In this paper, we re-evaluate the basic strategies for label correctingalgorithms for the multiobjective shortest path (MOSP) problem, i.e., node andlabel selection. In contrast to common believe, we show that---when carefullyimplemented---the node-selection strategy usually beats the label-selectionstrategy. Moreover, we present a new pruning method which is easy to implementand performs very well on real-world road networks. In this study, we test ourhypotheses on artificial MOSP instances from the literature with up to 15objectives and real-world road networks with up to almost 160,000 nodes.
展开▼
