Network-Based Analysis of Dynamical Systems: Methods for Controllability and Observability Analysis, and Optimal Sensor Placement (SpringerBriefs in Computer Science)

✍ Scribed by Dániel Leitold, Ágnes Vathy-Fogarassy, János Abonyi

• Publisher: Springer
• Year: 2020
• Tongue: English
• Leaves: 119
• Category: Library

✦ Synopsis

This book explores the key idea that the dynamical properties of complex systems can be determined by effectively calculating specific structural features using network science-based analysis. Furthermore, it argues that certain dynamical behaviours can stem from the existence of specific motifs in the network representation.

Over the last decade, network science has become a widely applied methodology for the analysis of dynamical systems. Representing the system as a mathematical graph allows several network-based methods to be applied, and centrality and clustering measures to be calculated in order to characterise and describe the behaviours of dynamical systems.

The applicability of the algorithms developed here is presented in the form of well-known benchmark examples. The algorithms are supported by more than 50 figures and more than 170 references; taken together, they provide a good overview of the current state of network science-based analysis of dynamical systems, and suggest further reading material for researchers and students alike. The files for the proposed toolbox can be downloaded from a corresponding website.

✦ Table of Contents

Preface
Contents
Symbols and Acronyms
Symbols
Acronyms
1 Introduction
1.1 Network Science-Based Analysis of Complex Systems
1.2 Formalisation of the Network-Based Controllability and Observability Analysis
1.3 Recent Trends in Network-Based Dynamical System Analysis
1.4 Motivation and Outline of the Book
References
2 Structural Controllability and Observability Analysis in Complex Networks
2.1 Critical Evaluation of the Methodology
2.2 Connections Between the State Variables and Their Effect on Controllability and Observability
2.3 Benchmark Examples
2.4 Effect of the Connection Types
2.5 Discussion
References
3 Reduction of Relative Degree by Optimal Control and Sensor Placement
3.1 Problem of Sensor Placement
3.2 Relative Degree-Based Cost Functions
3.3 Fast and Robust Output Configuration Design
3.3.1 Centrality Measures and Set Covering-Based Methods
3.3.2 Simulated Annealing and Fuzzy Clustering-Based Methods
3.4 Sensor Placement Case Studies
3.4.1 Description of the Case Studies
3.4.2 Performance of the Optimisation Algorithms
3.4.3 Convergence Analysis
3.5 Discussion
References
4 Application to the Analysis of Heat Exchanger Networks
4.1 Importance of Process Integration
4.2 Network-Based Complexity and Operability Analysis of HENs
4.3 Studied Benchmark Problems
4.3.1 Analysis of the 9sp-al1 Problem
4.3.2 Correlation Between Structural and Dynamical Properties
4.4 Discussion
References
5 The NOCAD MATLAB/Octave Toolbox Developed for the Analysis of Dynamical Systems
5.1 Existing Tools for Network Analysis
5.2 The Workflow Related to the Application of the Toolbox
5.3 Structure of the Toolbox
5.4 Use Cases
5.5 Discussion
References
6 Summary
6.1 Conclusion
Appendix A Effect of Connection Types
A.1 The Path-Finding Method
A.2 The Studied Networks
A.3 Calculated Measures of the Benchmark Networks
Appendix B Reduction of Relative Degree by Additional Drivers and Sensors
B.1 Details of the Case Studies
B.2 Network Representations of the Case Studies
B.3 Convergence of the Solutions
Index