Sep 22, 2019   5:34 a.m. Móric
Academic information system

Final thesis assignment form – Ing. Karol Kiš – FCFT I-AICHP den [term 4, year 2]

This application helps you to display preview of the assignment of the final thesis. Preview is for information only and need not correspond to the print output.

Slovak University of Technology in Bratislava
Department of Information Engineering and Process Control
 Faculty of Chemical and Food Technology



Author of thesis:Bc. Karol Kiš
Study programme:Automation and Information Engineering in Chemistry and Food Industry
Study field: Cybernetics
Registration number:FCHPT-5414-76954
Student's ID:76954
Thesis supervisor:Ing. Martin Klaučo, PhD.
Title of the thesis:Machine Learning Approaches Applied to Generation of Explicit Control Laws
Language of thesis:English
Topic specifications:The main aim of this thesis is to analyze the possibilities of generating optimal explicit control laws (as explicit model predictive control - EMPC) for linear systems subject to state, input and output constraints. The challenge is to use a machine learning approach, specifically neural networks (NN), which can generate an explicit control law for higher order systems with large predictions horizons, which is not computationally tractable using EMPC generation techniques.

Individual tasks include:
1. Review of machine learning approaches which can substitute control laws.
2. Overview of neural networks and their applications as explicit controllers.
3. Simulation case studies and overview of NN-based controllers.
4. Experimental application and verification of NN-based controller on a laboratory process.
Length of thesis:60
  1. J. Drgoňa – D. Picard – M. Kvasnica – L. Helsen: Approximate model predictive building control via machine learning. Applied Energy, zv. 218, str. 199–216, 2018.
Date of entry:29. 03. 2019
Date of submission:12. 05. 2019

Ing. Karol Kiš
doc. Ing. Michal Kvasnica, PhD.
Head of department
 prof. Ing. Miroslav Fikar, DrSc.
Study programme supervisor