Ranking multiobjective shortest paths

Abstract

This paper is concerned with the ranking of multi-objective shortest paths accordingly to an order relation verifying certain conditions such is the case, for instance, of the lexicographic order. We present a new labelling algorithm that makes use of shortest deviation paths for obtaining the set of Pareto solutions for the multi-objective shortest path problem. The computational experience reported at the end of the paper shows that the new algorithm clearly outperforms the previous approaches when one looks for the `-th shortest non-dominated paths.