WeatherRouting Bench 1.0 Now Live!
Published:
The Weather Navigation project is excited to announce the release of WeatherRouting Bench 1.0, the first comprehensive benchmark for weather routing. This tool addresses a significant gap in weather routing research: the absence of a standardized method for comparing algorithms.
Our benchmark provides a website with Python code to help researchers utilize weather data, test routes, and apply a consistent cost function for evaluating different methods. We also present optimization problems based on real shipping routes, each including a standard route as a baseline for comparison.
In the forthcoming paper accompanying this benchmark release, we explore how different problem settings impact optimization. We analy how the average cost of a journey varies with sailing speeds, ocean regions, and times of the year. By offering this benchmark, we aim to simplify the comparison of research efforts in weather routing.
Explore WeatherRouting Bench 1.0
As a first entry to the benchmark, our next paper introduces A⋆-FMS: an algorithm that combines A⋆ graph search with a discrete Newton-Jacobi method. This hybrid approach leverages the global search capabilities of graph methods and the local optimization strength of gradient descent methods. With optimal configuration, A⋆-FMS reduces travel time by 1.4% when sailing at 12 knots and 3.6% at 6 knots, compared to the orthodromic route.
Stay tuned for the release of our next paper!
This research is supported by:
BBVA Foundation via the project “Mathematical optimization for a more efficient, safer and decarbonized maritime transport”.
Spanish Agencia Estatal de Investigación under grant TED2021-129455B-I00, “Optimization of maritime routes with real time oceanographic and meteorological data”.
