Speaker: Dr. Jan Drgona

Topic: Scientific Machine Learning for Modeling, Optimization, and Control

Abstract: This talk presents a control-oriented perspective on Scientific Machine Learning (SciML) for modeling, optimization, and control of dynamical systems. SciML provides a unifying computational paradigm that integrates physics-based models, optimization algorithms, and control policies within a differentiable programming framework. This synthesis enables computation of structured gradients for constrained system identification, learning-to-optimize, and learning-based control while preserving interpretability, stability, and physical consistency.

Three recent advances will be highlighted. First, differentiable predictive control, a SciML approach that merges model predictive control with gradient-based learning to enable scalable, self-supervised training of explicit control policies suitable for real-time deployment on embedded hardware. Second, an operator-splitting formulation for neural differential-algebraic equations that integrates mechanistic dynamics with neural components to achieve robust extrapolation in systems with implicit constraints and conservation laws. Third, a self-supervised learning-to-optimize framework for mixed-integer nonlinear programs that provides feasibility guarantees and high-quality approximate solutions in milliseconds.

Together, these advances demonstrate how SciML can unlock new capabilities for the modeling, optimization, and control of complex dynamical systems, with applications in power grid and building energy management.

Biography: Ján Drgoňa is an associate professor in the department of Civil and Systems Engineering and a member of the Ralph S. O’Connor Sustainable Energy Institute.

His innovative research centers on differentiable programming and scientific machine learning (SciML) for dynamical systems, optimization, and control. He has particular experience in the deployment of machine learning and advanced control methods for real-world applications, including building energy systems and industrial process control.

Previously, Drgoňa was a principal investigator and research data scientist at Pacific Northwest National Laboratory (PNNL) where he served as the lead software developer of Neuromancer SciML library for learning to solve constrained optimization, physics-informed machine learning, and optimal control problems. Within two years, the library became the most popular open-source repository released by PNNL.

Drgoňa is a member of the Institute of Electrical and Electronics Engineers and the Association for Computing Machinery. He regularly serves as a reviewer for related journals including Applied Energy, Automatica, IEEE Control Systems Letters, IEEE Transactions on Control Systems Technology, IEEE Transactions on Industrial Informatics, Control Engineering Practice, Journal of Process Control, Energy and Buildings, Journal of Control Automation and Electrical Systems, and Electric Power Systems Research.

Drgoňa earned his BSc, MSc, and PhD in control engineering from the Institute of Information Engineering, Automation, and Mathematics at the Slovak University of Technology. Prior to his work with PNNL, he held a postdoctoral position in the Mechanical Engineering Department at KU Leuven in Belgium.