Answers are grounded in this paper and cite their sources.
The performance of predictive control strategies often degrades over time due to growing plant-model mismatch. Closed-loop performance restoration typically requires some form of model maintenance to reduce model uncertainty. This paper presents a stochastic predictive control approach with integrated experiment design for nonlinear systems with probabilistic modeling uncertainties. The integration of predictive control with experiment design enables enhancing the information content of closed-loop data for online model adaption. The presented approach considers control-oriented experiment design to ensure adequate model adaptation (in probability) in terms of an admissible control performance level. The stochastic optimal control approach is demonstrated on a continuous bioreactor case study.
No takes yet. Share an insight, caveat, or question.
Bavdekar et al. (2016) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: