UQ-ACE: Stability-Aware Counterfactual Controller Tuning via Bayesian SINDy
Abstract
Retuning feedback controllers in critical processes is challenging when experiments are costly, open-loop excitation is infeasible, and candidate controllers must be screened before deployment. The existing Adaptive Sampling for Counterfactual Explanations (ACE) method provides a sample-efficient approach for finding minimal controller-parameter changes that recover prescribed closed-loop performance, but it does not explicitly account for stability of candidate controllers proposed during the search, which may be destabilizing under plant uncertainty. This paper introduces UQ-ACE, a stability-aware retuning method based on Bayesian SINDy. In UQ-ACE, closed-loop trajectories are used to infer a posterior distribution over local dynamical models, from which candidate controllers are screened for stability before implementation. Extensive simulations show that UQ-ACE achieves valid retunings with few closed-loop experiments while preventing deployment of unstable candidates.
Related
- Performance Evaluation of Metaheuristic Algorithms for Fractional-Order PID Controller Tuning in Non-Linear Systems
- A particle swarm optimization with a reduced-subspace trust-region safeguard for contact-rich controller tuning
- How Low Can You Go: Fine-Tuning LLMs with Bayesian Optimization (LlMBO)
- Risk-Constrained MPPI Tuning with Update-Aware MA-ES Racing
- Tuning-Free Covariance-Aware Sequential Shrinkage for Multi-Source Estimation
BibTeX
@misc{guerrero2026uqace,
title = {{UQ-ACE: Stability-Aware Counterfactual Controller Tuning via Bayesian SINDy}},
author = {Margarita A. Guerrero and Rodrigo A. Gonzalez and Cristian R. Rojas},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.18156},
eprint = {2610.18156},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.18156},
}
RIS
TY - UNPB
TI - UQ-ACE: Stability-Aware Counterfactual Controller Tuning via Bayesian SINDy
AU - Margarita A. Guerrero
AU - Rodrigo A. Gonzalez
AU - Cristian R. Rojas
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.18156
UR - https://arxiv.org/abs/2610.18156
AB - Retuning feedback controllers in critical processes is challenging when experiments are costly, open-loop excitation is infeasible, and candidate controllers must be screened before deployment. The existing Adaptive Sampling for Counterfactual Explanations (ACE) method provides a sample-efficient approach for finding minimal controller-parameter changes that recover prescribed closed-loop performance, but it does not explicitly account for stability of candidate controllers proposed during the search, which may be destabilizing under plant uncertainty. This paper introduces UQ-ACE, a stability-aware retuning method based on Bayesian SINDy. In UQ-ACE, closed-loop trajectories are used to infer a posterior distribution over local dynamical models, from which candidate controllers are screened for stability before implementation. Extensive simulations show that UQ-ACE achieves valid retunings with few closed-loop experiments while preventing deployment of unstable candidates.
ER -