Extremum seeking in control: Methods, Theory and Applications
25 Apr 2023 - 27 Apr 2023
Extremum seeking control (ESC) has grown to become the leading approach to solve realtime optimization problems. Following 20 years of research developments, this strikingly general and practically relevant control approach is now equipped with an established and well understood control theoretical framework. The objective of this course is to provide a detailed introduction to the most recent fundamental developments in this field.
Practical information:
25 Apr 2023 - 27 Apr 2023
UCLouvain - Euler building (room A.002) Avenue Georges Lemaître, 4 - 1348 Louvain la Neuve
English
Target audience: Researchers
Objectives
The focus of the course is on the design of ESC systems using various leading methodologies which include classical perturbation based methods, estimation based methods and Lie-bracket averaging techniques.
- The first objective of the course is to present recent contributions to the design of fast ESC systems that can overcome the performance limitations of the classical approaches. The advent of fast techniques have sparked developments in new areas of applications such as distributed optimization, model-free control, observer design and deep learning.
- The second objective of the course is to present many important generalizations of ESC design techniques. These include the design of ESC for:
- multivariable systems
- systems subject to actuator limitations
- delay systems
- optimal output regulation
- constrained optimization problems
- The third objective of the course is to present several emerging applications of ESC in energy systems, biotechnology and power systems.
Outline of the course
- Basic ESC Loop: Analysis and Performance
- Fast ESC: Methods and Analysis
- ESC Generalizations: Challenges and Methods
- ESC in practice: Applications and Case Studies
- Beyond ESC: Adaptation and learning