Reasoning Under Uncertainty in Cyber-Physical Systems: Toward Efficient and Secure Operation

Abstract:

The increased sensing, processing, communication, and control capabilities introduced by cyber-physical systems bring many potential improvements to the operation of society's systems. With these improvements come many questions as to how one can ensure their efficient and secure operation. This dissertation investigates three questions related to decision-making under uncertainty in cyber-physical systems settings: (1) In the context of power systems and electricity markets, how can one obtain socially-optimal outcomes, while obeying the physics of the system, when the information necessary for making the optimal decision is distributed across many, potentially self-interested, agents? (2) When a system is under attack from a malicious agent, what models are appropriate for performing real-time and scalable threat assessment and response selection when we only have partial information about the attacker's intent and capabilities? (3) In partially observable sequential decision-making environments, specifically partially observable Markov decision processes, under what conditions do optimal policies possess desirable structure? The answers to these questions involve investigating the structure of how information exists in the system and properties of how it is revealed to decision-makers, in order to design models and algorithms that can efficiently translate the available information into decisions, while enabling tractable computation in realistic domains.