FASTEST PATHS IN TIME-DEPENDENT NETWORKS FOR IVHS APPLICATION

The article considers the problem of individual route guidance in an Intelligent Vehicle Highway Systems (IVHS) environment, based on time-dependent forecasts of link travel time. It proposes a consistency condition which deterministic forecasts should be constrained to satisfy, and shows that under consistency, the time-dependent fastest path can be calculated with exactly the same computational complexity as for static fastest paths, regardless of the time index set. In particular, consistency reduces computational complexity by O(T) in a T-period discrete time problem. A comment is made on the stochastic fastest path problem, suggesting modifications to a general-case algorithm from the literature.