Operational Resilience to Extreme Events in Networked Dynamical Systems through Real-time Adaptation
Abstract
This proposal will develop foundational theory and algorithms for ensuring real-time operational resilience of networked dynamical systems to extreme events that push the system far-from-equilibrium. Networked dynamical systems such as infrastructure networks, supply chains, biological networks, social networks, and soon, are central to our lives. Yet, they often fail catastrophically when faced with unexpected disturbances. Methods that can realize operational resilience in networked dynamical systems by allowing them to respond to inevitable disturbances and shocks by sensing and adapting, and learn from such experiences to become better at responding over time by reasoning about them, will be transformative in a number of domains. However,many factors make operational resilience for networked dynamical systems challenging beyond stand-alone systems.These includethe need for algorithms to be hierarchical and distributed, the need to consider the network structure and physical properties-constraints, the difficulty of obtaining system-level guarantees with local controllers, and computational complexity as the network size increases.This proposal will solve these challenges by integrating perspectives and tools from systems and control theory, physics, game theory,data science,and statistics. The intellectual bulwark of this proposal is inter-disciplinary research to realize networked dynamical systems that achieve operational resilience through real-time adaptation to changes in operating conditions, especially during extreme events--socks that take systems very far from the equilibria or nominal regimes under which they are typically designed to operate. The proposed approach will (i) identify control-relevant physics-constrained models of networked dynamical systems in far-from-equilibrium regimes, and (ii) design model-based distributed controllers that adapt to changing operating conditions to prevent cascading failures and guarantee continued performance and safety under large disturbances.The proposal will develop a unifying framework that integrates advances in these themes. Our approach can enable networked dynamical systems to move beyond a dichotomy of efficiency vs.robustness that is baked in during design and instead lead to systems that adapt to large disturbances and become better at adaptation over time.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310492
Entities
People
- Sivaranjani Seetharaman
Organizations
- Air Force Office of Scientific Research
- Office of the Secretary of Defense
- Purdue University