Efficient Automatic Tuning for Data-driven Model Predictive Control via Meta-Learning

arXiv:2404.00232v1 Announce Type: cross
Abstract: 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 …

abstract arxiv bayesian control cs.lg cs.ro data data-driven however meta meta-learning optimization package paper portfolio predictive python search spaces type via