A vehicle-centric logic for decentralized and hybrid route guidance

The paper presents a decentralized route guidance algorithm that can function alone or as a component of a hybrid centralized/decentralized route guidance system. The focus of the algorithm´s design is on vehicle-level communication and computation, for the purpose of developing vehicle-centric route guidance. The algorithm has been implemented and tested on a sample network in which both guided and unguided drivers are modeled. Results are given which show that the algorithm is successful at responding to small amounts of network stochasticity as well as to congestion caused by incidents. The results also illustrate the potential of tuning the algorithm´s parameters in real-time, using vehicle-level intelligence.