Recovery centrality of system resilience based on network structure and dynamics

System resilience is a topic of great concern in complex networks. Recovery of system resilience in the face of damage is a key issue. This study proposes the concept of recovery centrality based on the characteristics of the network structure and dynamic behavior to facilitate the recovery of system resilience. Based on reconstructing the damaged network structure, the system is recovered from non-resilient state to resilient state by dynamic micro-interventions. The results show that the proposed recovery centrality index (RCI) can distinguish the recovery capabilities of nodes. The node with the largest RCI can better realize the recovery of system resilience. Compared to other centrality indices (degree centrality, betweenness centrality, and eigenvector centrality), the proposed RCI can better capture the nodes that can recover the system resilience. This study quantifies the influence of nodes on system recovery from the perspective of resilience, which is conducive to formulating more favorable methods for system functionality recovery.