ERI: Information Management Policies for Secure Resource Allocation

Ensuring the security of societal infrastructures such as power grids, public transit, and cyber-physical systems is a problem of critical importance. These systems are comprised of multiple components that are all subject to adversarial interference. The scale of these systems and limited defensive resources simply do not allow for complete security coverage at all times. Moreover, adequate security is increasingly reliant on a defender’s informational capabilities. Knowledge about an adversary’s abilities, objectives, and actions are essential pieces of information that can be used to mitigate the harm from potential attacks. This project investigates how the security of these systems can be maximized through the strategic management and control of informational advantages. Key questions to be addressed in this project are: What type of information is most valuable given costs for acquisition? How much more secure does the system become with critical pieces of information? Can obfuscating the system’s capabilities from an adversary render their attacks less effective? The research will develop a theoretical framework on which to address these questions and to inform effective and information-aware security policies that may be applied across domains. In particular, the project is well-positioned to interface with the highly active cybersecurity community in the Colorado Springs area. The research will also be leveraged to enhance curriculum and research opportunities for undergraduate students. This project seeks to analyze strategic decision-making in adversarial environments, with an emphasis on the role that informational capabilities play. These scenarios are most readily studied using tools from game theory and optimization. The research will leverage and adapt popular models of competitive resource allocation, known as “Colonel Blotto” and “General Lotto” games. The central goal of the project is to incorporate various elements of asymmetric information, and to investigate their impacts on the players’ optimal allocation strategies. This enables a principled approach to study the interplay between information and performance, as well as to guide the design of effective information-gathering and manipulation policies. The findings from this project will advance fundamental knowledge along two directions. First, they will provide new analyses regarding the strategic control of information in adversarial settings. Standard game-theoretic analyses primarily consider the informational environment to be static, i.e., unable to be altered through the efforts of the agents that it affects. Second, the project will develop principled approaches for system planners to implement security policies in a variety of environmental and adverse conditions. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.