基于网络模体特征攻击的网络抗毁性研究

Network Destruction Resistance Based on Network Motif Feature

真实网络如生物神经网络、食物链网络中含有模体结构。但对于这种含有模体特征的网络,并没有一种针对性的攻击策略与之对应,在此基础上提出了模体度代数算法和一种模体攻击失效方式,设计了不同于传统攻击的模体攻击策略。对已检验出的具有明显模体特征的2个仿真网络,5个不同规模的实证网络利用该策略进行模体攻击,并与传统的点攻击方式进行对比研究。仿真结果表明:含有模体特征的网络在模体攻击下的抗毁性明显低于传统的点攻击策略,且在模体特征较明显的网络中模体攻击策略对网络的破坏性更加显著。

Real networks such as biological neural networks,food chain networks contain motif structure.But for this kind of network with the motif characters,there is no corresponded attack strategy.In this paper,we propose a motif degree algebraic algorithm and a motif attack failure mode for this kind of network with the characteristics of the motif structure.We design a model attack strategy which is different from the traditional attack.By using this strategy,two simulation networks and five different size real networks with distinct motif characteristics are simulated and compared with the traditional point attack strategy.The simulation results show that the survivability of the network with the motif feature under the motif attack strategy is obviously lower than traditional point attack strategy.And the motif attacking strategy is more significant to the network with obvious motif characteristics.