Supervisory output prediction for bilinear systems by reinforcement learning

Authors Georgios C. Chasparis
Thomas Natschläger
Editors
Title Supervisory output prediction for bilinear systems by reinforcement learning
Type article
Journal IET Control Theory & Applications
Number 10
Volume 11
DOI 10.1049/iet-cta.2016.1400
ISSN 1751-8644
Month June
Year 2017
Pages 1514-1521
SCCH ID# 1471
Abstract

Online output prediction is an indispensable part of any model predictive control implementation. For several application scenarios, operating conditions may change quite often, while designing the data collection process may not be an option. To this end, this paper introduces a supervisory output prediction scheme, tailored specifically for input-output stable bilinear systems, that intends on automating the process of selecting the most appropriate prediction model during runtime. The selection process is based upon a reinforcement-learning scheme, where prediction models are selected according to their prior prediction performance. An additional selection process is concerned with appropriately partitioning the control-inputsaAZ domain in order to also allow for switched-system approximations of the original bilinear dynamics. We show analytically that the proposed scheme converges (in probability) to the best model and partition. We also demonstrate these properties through simulations of temperature prediction in residential buildings.