Efficient Automatic Tuning for Data-driven Model Predictive Control via Meta-Learning
This work addresses efficiency and stability issues in data-driven model predictive control tuning for control systems, but it is incremental as it builds on existing AutoMPC and BO methods.
The paper tackled the computational expense and instability of AutoMPC when using Bayesian Optimization for tuning, by proposing a meta-learning approach called Portfolio that warmstarts BO with optimized initial designs from previous tasks, resulting in improved efficiency and stability on 12 benchmarks.
AutoMPC is a Python package that automates and optimizes data-driven model predictive control. However, it can be computationally expensive and unstable when exploring large search spaces using pure Bayesian Optimization (BO). To address these issues, this paper proposes to employ a meta-learning approach called Portfolio that improves AutoMPC's efficiency and stability by warmstarting BO. Portfolio optimizes initial designs for BO using a diverse set of configurations from previous tasks and stabilizes the tuning process by fixing initial configurations instead of selecting them randomly. Experimental results demonstrate that Portfolio outperforms the pure BO in finding desirable solutions for AutoMPC within limited computational resources on 11 nonlinear control simulation benchmarks and 1 physical underwater soft robot dataset.