Optimizing multi-modal urban traffic flow: Utilizing macroscopic fundamental diagram and Model Predictive Control

Abstract

Urban transportation systems, characterized by multiple modes and complex dynamics, present significant challenges for the efficient management and optimization of traffic. Addressing these challenges, this study utilizes the Macroscopic Fundamental Diagram to develop and implement Model Predictive Control (MPC) strategies aimed at optimizing traffic flow across multiple urban reservoirs. By designing optimal controllers that regulate the transfer flow of trucks and passenger vehicles, this study aims to maintain vehicle accumulation at a critical level. For this purpose, Centralized Model Predictive Control (C-MPC) and Decentralized Model Predictive Control (DC-MPC) approaches have been formulated to maximize the accumulation of passenger vehicles while reducing the number of trucks in the reservoir system. The findings reveal that the unified approach of C-MPC effectively reduces truck traffic but results in a higher change in passenger travel time. The outcome for segmented C-MPC shows a slower rate of change in vehicle accumulation. While DC-MPC offers a better balance and keeps accumulation for both trucks and passenger vehicles within predefined limits. It contributes to the theoretical understanding of traffic flow optimization and practical insights for city planners and engineers seeking to implement advanced traffic management solutions. Future work can explore the scalability of these controllers and their adaptation to real-time traffic data.