Keywords: critical infrastructures, infrastructure interdependencies, vulnerabilities, network analysis, giant components, NetLogo, modelling, simulation, critical components, multi–agent systems, MAS, agent–based systems, node failures, infrastructure protection, recovery strategies
Identifying vulnerabilities in critical infrastructures by network analysis
By modelling the relationships and how disruption cascades within a complex infrastructure network, vulnerabilities can be identified and recovery/response efforts planned. A simulation approach is applied in this paper to identify critical components in a given infrastructure system represented as a graph comprising nodes and links. The network is modelled using NetLogo, a multi–agent programming environment. One– and two–node attack scenarios are programmed to investigate the impact of different node failures in the network, with the ensuing disruption quantified by a giant component metric. In the one–node attack, the minimum, final and average sizes of the giant component unanimously identified the same critical node while the two–node attack simulation identified different node pairs as critical. Which pair should be considered most critical will depend on policy goals to be met. This simulation approach can aid policymakers in deciding which infrastructures require higher protection and how recovery strategies should be planned.