Congestion Control and Evolutionary Games in Traffic Management

University Of California-Los Angeles, Los Angeles CA

Investigators

Abstract

This project explores and develops tools for traffic management strategies based on recent advances in data network congestion control and evolutionary game theory. The objective is to develop constructive methods to influence traffic distribution patterns in road and freeway systems in order to improve efficiency and reduce congestion. The intent is to derive traffic management strategies based on localized feedback and incentives, such as advanced signaling, metering and tolling, for both dynamic incident management and long-term traffic behavior. The research path stems from recent successes in applying analytical tools for the control of telecommunication networks and new developments in the theory of evolutionary games. Project personnel have been involved in a new formulation for the Internet congestion control problem based on user utility functions and "congestion pricing". Project personnel also have been involved in new methods for evolutionary game theory to derive mechanisms for the evolution of strategies that assure a group of interacting decision makers will converge to an equilibrium without explicitly sharing individual utility functions. This project explores, both analytically and by simulation, to what degree these advances can have a bearing on transportation networks for the specific problems of 1) System level mandatory routing mechanisms for incident congestion, 2) Vehicle level incentive routing for incident congestion, and 3) Vehicle level incentive routing for recurring congestion. A potential long-term societal impact of this project is a greater efficiency of the transportation system and enhancing the ability of individual users to improve their traffic experience. A more immediate educational impact will be to train students on a multi-disciplinary viewpoint suitable for complex networks through a seminar course to be developed at UCLA. As an additional outreach effort, the project will develop an educational Website where visitors can interact with a graphical interface and see the effect of policy decisions on traffic patterns in a sample freeway system.