Multi-Agent Framework for the Complex Adaptive Modeling of Interdependent Critical Infrastructure Systems

This study presents a novel multi-agent system (MAS) framework for conceptualizing, modeling, and analyzing interdependent critical infrastructure (ICI) as complex adaptive systems (CAS) to enable resilience analysis. Unlike conventional approaches that consider interdependent infrastructure as networks of physical assets, the MAS framework conceptualizes interdependent infrastructure as interacting networks of social and physical entities and processes at different levels of hierarchical organization. The interaction between physical and social entities is captured at different levels of interaction proposed in System-of-System taxonomy and modified for modeling critical infrastructure systems and social agents. The proposed MAS framework is capable of abstraction, realization, and implementation of interdependent critical infrastructure systems as complex adaptive systems. The implementation of the proposed MAS framework is conducted through a nested high-level Petri-net method capable of: (1) abstraction of dynamic behaviors and interactions between social actors and physical networks, and (2) multi-level aggregation of dependencies in modeling critical infrastructure systems. The contributions of the proposed framework are threefold: (1) abstraction of dynamic behaviors and interactions between social and technical entities; (2) multi-level aggregation and implementation of interdependent critical infrastructure (ICI) networks as complex and layered systems; and (3) formalized modeling and simulation of infrastructure's multi-agent systems (MAS) using high-level petri net.