Paratransit Vehicle Routing and Scheduling

• Dial-A-Ride Paratransit Operations Using Real-time Information Technologies

The primary concern of this research is the advancement of the critical knowledge in the development and application of advanced information technologies such as automated vehicle location systems (AVL), digital telecommunication and mobile computers, to improve the productivity and performance of dial-a-ride paratransit systems. Its intent is to initialize an effort for addressing fundamental questions that arise in implementing and deploying such technologies, and to develop appropriate methodologies, tools and guidelines necessary to plan, design and operate these high technology-based paratransit systems.

The objectives of this research consist of the following: 1). review the current practice of common dial-a-ride paratransit systems, and identify functional needs of these systems for real-time information; 2). review available technological options for use in dial-a-ride paratransit systems, with respect to performance, reliability, and cost; 3). identify the essential technological and methodological questions that determine the cost-effectiveness of applying real-time information technologies; 4). develop and test models and methods for the representation and utilization of real-time information under a variety of operating environments; and 5). apply the methodologies to evaluate the cost of and benefits from real-time information systems, and develop guidelines for planning, designing and operating paratransit systems with such information technologies.

• Development of On-line and Off-line Routing and Scheduling Systems for Dial-A-Ride Paratransit Systems

The goal of this research was set to improve the responsiveness, reliability and productivity of dial-a-ride paratransit services through integration of routing and scheduling principles and practical experience and explicit consideration of travel time variability in urban roadway networks. Such extensive and complex integration has been made possible by improved data acquisition and processing capabilities of computer, telecommunication and vehicle location technologies. An object-oriented, client/server based Windows application software package called FirstWin has being developed to provide decision-support for both on-line and off-line operations. The next step of this research will focus on the evaluation and improvement of the developed algorithms for solving real-life scheduling problems.

On-line Slide Show

Travel Time Modeling and Estimation

• Modeling and Estimation of Travel Time Variability

The ability to accurately predict future link travel times in road traffic networks is a critical component for many Intelligent Transportation Systems (ITS) applications such as in-vehicle route guidance systems (RGS) and advanced traffic management systems (ATMS). Due to random fluctuations in travel demands, interruptions caused by traffic controls, unpredictable occurrences of traffic incidents and changes in weather conditions,   travel time in an urban traffic environment is highly stochastic and time-dependant. It has been increasingly recognized that, for many ITS applications, estimates of the variability of travel times are as important as estimates of the expected average travel times.  For example, having knowledge of the variability of travel times on individual links makes it possible to explicitly consider the reliability of alternative routes in identifying optimal routes. By considering the travel time variability in fleet vehicle routing and scheduling process, more reliable schedules may be generated resulting in improved quality of service.  The objective of this research is to develop models to represent and estimate the variability of travel time on various roadway facilities including freeway and signalized arterials.

• Real-time Prediction of Incident Delay

Rapid deterioration of travel conditions in many urban areas has generated a resurgent interest in more effectively managing congestion caused by traffic incidents. Effective incident management necessitates a thorough understanding of incident characteristics and better models for detecting and verifying the occurrence of incidents and for predicting the incident evolution process. Reliable and timely detection, verification and prediction allow effective response to incidents through adjustment of traffic control strategies and dissemination of incident information to public. The focus of this research is on investigating the evolution patterns of traffic incidents as related to factors such as incident types, severity, location and traffic condition, and to develop appropriate models and tools for real-time prediction of incident duration and incident delay.

On-link Silde Show

• Modeling and Estimation of Origin-Destination Travel Time in Urban Traffic Networks

Improved origin-destination (O-D) travel time estimation has significant implication in applications such as paratransit operation, carrier fleet management and emergency vehicle dispatching. The objective of this research is to develop improved methods for estimating O-D travel times that are time-dependent (dynamic) and stochastic. A new estimation method based on artificial neural network technique has been developed. A Windows application program called NeuralBase has been developed and tested to provide both off- and on-line estimation of dynamic and stochastic origin-destination travel times. The next step of this research will focus on the evaluation of the proposed technique using field travel time data under a wide range of urban traffic conditions, and on the application of this method to an actual vehicle routing and scheduling process such as dial-a-ride paratransit scheduling systems.

On-line Slide Show

Shortest Path Problems and Algorithms with ITS Applications

• Heuristic Shortest Path Algorithms and Their Potential ITS Applications

This research is motivated by the need for more efficient shortest path algorithms for real-time applications such as in-vehicle route guidance systems. By combining heuristic searching strategies from Artificial Intelligence (AI) with traditional labeling algorithms, a variaty of heuristic shortest path algorithms can been developed. Future research will focus on development and evaluation of such algorithms for solving realistically sized network problems.

On-line Slide Show

• Shortest Path Routing in Dynamic and Stochastic Networks

The dynamic and stochastic shortest path problem (DSSPP) is defined as finding the expected shortest path in a traffic network where the link travel times are modeled as a continuous-time stochastic process. The objective of this research is to examine the properties of the problem and to develop algorithms that can be used to solve the DSSPP given information that will be available in networks with Intelligent Transportation System (ITS) capabilities.  When both the dynamic and stochastic nature of link travel times are modeled explicitly, the optimal shortest path algorithms can become computationally inefficient and/or impractical for use within an actual application. The objective of this research is to investigate the problem of finding the expected shortest path in a traffic network where the dynamic and stochastic nature of link travel times is modeled explicitly and to develop an algorithm which can provide improved solutions without significantly adding to the overall computation time.