Papers

A decision-focused learning framework that guides recursive linearization for fast, high-quality day-ahead scheduling of underground pumped hydro energy storage.

AuthorsHonghui Zheng, Pietro Favaro, Yury Dvorkin, Ján Drgoňa

A self-supervised learning-to-optimize framework for constrained bilevel control co-design with enforced coupling constraints.

AuthorsJames Kotary, Himanshu Sharma, Ethan King, Draguna Vrabie, Ferdinando Fioretto, Ján Drgoňa

A learning-to-optimize framework for parametric MINLP that combines adaptive integer correction with projections to guarantee feasibility.

AuthorsBo Tang, Elias B. Khalil, Ján Drgoňa

An end-to-end framework combining Time-Integrated DeepONets with offline, self-supervised optimization of PDE control policies.

AuthorsDibakar Roy Sarkar, Ján Drgoňa, Somdatta Goswami

A self-supervised mixed-integer control policy for nonlinear pumped-hydro scheduling, using differentiable simulation, Transformers, and Gumbel-Softmax decisions.

AuthorsHonghui Zheng, Ján Boldocký, Yury Dvorkin, Ján Drgoňa

A receding-horizon neural battery-dispatch controller with differentiable degradation physics and a safety filter that guarantees recursive feasibility.

AuthorsEshagh Safarzadeh Ravajiri, Ján Drgoňa, Benjamin F. Hobbs

A mixed-integer differentiable predictive control framework for real-time optimization of nonlinear multi-chiller plants.

AuthorsJán Boldocký, Cary Faulkner, Elad Michael, Martin Gulan, Aaron Tuor, Ján Drgoňa

Near-optimal control for input- and state-constrained parametric mixed-integer problems using differentiable predictive control.

AuthorsJán Boldocký, Shahriar Dadras Javan, Martin Gulan, Martin Mönnigmann, Ján Drgoňa

Decentralized neural-operator control policies for PDE systems that generalize zero-shot across different swarm cardinalities.

AuthorsPietro Zanotta, Dibakar Roy Sarkar, Honghui Zheng, Somdatta Goswami, Ján Drgoňa

Function Encoder-based neural ODEs integrated with differentiable predictive control for zero-shot control of nonlinear systems.

AuthorsHassan Iqbal, Xingjian Li, Tyler Ingebrand, Adam Thorpe, Krishna Kumar, Ufuk Topcu, Ján Drgoňa

A smooth, projection-free ODE formulation of Lasso with a prescribed convergence time that is independent of the problem data.

AuthorsLiang Wu, Yunhong Che, Wallace Gian Yion Tan, Efstathios Iliakis, Richard D. Braatz, Ján Drgoňa

A Koopman lifting framework that turns nonlinear safety-constrained MPC with discrete-time control barrier functions into an efficient quadratic program.

AuthorsShuo Liu, Liang Wu, Dawei Zhang, Ján Drgoňa, Calin A. Belta

A polynomial-chaos-parametrized Koopman model with a compact deterministic reformulation for efficient stochastic MPC of nonlinear systems.

AuthorsEfstathios Iliakis, Wallace Gian Yion Tan, Liang Wu, Ján Drgoňa, Richard D. Braatz

A Koopman-based NMPC formulation that yields structured box-constrained quadratic programs with certified execution times.

AuthorsLiang Wu, Yunhong Che, Bo Yang, Kangyu Lin, Ján Drgoňa

A variable-splitting nonlinear MPC solver designed for efficient parallel execution directly on system matrices.

AuthorsLiang Wu, Bo Yang, Junheng Li, Xu Yang, Yilin Mo, Yang Shi, Aaron D. Ames, Ján Drgoňa

A convex multi-step Koopman framework for long-horizon dynamics learning that avoids compounding single-step prediction errors.

AuthorsLiang Wu, Wallace Gian Yion Tan, Leqi Zhou, Richard D. Braatz, Ján Drgoňa

A predictor-corrector interior-point method with deterministic execution-time and iteration-complexity guarantees for real-time MPC.

AuthorsLiang Wu, Yunhong Che, Richard D. Braatz, Ján Drgoňa

A scalable statistical framework combining stochastic sampling and extreme value theory to certify regions of attraction in high-dimensional nonlinear systems.

AuthorsPietro Zanotta, Panos Stinis, Ján Drgoňa

A stability-enhanced Gaussian-process VAE that learns interpretable, semi-contracting latent linear dynamics from high-dimensional video data.

AuthorsCarl R. Richardson, Jichen Zhang, Ethan King, Ján Drgoňa

A framework that combines causal block diagrams, differentiable learning, and contract-based verification.

AuthorsThomas Beckers, Ján Drgoňa, Truong X. Nghiem

A tutorial overview of physics-informed machine learning methods for safe modeling and control of dynamical systems.

AuthorsJán Drgoňa, Truong X. Nghiem, Thomas Beckers, Mahyar Fazlyab, Enrique Mallada, Colin N. Jones, Draguna L. Vrabie, Steven L. Brunton, Rolf Findeisen

An operator-splitting integration scheme for learning neural differential-algebraic equation models from data.

AuthorsJames Koch, Madelyn Shapiro, Himanshu Sharma, Draguna Vrabie, Ján Drgoňa