Managing Uncertainty Using CISIApro 2.0 Model

Critical infrastructures are large and complex systems showing physical, geographical, cyber, and logical interdependencies. Assessing the consequences of faults, cyber-attacks, or natural events on critical infrastructures is fundamental to mitigate the impact of these events and to improve the resilience of the entire “system of systems”. Modeling critical infrastructures is a way to understand how networks are interconnected and increase awareness. Our paper proposes a well-established framework for decomposing each infrastructure into its meaningful elements and their interconnections using the Mixed Holistic-Reductionist approach. CISIApro 2.0 is an agent-based simulator that implements the previous methodology. In this paper, we improve the framework using possibility theory to handle epistemic uncertainty. In contrast to randomness, epistemic uncertainty is caused by incomplete, ambiguous, or inaccurate knowledge. This is a problem in the modeling of critical infrastructures, due to the complexity of the physical process, the vagueness of telecommunications, and imprecise information. Including possibility and necessity measures in the CISIApro 2.0 simulator allows us to handle uncertainty in the propagating information in the model. The synthetic model is composed of eight interconnected infrastructures. The model output demonstrates the reasonableness of the proposed approach, although there is a need for further improvements, especially with regard to parameter tuning.