Cargando…

Enhancing structural robustness of scale-free networks by information disturbance

Many real-world systems can be described by scale-free networks with power-law degree distributions. Scale-free networks show a “robust yet fragile” feature due to their heterogeneous degree distributions. We propose to enhance the structural robustness of scale-free networks against intentional att...

Descripción completa

Detalles Bibliográficos
Autores principales: Wu, Jun, Tan, Suo-Yi, Liu, Zhong, Tan, Yue-Jin, Lu, Xin
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2017
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5548747/
https://www.ncbi.nlm.nih.gov/pubmed/28790416
http://dx.doi.org/10.1038/s41598-017-07878-2
Descripción
Sumario:Many real-world systems can be described by scale-free networks with power-law degree distributions. Scale-free networks show a “robust yet fragile” feature due to their heterogeneous degree distributions. We propose to enhance the structural robustness of scale-free networks against intentional attacks by changing the displayed network structure information rather than modifying the network structure itself. We first introduce a simple mathematical model for attack information and investigate the impact of attack information on the structural robustness of scale-free networks. Both analytical and numerical results show that decreasing slightly the attack information perfection by information disturbance can dramatically enhance the structural robustness of scale-free networks. Then we propose an optimization model of disturbance strategies in which the cost constraint is considered. We analyze the optimal disturbance strategies and show an interesting but counterintuitive finding that disturbing “poor nodes” with low degrees preferentially is more effective than disturbing “rich nodes” with high degrees preferentially. We demonstrate the efficiency of our method by comparison with edge addition method and validate the feasibility of our method in two real-world critical infrastructure networks.