CAREER: Exploiting Symmetry for Extreme-Scale Constrained Dynamical Systems

Automation and artificial intelligence are transforming society, increasing productivity by improving the speed, efficiency, and reliability of conducting complex tasks. Despite these advances, current approaches to autonomous decision-making require enormous computational resources, driving the expansion of data centers and placing growing strain on national electricity and water infrastructure. These high computational demands also limit where autonomous systems can be deployed, preventing low-cost or resource-limited applications from benefiting. This Faculty Early Career Development Program (CAREER) grant supports research that will create novel algorithms to dramatically reduce the computational demands of autonomous decision-making. The research will exploit symmetries, which are repeated patterns that frequently occur in large-scale human-engineered systems built from many similar components, including energy storage systems and logistics networks. By identifying and leveraging these patterns, this project will develop computational tools that enable symmetry-aware algorithms for extreme-scale autonomous decision-making, reducing both computing and memory requirements. Applications such as active battery balancing and resilient manufacturing logistics will strengthen national infrastructure by reducing the economic and environmental costs of computation. All algorithms and tools will be released as open-source software, broadening access and fostering innovation. Educational and outreach activities will promote interdisciplinary collaboration by bringing concepts from autonomy and artificial intelligence into curricula for students in majors outside control systems engineering. Project-based educational materials will be drawn from a broad range of real-world applications to engage students and demonstrate how these emerging tools are reshaping control and automation. This project will exploit symmetry to reduce the computational demands of extreme-scale autonomous decision-making. Extreme-scale refers to decision-making problems whose state and action spaces and constraint sets are so large, and whose decision times are so short, that conventional algorithms cannot meet real-time requirements on practical hardware. Central to this project is the notion of symmetry: transformations of the decision variables under which the decision problem remains invariant. These methods systematically identify and exploit these symmetries to reuse information across symmetric components rather than treating each as unique, enabling scalable autonomy with substantially lower computational and memory requirements. Although conceptually straightforward, efficiently exploiting symmetry in real-world applications is non-trivial and requires advances at the intersection of mathematics, optimization, and autonomous systems. The transformative potential of this research lies in applications such as vehicle electrification, smart grids, and resilient manufacturing, where it mitigates the growing economic and environmental costs of computation. 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. NSF Award ID: 2542350 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Claus Danielson | Institution: University of New Mexico, ALBUQUERQUE, NM | Award Amount: $619,800 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2542350 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2542350.html