National Science Foundation Expeditions in Computing for Computational Decarbonization of Societal Infrastructures at Mesoscales

University Of Massachusetts Amherst, Amherst MA

Investigators

Abstract

Our society stands at a critical inflection point with its rapidly accelerating demand for energy due to growth in domestic manufacturing, datacenters, artificial intelligence (AI), electric vehicles, and electric heat pumps. Maintaining this growth will require us to focus on improving energy-efficiency. The goal of this Expedition is to develop new computational techniques to unify, transform, and accelerate efficiency across multiple use domains. To this end, the Expedition focuses on optimizing and reducing the lifecycle costs of complex computing and societal infrastructure systems. Specifically, the CoDec Expedition will address an important class of problems that arise from interdependencies across multiple infrastructure domains, including computing, transportation, the built environment, and the electric power grid. Over five years, the Expedition will develop novel computational techniques, algorithms, systems, and AI methods that sense, optimize, and reduce the operational, embodied, and lifecycle impact on societal infrastructure over long temporal and spatial scales. This work will enable new scientific discoveries while ensuring the sustainable growth of our energy consumption that is necessary for advancing technology, growing our economy, and strengthening national security. To prepare our nation’s workforce in this emerging field, the Expedition will develop new educational programs and design new teaching curricula. CoDec will establish an industry consortium to share its research results and transition them to practice. It will also undertake a range of activities to broaden participation in computing through undergraduate research and K-12 outreach. Since the decarbonization of complex systems can be viewed as a sense-optimize-reduce problem that operates over long times scales of years or decades and over large geographical scales, it will require new types of optimization techniques and algorithms that are designed for these mesoscales. To address the problem, CoDec will develop novel (i) sensing approaches to provide visibility into systems' efficiency, (ii) optimization methods grounded in theory and AI to exploit new dimensions of flexibility, which are emerging in modern infrastructure, and (iii) software-defined interfaces and systems for programmatically deploying optimizations. These techniques will be verifiable, robust to uncertainties in energy's supply and demand, and transferable across multiple domains and economic sectors through intuitive interfaces, while also resolving human-in-the-loop concerns of privacy, fairness, equity, incentives, and user preference. 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.