Resilience of critical infrastructure systems is a key requirement in the vision of smart cities. These systems work continuously to enable the essential services such as water, gas, and electricity. They utilize diverse components organized as physical networks, and operated through heterogeneous and connected cyber elements. Many service utilities routinely face reliability concerns due to aging infrastructure, and lack the operational readiness that is needed to respond failures caused by natural disasters. Moreover, recent incidents have demonstrated that malicious entities can disrupt or gain control of these systems by exploiting cyber insecurities and/or physical faults. Indeed, sophisticated cyber intrusions and a number of successful physical attacks all confirm the insufficiency of the existing protection solutions. Such incidents can result in huge economic losses, and also pose threat to human lives. Since resiliency was not considered at the design stage of existing infrastructure systems, they continue to face significant risks from natural disasters and security attacks.
This talk is motivated by the need for a foundational approach for strategic security planning and operational response design, so that our infrastructure systems can better withstand, recover from, and adapt to both random and adversarial disruptions. The main agenda is to discuss how recently developed secure and distributed algorithms for network sensing and control can be implemented in practice to improve the resilience of large-scale infrastructure systems. These algorithms use ideas from control theory and large-scale optimization, along with game-theoretic analysis of strategic interaction between network operators and attackers. Through real-world case studies, we demonstrate that our algorithms can provide substantial improvements in strategic inspection and operational response capabilities of electricity and natural gas utilities facing risks of correlated disruptions.