Title
Improving Resilience of Complex Networks Facing Attacks and Failures through Adaptive Mechanisms.
Abstract
Studies have revealed that some topological properties of complex networks are robust to random node failures, but highly sensitive to failures in nodes of high centrality (i.e., attacks). This work proposes mechanisms based on local information for detecting vulnerable network configurations and for promoting changes in the topology to mitigate the impact on network connectivity of successive node losses due to attacks and failures. Two scenarios were evaluated: with maintenance of the number of links in the network, and with the creation of new links. For the first case, we show that the removal of the most central nodes affects networks in such a way that is also difficult to preserve their main topological properties, although improvement in the values of global efficiency at the expense of a reduction of the local efficiency was observed. For the second scenario, there was a significant decrease in the impact from attacks and failures. Notably, for failures the connectivity properties not only remained almost unchanged, but in some cases considerably increased, thus improving the overall network performance. The results were also verified in some benchmark real-network topologies, and a comparative performance evaluation with a random self-regenerating process was also analyzed. The study demonstrates the importance and feasibility of local adaptation mechanisms for link rewiring based on the concept of vulnerability.
Year
DOI
Venue
2014
10.1142/S021952591450009X
ADVANCES IN COMPLEX SYSTEMS
Keywords
Field
DocType
Failure and attack tolerance,adaptive mechanisms,network connectivity
Psychological resilience,Network connectivity,Centrality,Network topology,Complex network,Mathematics,Distributed computing,Network performance,Vulnerability
Journal
Volume
Issue
ISSN
17
2
0219-5259
Citations 
PageRank 
References 
1
0.36
4
Authors
2
Name
Order
Citations
PageRank
Cinara Guellner Ghedini192.90
Carlos H. C. Ribeiro290.90