BNART: A Novel Centralized Traffic Management Approach for Autonomous Vehicles

Traffic management poses a critical challenge in modern metropolitan areas, where efficient management can significantly reduce travel time and vehicle emissions. With the vision of smart cities in mind, the increasing prominence of autonomous cars is anticipated to revolutionize road networks under the control of a central supercomputer known as the Central Traffic Control Unit, which maintains real-time location information for each vehicle within its jurisdiction and assumes responsibility for path planning. This paper investigates the problem of traffic management for self-driving vehicles, focusing on the utilization of essential information, such as source and destination locations, to guide vehicles along the most efficient routes, mitigating road congestion and ensuring prompt arrivals. To address this challenge, we formulate the problem as a mixed-integer linear programming model. Furthermore, due to its complexity, we propose a heuristic method called the Best Neighbor Algorithm for Routing Traffic (BNART). We extensively evaluate the performance of our proposed heuristic approach against optimal solutions on small-scale instances and other state-of-the-art methods such as the shortest path algorithm on large-scale instances. Through simulation, we demonstrate that our proposed method outperforms other methods in terms of average travel time, exhibiting a negligible gap to the optimal solution.