A Shortest Path Problem in an Urban Transportation Network Based on Driver Perceived Travel Time

Abstract

This paper proposes a method to solve shortest path problems in route choice processes when each link's travel time is a fuzzy number, called the Perceived Travel Time (PTT). The PTT is a subjective travel time perceived by a driver. The algorithm solves the fuzzy shortest path problem (FSPA) for drivers in the presence of uncertainty regarding route travel time. For congested networks, the method is able to nd the shortest path in terms of perceived travel time and degree of saturation (congestion) along routes at the same time. The FSPA can be used to support the fuzzi cation of tra c assignment algorithms. The applicability of the resulting FSPA for the tra c assignment was tested in conjunction with incremental tra c loading and was applied to a large-scale real network. The results of the tra c assignment based on the FSPA, User Equilibrium (UE) and a stochastic loading network model (Dial's assignment algorithm) were compared to the observed volume for certain links in the network. We conclude that the proposed method o ers better accuracy than the UE or Dial's assignment algorithm for the network under testing.