Real networks are often subjected to failures and attacks. Real networks are found to be robust to errors and vulnerable to attacks. This behavior of real networks is attributed to their non-homogeneous degree distribution. Non-homogenous networks are also known as scale free networks. Real networks along with the scale free property show high modularity.; Many network models have been proposed to model real networks. Erdos-Renyi Random network model is the first attempt, but fails to incorporate both properties - scale free as well as the high modularity of the real networks. Small world model shows high clustering but lacks non-homogeneous distribution. The scale free network model has non-homogeneous degree distribution but lacks the modularity. In 2002, Ravasz and Barabasi proposed a Hierarchical Network model that combines non-homogeneous degree distribution as well as high modularity showed by real networks.; The objective of this research is to study the error and attack tolerance of different network models. The static as well as dynamic tolerances of attacks are analyzed. The effect of an attack on the network model is quantified by considering the dynamic flows of quantities and using the impact factors. The results of the study show that though the scale free as well as the hierarchical models are vulnerable to attacks, the performance can be highly secured by protecting some key nodes.
展开▼
