Data-driven switching modeling for MPC using Regression Trees and Random Forests. (May 2020)

• Record Type: Journal Article
• Authors: Smarra, Francesco; Di Girolamo, Giovanni Domenico; De Iuliis, Vittorio; Jain, Achin; Mangharam, Rahul; D'Innocenzo, Alessandro
• Is Part Of: Nonlinear analysis. Volume 36(2020)
• Publication Date: 2020-05
• Subjects: Regression Trees -- Random Forests -- Model predictive control -- Switching systems -- Markov Jump Systems; Nonlinear functional analysis -- Periodicals; Analyse fonctionnelle non linéaire -- Périodiques; Nonlinear functional analysis; Periodicals; 515.7248
• DOI: 10.1016/j.nahs.2020.100882
• Languages: English
• ISSNs: 1751-570X

Abstract

Abstract: Model Predictive Control is a well consolidated technique to design optimal control strategies, leveraging the capability of a mathematical model to predict a system's behavior over a time horizon. However, building physics-based models for complex large-scale systems can be cost and time prohibitive. To overcome this problem we propose a methodology to exploit machine learning techniques (i.e. Regression Trees and Random Forests) in order to build a Switching Affine dynamical model (deterministic and Markovian) of a large-scale system using historical data, and apply Model Predictive Control. A comparison with an optimal benchmark and related techniques is provided on an energy management system to validate the performance of the proposed methodology.