Optimal Coordination of Ramp Metering Via Iterative Dynamic Programming

Abstract

This paper presents an online approach for the optimal coordination and control of ramp metering. The ramp metering problem is formulated in a decision network and subsequently solved by iterative dynamic programming (IDP), where a near optimal ramp metering policy is obtained by minimizing the total travel time spent (TTS). The optimal metering policy is implemented under the framework of receding horizon control. In this study, by incorporating METANET for traffic flow model, a study location is simulated and the performance of the proposed algorithm is measured and compared to those without ramp metering and with a local uncoordinated ramp metering. Results show that the proposed method is able to improve traffic conditions and prevent recurrent congestion under certain ramp queue constrains.