Study On Optimization Of Network Robustness Under Attacks

As Internet technology flourishes and it is widely applied in our daily life,endless events of attacks are reported,such as information leakage of bank’s customers,hot sale of Lenovo’s pad in wrong low price,fault of root server and so on.So it is of great concern to strengthen the protection of the networks and build resistant networks under attacks.Complex network can depict the structure characteristics and typical dynamical behaviors of real complex systems.Research on attacks and optimization of complex network have become a hot point in relevant fields.Under the general background of complex network,integrating mathematical statistics,control,numerical simulations and the same kind of subjects,research efforts have been laid on “how to improve the network’s resilience”,“the relationship between network’s resilience and synchronizability” and “the effect of clustering coefficient on interdependent network’s structural resilience”.The innovative results are listed as follows.(1)Optimization of network resilience under attacks based on Simulated AnnealingBased on enhancing network resilience under attacks,degree-preserving cross-link rewiring is adopted to adjust the topological structure.Then,the problem of optimization of network resilience is transformed into a single-objective combinatorial optimization problem.The optimization model is established,in which the node robustness index is used as the objective function.An optimization algorithm based on Simulated Annealing is proposed,aiming at searching for the optimal solution in the global scope.Lastly,numerical simulations are investigated to verify the effectiveness of the proposed optimization method.Also,during optimization the change of several important structural parameters is recorded and analyzed.(2)Reduced synchronizability of dynamical scale-free networks with onion-like topologiesFirstly,some steps are adopted to improve network robustness,finding that the optimized networks topology presents with ‘‘onion-like’’,namely high-degree nodes form a core surrounded by rings of nodes with decreasing degree.Then we investigate the synchronization behavior of dynamical scale-free onion networks.Large quantities of numerical simulations indicate that solely enhancing the network robustness can lead to the reduction of the ability to achieve synchronization.Furthermore,two important structural properties of ‘‘onion-like’’topologies,the increase of average path length and the increase of assortative coefficient,are analyzed and considered to be responsible for the decease of synchronizability.(3)Effect of clustering on attack vulnerability of interdependent scale-free networksClustering coefficient is one of important parameters affecting network’s performance,then comes up with interdependent scale-free network model with tunable clustering coefficient and adopts a more realistic attacking strategy---targeted attack aiming at nodes whose degreesare much higher.The effects of coupling strength and coupling preference on attack vulnerability are explored.Numerical simulation results demonstrate that interdependent links between two networks make the entire system much more fragile to attacks.Also,it is found that clustering significantly increases the vulnerability of interdependent scale-free networks.Moreover,for fully coupled network,disassortative coupling is found to be most vulnerable to random attacks,while the random and assortative coupling have little difference.Additionally,enhancing coupling strength can greatly enhance the fragility of interdependent networks against targeted attacks.