This course overviews decentralized and distributed control theory, developed from 1980’s in systems and control engineering. For analysis and synthesis of large-scale network systems, such as energy management systems and transportation systems, it is not realistic to apply a standard centralized control theory, intended mainly for mechanical systems. In particular, it is crucial to develop control design methods on the premise that multiple independent entities may perform the design and operation on their own responsibility. First, typical decentralized and distributed control problems with their solution methods are introduced in this course. Then, students learn a modelset-based control design method, relevant to dissipativity theory, in terms of, e.g., the Nyquist stability criteria in classical control theory. Furthermore, they learn basic principles of distributed optimization-based control and retrofit control as control design approaches on the premise of decentralized decision-making.
This course educates the concepts, problem formulations, solutions and basic principles for network control systems. Through this course, students are expected to learn typical control problems, applications and basic principles of network control systems. This would help students have access to the state-of-the-art in systems and control, necessary for new research on this topic. Students will acquire several background knowledge such as dissipativity theory, optimization theory and retrofit control theory.
Network control systems, Decentralized and distributed control, Multi-agent systems, Graph theory, Dissipativity theory, Finite-frequency analysis, Distributed optimization, Retrofit control theory
✔ Specialist skills | Intercultural skills | Communication skills | Critical thinking skills | Practical and/or problem-solving skills |
* Real-time lectures (Zoom) and on-demand lectures (video viewing) are given alternately.
* An exercise report is asked in every on-demand lecture.
* The exercise report is to be handed in at the next real-time lecture, based on which the attendance is to be checked.
* An explanation of the exercise is provided in the real-time lecture.
* Questions about the on-demand lecture are replied to also in the real-time lecture, which follows the on-demand lecture.
Course schedule | Required learning | |
---|---|---|
Class 1 | Network Control Systems | * Make sure what the course aims at by their learning portfolio. * Understand the basics of decentralized/distributed control by watching the first on-demand lecture and working on the exercise report. |
Class 2 | Modelset-Based Systems Analysis and Synthesis (1) | * Understand the basics of passivity by watching the second on-demand lecture and working on the exercise report. |
Class 3 | Modelset-Based Systems Analysis and Synthesis (2) | * Understand the basics of finite frequency analysis by watching the third on-demand lecture and working on the exercise report. |
Class 4 | Network Systems Design Based on Distributed Optimization | * Understand the basics of distributed optimization by watching the fourth on-demand lecture and working on the exercise report. |
Class 5 | Retrofit Control for Modularity-in-Design (1) | * Understand the basics of retrofit control by watching the fifth on-demand lecture and working on the exercise report. |
Class 6 | Retrofit Control for Modularity-in-Design (2) | * Understand the basics of retrofit control by watching the sixth on-demand lecture and working on the exercise report. |
Class 7 | Course Summary | * Summarize the contents of the course. |
Questions in an on-demand lecture are replied to in the real-time lecture that follows the on-demand lecture. Students are encouraged to organize their questions in advance.
Necessary documents are provided in the lecture.
Necessary documents are provided in the lecture.
* Students are assessed on their understanding of the concept of network control systems, theory, solution and their applications.
* The course scores are given based on the exercises, each of which is to be handed in at a real-time lecture.
* An exercise is asked in every on-demand lecture.
Students should have completed SCE.C.202, SCE.C.301, SCE.C.402, SCE.C.531, SCE.C.501, and SCE.C.502 or have equivalent knowledge.
tel: 03-5734-3179, email: ishizaki[at]sc.e.titech.ac.jp