A holistic method for reliability performance assessment and critical components detection in complex networks

Many infrastructures are now considered to be critical for both the economic development and general functioning of modern societies. Thus, understanding their performance is important as a basis to develop intelligent and cost-effective ways to protect these networks. In this article, a critical infrastructure is modeled as a complex network for which a new metric is defined to understand its reliability. This metric called reliability Π describes the average reliability between every pair of nodes in a complex network. As such, it is related to the two-terminal reliability concept in the traditional network context. Furthermore, in an effort to identify the most critical components that affect reliability Π, a multi-objective optimization problem, known as the critical component detection problem, is introduced. The solution to this problem provides two important insights about the behavior of a complex network: (i) an approximation to the set of optimal solutions that identifies the most critical components; and (ii) a quantitative assessment of how these failures affect the complete complex network.