The α-reliable path problem in stochastic road networks with link correlations: A moment-matching-based path finding algorithm

Most existing studies on routing guidance only paid attention to the average path travel time, which failed to consider travel time reliability (TTR) preferences by different travelers. In this study, a moment-matching-based hybrid genetic algorithm (MHGA) is proposed to search the reliable shortest path (RSP) in stochastic road networks with link correlations. First, the goodness-of-fit results based on field data reveal that lognormal distributions are more appropriate for characterizing link travel times. The impact of topological distance (measured by the number of links) and road type on link correlations is also scrutinized. Then, a moment-matching method (MOM) is utilized to determine the parameters of the approximate path travel time distribution (TTD) by accounting for link correlations. A local search algorithm is designed to improve the search ability of the path finding algorithm. In view of travelers' risk tolerance, the algorithm enables the provision of personalized routing guidance for individual travelers. Furthermore, to support path finding applications in a large-scale network, heuristic constraints are imposed to help reduce the computational workload and accelerate the convergence speed of the search process. Finally, numerical case studies based on synthetic networks and a real road network in Beijing are presented, and the results help demonstrate that the algorithm has good potential to solve RSP searching problems in a large-scale network with desirable efficiency.