The complex interaction between physical systems and people through cyber and economic layers has created new challenges in achieving efficiency while maintaining reliability in contemporary complex man-made systems.
Resilience or Systemic Risk are terms used to describe fragility in systems that results in cascades of failures due to either relatively small disturbances or even larger and malicious types of disruptions. The flash crash of 2010, the Great Recession of 2008, the New England power outage of 2003, or simply the 100 kilometer congestion in China that lasted more than 9 days are just a few of many examples of the systemic risk present in interconnected systems. Such disruptions have resulted in massive economic losses and decreased people’s confidence in such systems.
However, our ability to collect data from all layers in such interconnections presents an unprecedented opportunity to provide a data-driven platform, predicated on a new foundational science, for measuring, predicting, and containing systemic risk. Such development will emerge from an in-depth understanding of risk in financial and transportation systems.
We propose an integrative research program that combines expertise from these application domains, economics, operation research, system and control theory, probability and machine learning to provide a comprehensive solution to this problem. The research will address many of the overarching challenges associated with this problem such as the scarcity of failure data, the absence of full information and access to data by the many agents involved, the effects of the physical and social networks, the strategic behavior and lack of coordination between agents, and the absence of mechanistic models for human behavior. This holistic view will build on existing methods, our past research record, collaborations with domain experts, and various partnerships with industry and government.