Hierarchical Multi-objective Control of Nonlinear Systems with Dynamical Input Constraints

Authors

  • Ali Deeb Higher School of Cyber-physical Systems and Control, Peter the Great St. Petersburg Polytechnic University, Russia https://orcid.org/0000-0001-8369-1721
  • Vladimir Khokhlovskiy Higher School of Cyber-physical Systems and Control, Peter the Great St. Petersburg Polytechnic University, Russia https://orcid.org/0000-0002-5174-4363
  • Viacheslav Shkodyrev Higher School of Cyber-physical Systems and Control, Peter the Great St. Petersburg Polytechnic University, Russia https://orcid.org/0000-0001-6817-3440

DOI:

https://doi.org/10.47852/bonviewAIA52024314

Keywords:

multi-objective optimization, advanced control, boiler-turbine system, evolutionary machine learning, nonlinear control

Abstract

This work presents a multi-level hierarchical control strategy to address the problem of complex multi-objective optimization-based control in real time. Our suggested strategy utilizes evolutionary algorithms to solve the high-level optimization problem, providing a control policy under which a lower-level control loop handles the dynamics of the control values while respecting both regional and dynamical input constraints. Moreover, a real-time under-policy prediction phase is developed to absorb the latency of the computationally expensive policy search phase. The overall strategy is designed to leverage nonlinear systems without the need for further linearization or operating point approximations. Experimental results on a drum boiler-turbine unit simulation demonstrate the capabilities of our suggested strategy to steer the system outputs toward desired values with faster convergence compared to traditional methods. The proposed approach shows significant improvements in control performance, handling complex nonlinear control problems in real time, and providing optimized and fast control signals to guide the system outputs towards different operating points.

 

Received: 10 September 2024 | Revised: 8 February 2025 | Accepted: 24 February 2025

 

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

 

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

 

Author Contribution Statement

Ali Deeb: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization. Vladimir Khokhlovskiy: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing, Visualization, Supervision. Viacheslav Shkodyrev: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.


Metrics

Metrics Loading ...

Downloads

Published

2025-05-09

Issue

Section

Research Article

How to Cite

Deeb, A., Khokhlovskiy, V., & Shkodyrev, V. (2025). Hierarchical Multi-objective Control of Nonlinear Systems with Dynamical Input Constraints. Artificial Intelligence and Applications. https://doi.org/10.47852/bonviewAIA52024314