Restoring core-periphery structure of networks

Abstract

Core-periphery structure is commonly observed in diverse real-world networks, where nodes are categorized as core or peripheral nodes by connection patterns. It plays an important role in ensuring potential functionality and intrinsic order in networks. Networks are inevitably suffering from the malicious attacks and random errors, which could lead to the collapse of core-periphery structure, and in turn the performance degradation and even the total instability of network behaviors. Therefore, it is of great significance to rapidly recover such structure after destruction. However, few of the existing works about networks have put emphasis on core-periphery restoration. Here we propose a novel recovery index to characterize and measure the quality for the core-periphery restoration when a disaster happens. Several efficient algorithms guided by our index are devised to provide optimal or near-optimal restoration plans considering both the mesoscale core-periphery structure and global network connectivity. Numerical results are provided to demonstrate the efficacy of our framework that remarkably enhances core-periphery restoration.