Agent-Based Framework for Assessing Systemic Risk of Interdependent Sociotechnical and Infrastructure Systems

Sociotechnical systems consisting of infrastructures, businesses, and households are constantly expanding and evolving due to, among others, rapidly progressing economic development, urbanization, and globalization. These complex interdependent cybernetic systems become increasingly prone to both natural and man-made disruptions. Such complex systems exhibit emergent effects in response to any disruption which propagates throughout those systems. Consequently, using traditional methods of risk assessment of individual systems is insufficient to predict the emergent systemic impacts of disruptive events of the future. Hence, there is an urgent need to develop models assessing systemic risk of such complex interdependent sociotechnical systems. The aim of the presented study was to (1) present and apply a framework for modeling interdependencies between complex sociotechnical systems at different levels of detail, with a particular focus on urban areas; (2) develop disruption generators and devise the method of assessing impact of disruptions on the systems; and (3) demonstrate how the modeling framework can be applied to assess systemic risks. Our study resulted in the development of a simulation combining socioeconomic agents’ models, such as households and businesses, with physical infrastructure systems models to assess systemic risks, reliability and safety associated with interdependencies in such sociotechnical systems. The disruption cost served as a measure of performance of the system used to assess the risk. Our model was shown to capture the emergent behavior of complex sociotechnical systems in response to disruptions. Based on our analysis, the most risk was associated with low-reconfigurability infrastructure systems, particularly disruptions to the water supply system, resulted in largest impacts. We also identified areas for future research that focus on including a wider range of systems, more accurate disruption generators, and on applying the presented modeling framework to other complex networks and sociotechnical systems.