复杂网络能控性鲁棒性最优攻击序列研究

Study on the Optimal Attack Sequence for Controllability Robustness of Complex Networks

复杂网络的能控性在不同的攻击方式下会呈现不同的鲁棒性,寻找网络的最优攻击序列对复杂网络的能控性鲁棒性的保护和提升具有重要意义,本文使用遗传算法搜索针对复杂网络能控性的最优攻击序列,分析了在不同平均度情况下,遗传算法所求得的攻击序列和其他方法得到的攻击序列的特征,实验发现即使在不同类型和度分布的网络中,遗传算法生成的攻击序列都能取得比传统蓄意攻击方法更好的攻击效果在所用网络上,相比于其他蓄意攻击方法,遗传算法得到的最优攻击序列的节点度数排名更靠后,破坏性排名靠前;在不同类型的网络上,不同特征对最优攻击序列的重要程度不一样.

Complex networks performs different controllability robustness performances under different attack strategies.It is vital to search for the optimal attack strategy that causes the maximum destruction to network controllability,which is meaningful to protect or improve the controllability robustness of complex networks.In this paper,a genetic algorithm(GA)based aproach is used to search for the optimal sequence.Characteristics of the optimized attack sequences are analyzed and compared with those obtained by other methods under different average degrees.Experimental results show that the attack sequences optimized by GA are more destructive to controllability than other attack strategies on those networks have different topologies and different average degrees,Compared with other deliberate attack methods,the node degree of the optimal attack sequence obtained by the genetic algorithm ranks lower on the experimental network.Destructiveness ranks higher on the list of different methods.On different networks,different characteristics have different degrees of importance.