Distributed location of the critical nodes to network robustness based on spectral analysis

We propose a methodology to locate the most critical nodes to network robustness in a fully distributed way. Such critical nodes may be thought of as those most related to the notion of network centrality. Our proposal relies only on a localized spectral analysis of a limited neighborhood around each node in the network. We also present a procedure allowing the navigation from any node towards a critical node following only local information computed by the proposed algorithm. Experimental results confirm the effectiveness of our proposal considering networks of different scales and topological characteristics.