A class of multicriteria shortest path problems for real-time in-vehicle routing

In-route guidance systems fastest path routing has typically been adopted because of its simplicity. However, empirical studies on route choice behavior have shown that drivers use numerous criteria in choosing a route. The objective of this paper is to develop computationally efficient algorithms for identifying a manageable subset of the nondominated (i.e., Pareto optimal) paths for real-time in-vehicle routing. The basic notion of the proposed approach is that (i) enumerating all nondominated paths is computationally too expensive, (ii) obtaining a stable mathematical representation of the driver's utility function is theoretically difficult and impractical, and (iii) identifying the optimal path given a nonlinear utility function is a nondeterministic polynomial time (NP)-hard problem. Consequently, a heuristic two-stage strategy that identifies multiple routes and then selects the near-optimal path may be effective and practical. As the first stage, we relax the uniqueness of the utility function by measuring the context-dependent preference using an entropy model and propose a branch-and-bound technique that discards most of the nondominated paths. To make sure that the paths identified are dissimilar in terms of links used, the portion of shared links between routes is limited. The test of the algorithm in a large real-life traffic network shows that the algorithm can significantly reduce computational complexity while identifying reasonable alternative paths.