A Data-Driven Automatic Tuning Method for MPC under Uncertainty using\n Constrained Bayesian Optimization

2020/11/23 by Farshud Sorourifar, Sorourifar, Farshud, Georgios Makrygirgos +5 · 6 citations
Engineering · #Advanced Control Systems Optimization #FOS: Electrical engineering #Fault Detection and Control Systems #Process Optimization and Integration #Systems and Control (eess.SY) #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2011.11841

openalex publication_date 2020/11/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

The closed-loop performance of model predictive controllers (MPCs) is\nsensitive to the choice of prediction models, controller formulation, and\ntuning parameters. However, prediction models are typically optimized for\nprediction accuracy instead of performance, and MPC tuning is typically done\nmanually to satisfy (probabilistic) constraints. In this work, we demonstrate a\ngeneral approach for automating the tuning of MPC under uncertainty. In\nparticular, we formulate the automated tuning problem as a constrained\nblack-box optimization problem that can be tackled with derivative-free\noptimization. We rely on a constrained variant of Bayesian optimization (BO) to\nsolve the MPC tuning problem that can directly handle noisy and\nexpensive-to-evaluate functions. The benefits of the proposed automated tuning\napproach are demonstrated on a benchmark continuously stirred tank reactor\nexample.\n

Cited by

• Integrated Online Monitoring and Adaptation of Process Model Predictive Controllers
• Fine-Tuning of Neural Network Approximate MPC without Retraining via Bayesian Optimization
• Direct transfer of optimized controllers to similar systems using dimensionless MPC
• Economic zone data-enabled predictive control for connected open water systems
• Model Predictive Control with Reference Learning for Soft Robotic Intracranial Pressure Waveform Modulation
• Model-free Reinforcement Learning for Model-based Control: Towards Safe, Interpretable and Sample-efficient Agents

Related

• Model Predictive Control Tuning by Monte Carlo Simulation and Controller Matching
• MPC Controller Tuning using Bayesian Optimization Techniques
• Stability-informed Bayesian Optimization for MPC Cost Function Learning
• Data-Driven Scenario Optimization for Automated Controller Tuning with\n Probabilistic Performance Guarantees
• Closed-loop Performance Optimization of Model Predictive Control with Robustness Guarantees