Tackling Air Traffic Conflicts as a Weighted CSP : Experiments with the Lumberjack Method

Published 2020-01-30 3 authors

Attribution

This is the abstract and citation. Full text lives at arXiv — we link out rather than host. All credit to the authors and arXiv.

Abstract

Verbatim from arXiv. Not paraphrased, not summarized.

In this paper, we present an extension to an air traffic conflicts resolution method consisting in generating a large number of trajectories for a set of aircraft, and efficiently selecting the best compatible ones. We propose a multimanoeuvre version which encapsulates different conflict-solving algorithms, in particular an original "smart brute-force" method and the well-known ToulBar2 CSP toolset. Experiments on several benchmarks show that the first one is very efficient on cases involving few aircraft (representative of what actually happens in operations), allowing us to search through a large pool of manoeuvres and trajectories; however, this method is overtaken by its complexity when the number of aircraft increases to 7 and more. Conversely, within acceptable times, the ToulBar2 toolset can handle conflicts involving more aircraft, but with fewer possible trajectories for each.

Authors

Thomas Chaboud
Cédric Pralet
Nicolas Schmidt

Keywords

cs.AI

Citation: Thomas Chaboud, Cédric Pralet, Nicolas Schmidt (2020). Tackling Air Traffic Conflicts as a Weighted CSP : Experiments with the Lumberjack Method. arXiv ID 2001.11390. https://arxiv.org/abs/2001.11390 ↗