Browsing by Author "Di, Xuan"

Now showing 1 - 6 of 6
  • Thumbnail Image
    Item
    Boundedly rational user equilibrium: models and applications
    (2014-08) Di, Xuan
    Efficient transportation management requires good understanding of people's travel behavior. Most transportation planning models assume travelers are perfectly rational in decision-making. However, much of the empirical evidence from psychology, economics, and transportation has shown that perfect rationality is not realistic in modeling travelers' decision-making process. Thus existing transportation planning models may provide inaccurate predictions to transportation planners. Motivated by travelers' route choice changes in response to the reopening of the I-35W Bridge in Minneapolis, this dissertation shows that travelers are boundedly rational (BR) in making route choices. Though the BR travel behavioral model was proposed in the 1980's, empirical validation of such behavioral principle using real-world data along with a theoretical framework was non-existent. This study is dedicated to bridging these gaps from both empirical and theoretical perspectives.The first contribution of this dissertation is the empirical verification and estimation of boundedly rational route choice behavior. By analyzing recorded GPS trajectories from 143 commuters before and after the reopening of the I-35W Bridge in Minneapolis, we employ a probit model to estimate the bounded rationality parameters in Twin Cities. Despite the behavioral appeal of bounded rationality, a rigorous study of boundedly rational user equilibria (BRUE) solution has been lacking, partly due to its mathematical complexity. This research offers a systematic approach of deriving the BRUE solutions analytically on networks with fixed travel demands. Based on the definition of ε-BRUE, where ε is the indifference band for perceived travel times, we formulate the ε-BRUE problem as a nonlinear complementarity problem (NCP). With the increase of the indifference band, the path set that contains equilibrium flows will be augmented and the critical values of the indifference band to augment the path set can be identified by solving a sequence of mathematical programs with equilibrium constraints (MPEC). A novel solution method is provided to obtain the BRUE solution set and numerical examples are given to illustrate this finding. To provide guidelines to policy-makers for congestion mitigation, this research also explores an important phenomenon which should be avoided in transportation network design, i.e., Braess paradox. The classical Braess paradox was built upon the perfectly rational behavioral assumption. Under the framework of bounded rationality, each equilibrium flow pattern leads to a different total system travel time, resulting in non-unique network performance measures. Because of the non-uniqueness of BRUE solutions, which particular equilibrium pattern should be used to compare network performances before and after new roads are built remains a question. This dissertation aims to study the analytical properties of Braess paradox under bounded rationality by exploring the relationships between the occurrence of Braess paradox and the indifference band as well as the demand level. The unveiled relationships offer a guideline for transportation planners to prevent the occurrence of Braess paradox and pave the way for strategic transportation management under the bounded rationality assumption.
  • Thumbnail Image
    Item
    Development of Algorithms for Travel Time-Based Traffic Signal Timing, Phase I – A Hybrid Extended Kalman Filtering Approach for Traffic Density Estimation along Signalized Arterials
    (2010-12) Liu, Henry X.; Di, Xuan
    As technologies continue to mature, the concept of IntelliDrive has gained significant interest. Besides its application on traffic safety, IntelliDrive also has great potential to improve traffic operations. In this context, an interesting question arises: If the trajectories of a small percentage of vehicles (IntelliDrive vehicles) can be measured in real time, how can we use such data to improve traffic management? This research serves as a starting point that aims to produce a paradigm shift to optimize the traffic signal control from the use of the conventional fixed-point loop detector data to the use of mobile vehicle trajectory-based data. Since the change of density on arterials can help traffic engineers to track the queue length at intersections, which is important for traffic signal optimization, in this project we will focus on the estimation of traffic density on urban arterials when trajectories from a small percentage of vehicles are available. Most previous work, however, focuses on freeway density estimation based merely on detector data. In this research, we adopt the MARCOM (Markov Compartment) model developed by Davis and Kang (1994) to describe arterial traffic states. We then implement a hybrid extended Kalman filter to integrate the approximated MARCOM with fixed-point and vehicle-trajectory measurements. We test the proposed model on a single signal link simulated using VisSim. Test results show that the hybrid extended Kalman filter with vehicle-trajectory data can significantly improve density estimation.
  • Thumbnail Image
    Item
    Estimation of Crossing Conflict at Signalized Intersection Using High-Resolution Traffic Data
    (Minnesota Department of Transportation, 2017-03) Liu, Henry X.; Davis, Gary A.; Shen, Shengyin; Di, Xuan; Chatterjee, Indrajit
    This project explores the possibility of using high-resolution traffic signal data to evaluate intersection safety. Traditional methods using historical crash data collected from infrequently and randomly occurring vehicle collisions can require several years to identify potentially risky situations. By contrast, the proposed method estimates potential traffic conflicts using high-resolution traffic signal data collected from the SMART-Signal system. The potential conflicts estimated in this research include both red-light running events, when stop-bar detectors are available, and crossing (i.e. right-angle) conflicts. Preliminary testing based on limited data showed that estimated conflict frequencies were better than AADT for predicting frequencies of angle crashes. With additional validation this could provide a low-cost and easy-to-use tool for traffic engineers to evaluate traffic safety performance at signalized intersections.
  • Thumbnail Image
    Item
    Indifference Bands for Route Switching
    (2016) Di, Xuan; Liu, Henry; Zhu, Shanjiang; Levinson, David M
    The replacement I-35W bridge in Minneapolis saw less traffic than the original bridge though it provided substantial travel time saving for many travelers. This observation cannot be explained by the classical route choice assumption that travelers always take the shortest path. Accordingly, a boundedly rational route switching model is proposed assuming that travelers will not switch to the new bridge unless travel time saving goes beyond a threshold or indifference band. To validate the boundedly rational route switching assumption, route choices of 78 subjects from a GPS travel behavior study were analyzed before and after the addition of the new I-35W bridge. Indifference bands are estimated for both commuters who were previously bridge users and those who never had the experience of using the old bridge. This study offers the first empirical estimation of bounded rationality parameters from GPS data and provides guidelines for traffic assignment.
  • Thumbnail Image
    Item
    Multi-agent Route Choice Game for Transportation Engineering
    (Transportation Research Board, 2015) Di, Xuan; Liu, Henry; Levinson, David M
    In undergraduate transportation engineering courses, traffic assignment is a difficult concept for both instructors to teach and for students to learn, because it involves many mathematical derivations and computations. We have designed a multiplayer game to engage students in the process of learning route choice, so that students can visualize how the traffic gradually reach user equilibrium (UE). For one scenario, we employ a Braess' Paradox, and explore the phenomenon during the game-play. We have done the case-control and before-after comparisons. The statistical results show that, students who played the game improve their understanding of the Braess' Paradox more than those who did not play. Among game players, younger students benefit more in their learning; while those who are not comfortable with exploring a phenomenon on their own think this game not as effective as those who prefer hands-on learning experiences.
  • Thumbnail Image
    Item
    Unexpected versus expected network disruption: Effects on travel behavior
    (Elsevier, 2017-02) Danczyk, Adam; Di, Xuan; Liu, Henry X; Levinson, David M
    This paper discusses the observed evolution of traffic in the Minneapolis-St Paul (Twin Cities) region road network following the unexpected collapse of the I-35W Bridge over the Mississippi River. The observations presented within this paper reveal that traffic dynamics are potentially different when a prolonged and unexpected network disruption occurs rather than a preplanned closure. Following the disruption from the I-35W Bridge's unexpected collapse, we witnessed a unique trend: an avoidance phenomenon after the disruption. More specifically, drivers are observed to drastically avoid areas near the disruption site, but gradually return after a period of time following the collapse. This trend is not observed in preplanned closures studied to date. To model avoidance, it is proposed that the tragedy generated a perceived travel cost that discouraged commuters from using these sections. These perceived costs are estimated for the Twin Cities network and found to be best described as an exponential decay cost curve with respect to time. After reinstituting this calibrated cost curve into a mesoscopic simulator, the simulated traffic into the discouraged areas are found to be within acceptable limits of the observed traffic on a week-by-week basis. The proposed model is applicable to both practitioners and researchers in many traffic-related fields by providing an understanding of how traffic dynamics will evolve after a long-term, unexpected network disruption.