Risk-Averse Time-Dependent Route Guidance by Constrained Dynamic A Search in Decentralized System Architecture

Both the risk of concentration phenomenon for the navigation system and the risk of encountering unacceptable delay are reduced by guiding risk-averse drivers along reliable paths while avoiding high-risk links as far as possible. A constrained dynamic A* algorithm is proposed to search for a reliable path subject to a trip duration constraint in a time-dependent road network. The efficiency of the dynamic A* search is increased by taking advantage of information computed off-line. The algorithms proposed are implemented, and their computational performance is analyzed in simulation.