Community Detection in Action: Identification of Critical Elements in Infrastructure Networks

Abstract

Modern infrastructure systems form complex networks that are organized hierarchically in communities of tightly integrated elements. This paper presents three new community-based metrics to identify the critical elements of a network system. Two of these metrics assess intracommunity and intercommunity behavior for any community structure, and the third metric accounts for the multiple levels of community structure. First, these metrics are studied to establish their characteristics with different community structures, and then the Great Britain Railway Network is used as a case study to demonstrate the usefulness of these new metrics. The results show that an assessment of the system using these metrics leads to the identification of not only those elements that are critical at the global level, but also those that greatly affect the local performance of the communities. Such identification of the critical components at the community and global levels would enable a better understanding of system behavior by stakeholders with competing demands.