Dynamic Coordinated Routing and Charging Navigation System for Traffic Congestion Mitigation: A Game Theoretical Modeling and Asynchronous Distributed Optimization Approach

Abstract

As the world embraces sustainable transportation, electric vehicles (EVs) have emerged as a promising solution. However, their increasing popularity brings forth new challenges, particularly concerning congestion at charging stations and on roads. Recent advancements in wireless communication and sensing technologies enable travelers to access real-time traffic and charging station availability information via navigation systems, allowing them to make informed routing and charging decisions to avoid congestion. However, this real-time information can sometimes be counterproductive. If travelers react independently and selfishly to similar traffic information, it can exacerbate congestion due to the flash crowd effect. This study aims to alleviate such problems by developing a navigation system based on dynamic coordinated joint routing and en-route charging mechanism (DcRC) to provide coordinated routing and charging guidance for both EVs and internal combustion engine (ICE) vehicles, thereby reducing congestion on roads and at charging stations. Specifically, by incorporating traffic flow and charging station queue dynamics, the DcRC is formulated as a mixed strategy congestion game with an equivalent mathematical programming model, which generates equilibrium routing and charging decisions to mitigate the flash crowd effect and reduce traffic congestion without violating each vehicle's self-interest. To integrate the DcRC into online navigation services, this study develops an asynchronous distributed ADMM-aided Branch-and-Bound (DAB) solution algorithm to efficiently solve the DcRC with hundreds of participants. The DAB draws upon a customized branch and bound algorithm to decompose the complex mathematical program into manageable sub-problems, and utilizes an asynchronous distributed ADMM to solve each sub-problem, leveraging individual vehicles’ computing resources while sustaining robustness against their unstable and stochastic computing and communication performance. Numerical experiments confirm the DcRC's effectiveness in reducing congestion and system costs, as well as the DAB's efficiency in supporting real-time navigation services for hundreds of participants with unstable computation and communication performance.

Keywords: Congestion mitigation, Congestion game, Equilibrium routing mechanism, Distributed algorithm, Robust and resilient algorithm, EV charging