Resilience Patterns in Dynamic Aircraft-to-Aircraft Communication Networks

While resilience of infrastructure networks with fixed size and topology have been quantified using network science methods, the underlying mechanisms have not been elucidated for resilience of dynamic networks that change in size and topology. Using aircraft-to-aircraft (A2A) communication networks, we demonstrate whether and how changes in network size and topology over time affect the resilience of dynamic infrastructure networks. We simulate failure and recovery of spatially constrained A2A communication networks for an airspace region in the U.S. over multiple hourly time windows in a day. These networks were constructed by mapping aircraft locations to nodes and communication channels between aircrafts to edges based on spatial proximity and bandwidth availability. The results imply that the number and relative sizes of the connected network components are key factors affecting the variation in resilience of dynamic networks over time across diverse failure events. Our network science-based simulation approach captures failure patterns of A2A networks beyond an existing analytical model for transportation systems that is limited to locally tree-like random networks. Furthermore, network centrality-based mixed recovery strategies (i.e., betweenness and degree with switching over time) outperform random recovery across multiple hourly time windows. Thus, hourly variability in resilience patterns of A2A communication networks should be a key consideration for developing dynamic airspace transportation risk mitigation insights.