Robust Networks to Cascading Failures

Abstract

Nowadays, the ongoing progress of networking in essential utilities such as the Internet, the WWW, transportation networks, electrical power grid networks, etc., brings significant benefits to the quality of our life. However, networked systems hold a certain danger that a failure of a single node in the system may diffuse to all other nodes. This chain of failure is widely known as cascading failure. Examples of cascading failure include disease epidemics, traffic congestion, electrical power system blackouts, and so on. In these systems, if external shocks or excess loads at some nodes are propagated to other connected nodes due to failure, the domino effects often come with disastrous consequences. Thus, how to prevent cascading failures in complex networks becomes an important emergent issue. In this paper, we consider an overload-based cascading failure model and design a robust network structure against this type of cascading failure. Numerical simulations show that the proposed network, which consists of a complete cluster of connected hub nodes, and periphery nodes connecting one of hub nodes in the cluster, is least susceptible to cascading failures compared with other types of networks.