferability test statistic commonly used in mode choice modeling. Lastly, an experiment is presented to illustrate the implementation of the transferability tests, based on revealed preference data from two different case studies. Experiment results show that path generation techniques are totally transferable at the model specification level and partially transferable at the model parameter level, and that transferability is generally verified when parameters optimized for a larger network are successfully applied to a smaller network. Experiment results also show that not all route choice models are transferable at the model specification level, and none are transferable at the model parameter level.

Understanding the distribution and the determinants of search criteria thresholds helps in representing choice set formation and choice behavior. The development and estimation are presented of a two-stage model that jointly represents search criteria thresholds, within a noncompensatory choice set formation stage, and choice behavior, within a compensatory stage. Data were collected by using a custom-designed web-based choice experiment that seamlessly tracked the entire choice process by recording choice protocols. The model is applied to off-campus rental apartment choice by students as an example of a complex choice situation. The model combines three correlated ordered response models for the search criteria with a multinomial logit model for the choice among a realistically large realm of 200 alternatives. Estimates demonstrate that (a) the thresholds are correlated and their selection can be explained by individual characteristics, preferences, and perceptions and (b) the suggested methodology alleviates the computational complexity embedded in two-stage models by reducing the number of theoretically possible choice sets to the actually chosen ones.

A national transportation planning model is needed for several reasons, including heavy investments in transportation infrastructures and the need to perform formal cost-benefit analyses of all medium- and large-scale transport projects. The paper describes the formulation and development of the planning model. The model is unique for the data collected and is used for analyzing nationwide travel. After careful legal review regarding privacy laws, cellular phone (CP) data were obtained for sixteen 1-week samples of 10,000 phones. In total, data for 1.04 million person-days were obtained. Data records included the unique CP identification, the antenna serving the CP, and a time stamp (date, hour, minute, second). At the minimum, a record was written each time a moving CP changed its connecting antenna. To ensure privacy, neither information nor identification of the cellular phone owner was recorded. The paper describes the structure of the planning model. The CP survey data were used only in the models for constructing person trip tables. To the authors' knowledge, this is the first time that data obtained by wireless location technology (WLT) in large quantities have been used in transportation planning. The paper discusses the advantages and limitations of using WLT and presents directions for further research.

This article investigates the single-class static stochastic user equilibrium (SUE) problem with separable and additive link costs. A SUE assignment based on the Cross-Nested Logit (CNL) route choice model is presented. The CNL model can better represent route choice behavior compared to the Multinomial Logit (MNL) model, while keeping a closed form equation. The article uses a specific optimization formulation developed for the CNL model, and develops a path-based algorithm for the solution of the CNL-SUE problem based on adaptation of the disaggregate simplicial decomposition (DSD) method. The article illustrates the algorithmic implementation on a real size network and discusses the trade-offs between MNL-SUE and CNL-SUE assignment.

Passing manoeuvres on rural two-lane highways significantly affect highway capacity, safety and level of service. This article presents an analysis of data on drivers' passing decisions on two-lane rural highways that were collected with an interactive driving simulator. Measurements of the speeds and positions of all vehicles in several different scenarios were collected and processed to generate observations of gap acceptance behaviour. In addition, participants responded to a questionnaire which collected information on their socio-demographic and driving styles characteristics. These data were utilised to develop a model that explains the decision whether to pass or not, using variables that capture the impact of the road geometry, traffic conditions and drivers' characteristics. It was found that while the traffic related variables had the most important effect on passing decision, factors related to the geometric design and the driver characteristics also had a significant effect on these decisions.

Overtaking maneuvers on two-lane rural roads are difficult maneuvers which involve relatively complicated decisions. The main hypothesis tested in this paper is that the frequency of overtaking maneuvers on a driving simulator is associated with a faulty decision making style in the Iowa Gambling Task (IGT), a popular decision task employed for assessing cognitive impulsivity. In a controlled study, 36 participants drove a scenario involving multiple overtaking decisions in an interactive driving simulator (STISIM) and also completed the IGT. The results show a significant negative correlation of about 0.3 between the IGT performance and the number of overtaking maneuvers, the average driving speed, and the acceleration noise. We also found a positive correlation of 0.5 between IGT performance and the percent of aborted overtaking maneuvers. A cognitive modeling analysis shows that the associations appear to be modulated by weighting of gains compared to losses obtained during repeated play. These results demonstrate that the IGT has a potential to predict risk prone behavior in overtaking maneuvers and in driving in general.

This paper investigates the destination choice problem in transportation planning processes. Most models assume a Multinomial Logit (MNL) form for the problem. The MNL cannot account for unobserved similarities which exist among choice alternatives. The purpose of this paper is to investigate alternative destination choice model structures, focusing on closed-form models. The paper reviews recent GEV formulations and discusses the adaptation of these models to destination choice situation. In addition the paper presents a new model structure composed of three hierarchical levels: it assumes a choice process composed of a broad selection of zones based on a specific land use characteristic (in this case, presence of shopping center) and then a finer selection of zones based on a geographical characteristic (in this case, adjacent zones). To illustrate the similarity measures of selected GEV formulations and the new model structure the paper specifies, estimates and compares destination choice models for weekday shopping trips based on a revealed preference survey. The paper discusses the structure of the proposed choice models, similarity measures and implementation issues related to the GEV destination choice models.

The concept of path size attempts to capture correlations among routes in route choice modeling by including a correction term in the multinomial logit formulation. Several correction terms were proposed in the literature, yet no satisfactory derivation based on theoretical arguments is presented, raising doubts about the correct specification of the correction terms. This paper proposes the detailed and systematic derivation of a new formulation of the measure of path size and explicitly defines the assumptions involved in its derivation. The path size correction (PSC) factor results from the notion of aggregate alternative as well from the simplification of nested logit models. The new measure of path size offers a more natural interpretation of the correlation due to spatial overlap of alternative routes. Estimation of PSC-logit models in two real-world networks and calculation of predicted choice probabilities in synthetic networks allow comparison of the new path size measure with respect to the classic one. Estimates show similar performances between the models, and predictions illustrate better performances of the new version of the path size factor.

Typical daily decision-making process of individuals regarding use of transport system involves mainly three types of decisions: mode choice, departure time choice and route choice. This paper focuses on the mode and departure time choice processes and studies different model specifications for a combined mode and departure time choice model. The paper compares different sets of explanatory variables as well as different model structures to capture the correlation among alternatives and taste variations among the commuters. The main hypothesis tested in this paper is that departure time alternatives are also correlated by the amount of delay. Correlation among different alternatives is confirmed by analyzing different nesting structures as well as error component formulations. Random coefficient logit models confirm the presence of the random taste heterogeneity across commuters. Mixed nested logit models are estimated to jointly account for the random taste heterogeneity and the correlation among different alternatives. Results indicate that accounting for the random taste heterogeneity as well as inter-alternative correlation improves the model performance.

Transportation and urban land use maintain a complex, symbiotic relationship. Transportation affects land use by improving accessibility to urban functions, and the built environment affects travel through its distribution and density. It is important to explore these dynamics in order to understand the effects that determine urban and metropolitan daily travel. The main claim of this chapter is that land use characteristics account for at least some of the daily travel habits in the metropolis. Current practice in transportation demand modelling emphasises socioeconomic factors as explanatory variables for daily travel rates. Land use characteristics, however, are less considered, if at all, in such models. This chapter presents findings from a study conducted on Light Rail Transit (LRT) stations in Tel Aviv Metropolitan Area (TAMA), in which land use characteristics were explicitly included in travel generation models and forecasts. These land use variables were then tested in several land use development policies in terms of population and employment density and distribution. The analysis is organised as follows: The first part reviews the literature dealing with land use effects on urban travel trends. The general dynamics of land use and travel trends are discussed, followed by a review of land use variables-such as density and diversity, land use mix, and their impact on travel behaviour. Thereafter follows a brief overview of land use parameters involved in transportation modelling. The methodological section presents the trip-generation models and briefly describes the simulations of land use development policy scenarios for the LRT stations. The results of the estimation process and the comparison of land use policies is then presented, followed by a discussion of some key findings regarding several land use variables which seem to be especially influential in determining trip patterns. Finally, conclusions are drawn regarding land use variables' ability to represent intrinsic travel demand.

Few of the recently developed route choice models have actually been applied in traffic assignment problems. This paper discusses the implementation of selected route choice models in stochastic user equilibrium algorithms. The focus of the paper is on path-based assignment, which is essential in the implementation of route choice models. The paper analyzes the effect of choice set size and selected choice models on problem convergence, running time and selected results. The results presented in the paper indicate that for real-size networks, generation of a large number of alternative routes is needed. Furthermore, convergence properties greatly improve if the generated routes are sufficiently disjointed.

Most stochastic user equilibrium (SUE) model applications reported in the literature are based on the multinomial logit (MNL) model. This paper presents a SUE assignment based on the cross-nested logit (CNL) route choice model, which can better represent route choice behavior. The paper develops path-based algorithms to solve the CNL-SUE problem based on adaptation of the disaggregate simplicial decomposition method. The algorithms differ for the step-size determination; three different methods are considered. The algorithms are tested in two well-known networks. Two main tests are conducted: (a) the impact of the CNL model parameters on the assignment results is analyzed and (b) the differences between the CNL-SUE and the MNL-SUE solutions are investigated. The results indicate that the path-based algorithm with Armijo's step-size rule outperforms other step-size determinations. This paper indicates that depending on the model parameters, particularly the nesting coefficient, the CNL-SUE path flows may be quite different from the MNL-SUE path flows.

Most route choice models are related to revealed choice behavior and are estimated by adding alternative paths to observed routes. This paper focuses on the effects of choice set composition in route choice modeling by designing an experimental analysis of actual route choice behavior of individuals driving habitually from home to work in an urban network. The numerical analysis concentrates on a qualitative perspective, by considering path sets built with different generation techniques, and a quantitative perspective, by accounting for path sets constructed with sample size reduction from each initial choice set. Comparison of prediction accuracy across different choice sets suggests that a recently developed branch and bound algorithm generates heterogeneous routes that allow for estimating models with better prediction abilities with respect to the outcomes of the drivers' actual choices. Further, comparison of route choice models across different choice set compositions indicates that nonnested structures, such as C-logit and path size logit, yield more robust parameter estimates.

Recent developments in sensing and tracking technologies have enabled large geographical databases to be established that represent spatial dynamics of 'behavioral entities'. Within this type of dynamics there are several levels and modes of organization that need to be revealed. Clusters are high-level groupings of entities, where change in their location and form, including split and merge events, represents self-organization and functioning patterns. Such information may contribute for better understanding spatially complex dynamic patterns. The main objective of this article is to develop an adaptable methodology that facilitates exploration of spatial order and processes in point pattern dynamics. The approach presented here utilizes data-clustering at each snapshot of the moving pattern, and then involves pairwise linking between the clusters identified at each snapshot and those identified in the following snapshot. Such linking is based on a new methodology that defines well globally optimized solutions for numerous possible linking combinations based on Linear Programming. A preliminary assessment of the approach was conducted with an existing Ants' simulation tool, capable of creating data sets covering in detail a substantial portion of the nest's life cycle.

This paper presents research to evaluate drivers' passing maneuvers on two-lane rural roads using data collected with STISIM, an interactive driving simulator. In addition to the observations of driving behavior obtained in the simulator experiments, drivers' socioeconomic characteristics and indicators of their driving style were collected using self-reported questionnaires. The participants were asked to drive a 9.5 km twolane rural road section with no intersections. The positions and speeds of the subject vehicle and the other simulated vehicles were recorded at a resolution of 0.1 seconds. The data collected in the experiment was used to develop a model that explains drivers' passing decisions. The results indicate that the speed of the subject vehicle and its relations to the vehicle being passed are the most important factors affecting passing behavior. In addition, drivers' socio-demographic characteristics and driving styles also affect passing decisions.

This paper discusses choice set generation and route choice model estimation for large-scale urban networks. Evaluating the effectiveness of Advanced Traveler Information Systems (ATIS) requires accurate models of how drivers choose routes based on their awareness of the roadway network and their perceptions of travel time. Many of the route choice models presented in the literature pay little attention to empirical estimation and validation procedures. In this paper, a route choice data set collected in Boston is described and the ability of several different route generation algorithms to produce paths similar to those observed in the survey is analyzed. The paper also presents estimation results of some route choice models recently developed using the data set collected.

An algorithm to solve explicitly the path enumeration problem is proposed. This algorithm is based on the branch-and-bound technique and belongs to the class of deterministic methods along with existing approaches that combine heuristic or randomization procedures with shortest-path search. The branch-and-bound algorithm is formulated, and a methodology is designed for the application of deterministic approaches to a real case study. Path sets generated with different methods are compared for behavioral consistency, namely, the ability to reproduce actual routes chosen by individuals driving habitually from home to work. Choice set compositions for modeling purposes are determined for the consistency of the path generation process with the observed behavior. Further, model estimates and performance for different route choice specifications are examined for both path set compositions. Results suggest that the proposed branch-and-bound algorithm generates realistic and heterogeneous routes, reproduces better the observed behavior of the interviewed drivers, and produces a good choice set for route choice model estimation and performance comparison.

In recent years, interest in finding methods to measure and analyze airline passengers' preferences has been growing. The current state of the air transportation industry encourages the implementation of tools and methodologies for such purposes. Initial modeling results of research on the preferences of domestic airline users are presented. A field survey (computerized questionnaire on a laptop computer) was conducted in five of Israel's domestic airports. Respondents were asked to rank four alternatives as main level-of-service variables. Separate models were estimated for three passenger trip purposes: personal, tourism, and business. The results of single-choice and rank-ordered logit models are compared for all available observations. The results indicate that the combination of first- and second-rank data sets produces coefficient estimates that are efficient but not significantly better than those of the single-choice logit models.

This paper focuses on path-based solution algorithms to the stochastic user equilibrium (SUE) and investigates their convergence properties. Two general optimization methods are adapted to solve the logit SUE problem. First, a method that closely follows the Gradient Projection (GP) algorithm developed for the deterministic problem is derived. While this method is very efficient for the deterministic user equilibrium problem, we use a simple example to illustrate why it is not suitable for the SUE problem. Next, a different variant of gradient projection, which exploits special characteristics of the SUE solution, is presented. In this method the projection is on the linear manifold of active constraints. The algorithms are applied to solve simple networks. The examples are used to compare the convergence properties of the algorithms with a path-based variant of the Method of Successive Averages (MSA) and with the Disaggregate Simplicial Decomposition (DSD) algorithm.

In the context of route choice, modeling the process that generates the set of available alternatives in the mind of the individual is a complex and not fully explored issue. Route choice behavior is influenced by variables that are observable, such as travel time and cost, and unobservable, such as attitudes, perceptions, spatial abilities, and network knowledge. In this study, attitudinal data were collected with a web-based survey addressed to individuals who habitually drive from home to work. The paper proposes a methodology to conduct a proper application of factor analysis to the route choice context and describes the preparation of an appropriate data set through measures of internal consistency and sampling adequacy. The paper shows that, for the data set obtained from the web-based survey, six latent constructs affecting driver behavior were extracted and scores of each driver on each factor were calculated.

Several route choice models are reviewed in the context of the stochastic user equilibrium problem. The traffic assignment problem has been extensively studied in the literature. Several models were developed focusing mainly on the solution of the link flow pattern for congested urban areas. The behavioural assumption governing route choice, which is the essential part of any traffic assignment model, received relatively much less attention. The core of any traffic assignment method is the route choice model. In the wellknown deterministic case, a simple choice model is assumed in which drivers choose their best route. The assumption of perfect knowledge of travel costs has been long considered inadequate to explain travel behaviour. Consequently, probabilistic route choice models were developed in which drivers were assumed to minimize their perceived costs given a set of routes. The objective of the paper is to review the different route choice models used to solve the traffic assignment problem. Focus is on the different model structures. The paper connects some of the route choice models proposed long ago, such as the logit and probit models, with recently developed models. It discusses several extensions to the simple logit model, as well as the choice set generation problem and the incorporation of the models in the assignment problem.

Travel cost is one of the most important variable in travel demand choice models, particularly in mode choice models. Modelers use different data and assumptions to calculate the cost of auto trips and use them to estimate travel choices made by individuals. However, while people make their choices based on their perceived costs, the modeler estimates the choices using a cost function. Since the costs produced by the cost function may be different from the perceived costs, this may cause a bias in the model. The purpose of this paper is to investigate the individual's perception of auto travel cost and the parameters that affect this perceived cost. Perceived cost was studied through a survey designed for this purpose and the results were also compared with calculated values taken from local sources.

An adaptation of the general logit kernel (LK) model to the route choice context is presented. Recent intelligent transportation systems applications have highlighted the need for better models of the behavioral processes involved in route choice. Several route choice models have been developed recently. However, few studies concentrated on the model estimation and applications for large urban networks. How LK can be adapted to route choice situations by suitably defining the elements of the model is described. The model is estimated using a sample formed from a route choice survey combined with network variables. Preliminary estimation results are presented and discussed.

The route choice problem is complicated in typical transportation networks because of the size of the choice set and because of the overlapping problem since many routes share links. Well-known models like the probit and the logit were further developed in an attempt to overcome these problems. The logit model has the appeal of being relatively close to the probit model while keeping a convenient analytical closed form. However, the simple multinomial logit model cannot correctly represent route choice, especially with respect to the overlapping problem. Other hierarchical logit models can potentially overcome the overlapping problem. The recently developed generalized nested logit (GNL) model is found to be very suitable for route choice, as is the cross-nested logit (CNL) model. The inclusion of the congestion effect in the route choice problem is accounted for in stochastic user equilibrium (SUE) problems. The development of a SUE formulation for the GNL model is presented. In addition, how to adapt the GNL model to route choice in a way similar to that of the CNL model is shown. An equivalent SUE formulation for the GNL model is developed. In this way, a unified framework is presented to relate GNL-type models, which are derived from discrete choice theory, with aggregate entropy formulations. A preliminary algorithm is developed to illustrate the potential application of the GNL formulation for real networks.

Traffic assignment models can be classified according to the behavioral assumption governing route choice. The deterministic user equilibrium (UE), stochastic user equilibrium (SUE) and system optimum (SO) models have been studied extensively in the literature. The relationship between the UE solution and the SO solution for a given network is well known, as is the relationship between UE and SUE. The question that arises concerns the relationship between SUE and (deterministic) SO. The flow pattern obtained from the SO solution serves as a yardstick for comparison with the flow patterns obtained from the UE and SUE solutions. The investigation examines whether the stochastic equilibrium is `closer' than the deterministic user equilibrium to the system optimum. This paper compares the performance of the different solutions for simple networks. The comparison is made by evaluating the relative difference in total system times for UE and SUE solutions with respect to the SO solution. This paper also presents an extension of previous results to show that the Braess' paradox can occur for certain ranges of demand volumes in the case of stochastic equilibrium and non-linear cost functions.

Different effects in stochastic user-equilibrium formulations are compared. The starting point is the logit assignment formulation for the stochastic user-equilibrium model. Recently, extended logit-based models were developed as solutions of equivalent stochastic user-equilibrium problems. These extended logit models are theoretically superior to the simple logit model, because they take into account the similarity among routes. The similarity indexes were defined based on physical parameters of the network, such as link lengths, and therefore did not depend on congestion. The assumption that similarity coefficients are independent of congestion means that the similarity effect and the congestion effect are treated separately. However, both similarity and congestion effects are taken into account in the equilibrium formulations. Simple network examples are presented to illustrate the sensitivity of the link flow pattern (and path flow pattern) to the three different effects in stochastic assignment models: the congestion effect, the stochastic effect, and the similarity effect. The relative influence of those effects on the flow patterns is discussed.

New stochastic user-equilibrium formulations are presented. In the transportation literature, the 'logit assignment' stands for a stochastic user-equilibrium model in which the multinomial logit is the route-choice model. Efficient algorithms using this mathematical formulation were proposed to solve the logit assignment. However, the use of the logit function for route choice has some theoretical drawbacks. In typical transportation networks, many routes have common links, and the structure of the model is not able to account for these common links because the probability for choosing a route is computed based solely on the total route cost. Recently, extended logit-based models were proposed to overcome the overlapping problem and keep the analytical tractability of the logit function. It is demonstrated how extended logit models - such as the Cross-Nested Logit and the Paired Combinatorial Logit - can be derived from more general entropy-type formulations, thus allowing the use of existing (and yet under development) algorithmic solutions for the more general logit-family stochastic assignment model.

The network loading process of stochastic traffic assignment is investigated. A central issue in the assignment problem is the behavioral assumption governing route choice, which concerns the definition of available routes and the choice model. These two problems are addressed and reviewed. Although the multinomial logit model can be implemented efficiently in stochastic network loading algorithms, the model suffers from theoretical drawbacks, some of them arising from the independence of irrelevant alternatives property. As a result, the stochastic loading on routes that share common links is overloaded at the overlapping parts of the routes. Other logit-family models recently have been proposed to overcome some of the theoretical problems while maintaining the convenient analytical structure. Three such models are investigated: the C-logit model, which was specifically defined for route choice; and two general discrete-choice models, the cross-nested logit model and the paired combinatorial logit model. The two latter models are adapted to route choice, and simple network examples are presented to illustrate the performance of the models with respect to the overlapping problem. The results indicate that all three models perform better than does the multinomial logit model. The cross-nested logit model has an advantage over the two other generalized models because it enables performing stochastic loading without route enumeration. The integration of this model with the stochastic equilibrium problem is discussed, and a specific algorithm using the cross-nest logit model is presented for the stochastic loading phase.

A new link-nested logit model of route choice is presented. The model is derived as a particular case of the generalized-extreme-value class of discrete choice models. The model has a flexible correlation structure that allows for overcoming the route overlapping problem. The corresponding stochastic user equilibrium is formulated in two equivalent mathematical programming forms: as a particular case of the general Sheffi formulation and as a generalization of the logit-based Fisk formulation. A stochastic network loading procedure is proposed that obviates route enumeration. The proposed model is then compared with alternative assignment models by using numerical examples.

Congestion of aircraft traffic along airways and in the terminal areas of large airports has significantly increased in recent years. To handle the new levels of congestion in the skies a better understanding of air conflicts is necessary. This paper presents a stochastic model that evaluates the number of conflicts for a spatial geometry of two intersecting airways. The stochastic nature of any desired parameter that defines aircraft location or performance characteristics is accounted for by the simulation process. The structure of the model ensures its efficiency in terms of computer resources as well as its accuracy. The effect of shape of the arrival distribution functions on the shape of the conflict distribution function and its variance was investigated. The variance of the number of conflicts is dependent on the aircraft-arrival distribution. This association is not a simple one, since an airplane on one path can conflict with more than one airplane on the other path. Thus the covariances in the variance term are dependent on the interarrival time described by the distribution. In a similar way the shape of the distribution function of conflicts was found to be dependent on the aircraft-arrival rate distributions.