Hierarchical shortest path finding algorithm with probabilistic travel time

抄録

With the development of Route Guidance System (RGS), travelers' route choice behaviors change rapidly yet the existing modeling methods could not follow the speedy change. For example, most route choice models regard travel time as a deterministic value so the uncertainty of travel time is ignored. In addition, many of the previous route choice models were formulated without integrating the driver's attitude and preference in detail. In this paper, a new route choice model which can deal with the mentioned topics is proposed. With the combined concepts of hierarchical road classification and resource constraints, non-overlapping paths are enumerated in the real network. After then, the model calculates the path choice probabilities with respect to the given preferred arrival time. Using a real-size network, the proposed model was tested and the calculated path choice probabilities are compared to the True Success Rate (TSR) and multinomial logit (MNL) model.