Weather Routing Optimization: A New Shortest Path Algorithm

This paper presents an algorithm which solves the multiobjective shortest path problem in a time-dependent graph, taking advantage of the specificities of the weather routing problem. Multicriteria shortest path problems are widely studied in the literature, as well as monocriteria shortest path problems in time-dependent graphs. Their solving has numerous applications, especially in the transportation field. However, the combination of both these issues is not studied as much as each one separately. In this paper, we study the weather routing problem for cargo ships, which involves optimizing the ship routes following realtime weather information. For this problem, the arc weights on the graph have a low dispersion around their average value. We propose an extension of an algorithm (NAMOA*) taking advantage of this property. We study the validity of this new algorithm and explain why it solves efficiently the weather routing problem. Experiments done using real weather data corroborates the algorithm efficiency.