Optimal attack strategy with heterogeneous costs in complex networks

Abstract

The problem of network disintegration, such as suppressing the epidemic spreading and destabilizing terrorist networks, has broad applications and recently has received growing attention. We first presents a cost constrain model for attack strategy in complex networks, in which the attack cost for a node is heterogeneous. Then we introduce the unequal probability sampling into the network disintegration problem and present an optimization model for optimal attack strategy with limited attack costs in complex networks. We investigate the influence of cost-sensitive parameter and cost-constraint parameter on the optimal attack strategy in both model networks and real-world networks. We find that the low-degree nodes are attacked preferentially when the cost constraint is tight; however, the high-degree nodes are attacked preferentially when the cost constraint is relax. Moreover, we find there is a threshold for the cost-sensitive parameter, the high-degree nodes won't be attacked preferentially if the cost of single node is more than a threshold.