Agent-Based Modeling of a Critical Infrastructure System: A Complex Adaptive Systems Perspective

摘要

As systems become more complex the probability increases that they will exhibit emergent behavior that could lead to system failures or widespread and prolonged service interruptions. This research presents findings on the behavior of complex systems after they experience disruptive events that impact their performance. A Complex Adaptive System (CAS) approach is used to conceptualize a complex network system that has been impacted by disruptions and perturbations. A combination of network and maximum flow interdiction analysis, as well as agent-based modeling, is used to measure performance of the system as it responds to disruptive events and restoration efforts. This system level response, an emergent property of the complex network, is represented as system resilience. Various resilience measures are used to quantify system resilience and assess the effectiveness of strategies system owners employ to restore the system. We illustrate our techniques by characterizing a Critical Infrastructure System (CIS) network as a CAS, and applying an agent-based simulation with an adaptive algorithm. The use of agent based modeling and simulation allows for the consideration of dynamic system interdependencies such as human interactions necessary to restore the network. This research effort is to provide engineers and designers of complex networked systems, as well as system owners and operators, a method to better investigate and understand the complex adaptive nature of their systems; and to fully assess the effectiveness of restoration strategies they employ to restore system performance after a disruptive event.