Adaptive Route Guidance Based On Real-time Information In Stochastic Time-dependent Transportation Networks

Owing to time-varying and stochastic properties of transportation networks, modeling transportation networks as stochastic time-dependent (STD) graphs has became a trend. In this paper, by applying multi-lognormal distribution, we develop a mathematical framework that can incorporate the concept of real-time information into the discussion of adaptive route guidance in STD networks. We address some basic aspects about adaptive route guidance based on real-time information in STD networks, such as definition of real-time information, impact of real-time information and spatial-temporal Markov property and so on. Based on least mean-square estimation method, we propose two time-adaptive decision rules with consideration of real-time information. Simulation results on an artificial STD network are given and the effectiveness of our algorithms is verified. At last, some basic conclusions are also validated and the future work is discussed.