Smoothed traffic complexity metrics for airspace configuration schedules

David Gianazza (1)

1 DGAC - Direction Générale de l'Aviation Civile

This paper is a continuation of previous research on optimal airspace configuration. It is expected to improve the predictability and the flexibility of the airspace management process by computing realistic predictions of the sectors opening schedules in En-route ATC centers. In previous papers, we selected relevant complexity metrics to predict the controllers workload, using neural networks trained on recorded airspace configurations. We also introduced new algorithms to build optimally balanced airspace configurations, exploring all possible combinations of elementary sectors. As a result of this previous work, we were able to compute realistic schedules on a whole day of traffic, using complexity metrics that were computed from recorded radar tracks. The raw metrics, however, showed high variations in time which caused a "configuration switching" phenomenon. Although the number of control sectors in the computed schedule stayed globally close to the recorded number of sectors, the airspace was reconfigured much more often than in reality. The present paper shows how the input metrics can be smoothed in order to avoid this problem, and what may be the subsequent problems caused by the smoothing strategy.

Mots clés: air traffic complexity; dynamic density; neural networks; CASSOS

Domaines: Optimisation et contrôle [math.OC]

David Gianazza. Smoothed traffic complexity metrics for airspace configuration schedules. ICRAT 2008, 3rd International Conference on Research in Air Transportation, Jun 2008, Fairfax, United States. pp xxxx. ⟨hal-01020711⟩