Skip to main content

Studying structural change in the European Aviation Network

Paper ID

SIDs-2023-50

Conference

SESAR Innovation Days

Year

2023

Theme

Machine learning and artificial intelligence

Project Name

Keywords:

change dynamics, graph representation learning, network analysis

Authors

Juul Vossen, Enrico Spinielli and Rainer Koelle

DOI

https://doi.org/10.61009/SID.2023.1.29

Project Number

Abstract

Drastic loss of flight connections due to the COVID-19 pandemic has called for new approaches to accurately study structural change in the European Aviation Network. This study highlights the limitations of traditional centrality-based network approaches and proposes a diffusion-based graph embedding approach using the GraphWave algorithm. This new approach was validated using domain knowledge and tested in its ability to capture known events that occurred during and after the COVID-19 pandemic. The network is modelled based on all flights departing from and arriving to European airports in the period of 2019 through 2022. Flight connections were aggregated on a weekly basis to analyze structural embeddings and the structural role of airports. The temporal analysis supported the identification and assessment of changes to the role of airports and structural changes of the network. This study shows the potential of the approach by applying the model to uncover global, regional, and local change dynamics, and highlighting its potential as a valuable tool for researchers and practitioners studying the evolution of complex networks.