(562g) Exploring the Resilience and Sustainability of the US Aviation Sector Via Graph Theoretic Approaches

The US Department of Homeland Security identifies a list of 16 critical infrastructure sectors (CIS). These CIS provide the essential services to the US economy and a disruption on any of these sectors could have a debilitating impact on nation’s security, economy, health and safety. Given the critical role of the CIS to the US economy and their high interconnectedness with one another and other US industrial sectors at multiple scales, CIS offer an excellent context in which to understand infrastructure vulnerabilities and resilience. One of the essential critical infrastructures is the US aviation sector. The reliance of multiple CIS and other industrial sectors for their operations either directly or indirectly on the US aviation sector indicates the critical importance of this CIS to the US economy. Similar to natural ecosystems, CIS (including the US aviation sector) are complex adaptive systems with many interconnections and interdependencies and exhibit emergent properties. Central to developing resilient CIS is gaining an understanding of the interconnectedness and interdependencies between CIS including feedback loops. Such understanding is critical for identifying systemic vulnerabilities and hence designing for resilience.

Graph theory based methods have been extensively applied to understand the structure and behavior of networks. However, they remain largely unexplored for understanding resilience for CIS. Furthermore, graph theoretic approaches have many interesting insights to offer, however, they has been largely ignored in the existing sustainability literature. We utilize a graph theoretic framework to study the network structure and topology of the US aviation sector. The US aviation sector is modeled as directed, weighted graph with passenger flows between pair of airports representing the weights. A variety of network and node level metrics are developed and estimated to analyze the structure of the US aviation sector. This includes metrics like degree centrality, shortest path analysis, eigenvector centrality, and network efficiency. We also perform robustness analysis in response to partial (random) and complete disruption (targeted attacks) of the nodes. We compare and contrast the network topology and efficiency of the original network with the disrupted network to gain insights about the critical nodes in the network, and understand robustness of the US aviation sector to disruptions. Several different clustering and community detection algorithms are utilized to help uncover the pattern of connections and aid in understanding the cascading impact of disruptions and resilience implications in the US aviation network. The implications of the results for developing risk management strategies and guiding the resilient operation of the network will also be described.