Real-Time Optimization and Control Under Uncertainty

Real-Time Optimization and Control Under Uncertainty

We develop methods for optimization and control when the systems are subject to drifts, or to time-varying disturbances. We use iterative optimization-based methods, such as Bayesian optimization, Iterative Learning Control, MPC, and reinforcement learning.

Key topics:

• Modeling uncertainties in industrial processes
• Adaptive control strategies for dynamic environments (additive manufactuirng, fast motion stages)
• Real-time feedback and optimization loops
• Robustness under unpredictable changes and disturbances

Projects

• Meta-learning for Control: Manufacturing and Robotics Applications
• Adaptive Data-driven Optimization Algorithms for Manufacturing Components and Processing
• High-dimensional Optimization with Digital Twins

Funding

• SNF NCCR Automation

Publications

• Adaptive Bayesian Optimization for High-Precision Motion Systems, C König, R Krishnadas, EC Balta, A Rupenyan, IEEE Transactions on Automation Science and Engineering, 2025
• In-situ controller autotuning by Bayesian optimization for closed-loop feedback control of laser powder bed fusion process B Kavas, EC Balta, MR Tucker, R Krishnadas, A Rupenyan, J Lygeros, M. Bambach, Additive Manufacturing 99, 104641, 2025