“Advancing AI and Automation via Learning-based Online Optimization and Control”
Monday, Feb. 26 at 1:00pm
LAR 234
Add to Calendar
Abstract
Technological advances in artificial intelligence (AI) and automation aim to enhance autonomy, efficiency, and reliability in modern cyber-physical systems. A key goal is to empower them with decision-making capabilities to respond and adapt to environmental and operational changes, often without requiring human intervention. In this context, the presentation begins by providing an overview of the main research themes within Dall’Anese’s group, which are focused on advancing foundational methods for optimization and control of cyber-physical and network systems. Subsequently, the talk delves into innovative approaches for developing and analyzing learning-enabled optimization and control architectures by leveraging and seamlessly integrating tools from online optimization, mathematical control theory, dynamics, and supervised learning. A key ingredient for these new architectures is to convert principled optimization algorithms into feedback controllers that dynamically and autonomously steer the inputs and outputs of a system towards optimal solutions of well-posed optimization problems modeling performance, safety, and operational goals. These feedback optimization methods enable an optimal system operation in spite of stochastic conditions, unknown disturbances, and uncertain human-system interactions.
Architectures that are augmented with supervised learning methods are then presented, with the objective of unlocking the opportunities for perception-based and data-based optimization and control schemes. In particular, it is shown how learning methods—from the traditional least squares to deep neural networks—can be utilized to estimate problem inputs, system parameters, and the system state from data. Benefits and impact of these methods in the context of modern power systems and healthcare applications are discussed.
Biography
Dr. Emiliano Dall’Anese is an Associate Professor in the Department of Electrical, Computer, and Energy Engineering (ECEE) at the University of Colorado Boulder, where he is also an affiliate faculty member with the Department of Applied Mathematics. He received the Ph.D. in information engineering from the Department of Information Engineering, University of Padova, Italy, in 2011. He was with the University of Minnesota as a postdoc (2011-2014) and the National Renewable Energy Laboratory as a senior researcher (2014-2018). His research interests span the areas of optimization, control, and learning, with applications in energy systems and cyber-physical systems. He received the National Science Foundation CAREER Award in 2020, the IEEE PES Prize Paper Award in 2021, the Outstanding Faculty Researcher in the ECEE Department in 2022, the IEEE Transactions on Control of Network Systems Best Paper Award in 2023, and three additional best journal papers recognitions. He is currently serving as an Associate Editor for IEEE Control Systems Letters.