Optimizing network topology for cascade resilience

Abstract

Complex networks need resilience to cascades – to prevent the failure of a single node from causing a far-reaching domino effect. Such resilience can be achieved actively or topologically. In active resilience schemes, sensors detect the cascade and trigger responses that protect the network against further damage. In topological resilience schemes, the network’s connectivity alone provides resilience by dissipating nascent cascades. Designing topologically resilient networks is a multi-objective discrete optimization problem, where the objectives include resisting cascades and efficiently performing a mission. Remarkably, terrorist networks and other “dark networks” have already discovered how to design such networks. While topological resilience is more robust than active resilience, it should not always be pursued because in some situations it requires excessive loss of network efficiency.