Minimum entropy control for time-varying systems (Q2565152)

This book is aimed at researchers in mathematical control theory. Using concepts from operator theory, the authors provide a time domain notion of entropy for general classes of systems, which is equivalent to the usual frequency domain notion for linear time-invariant systems. They consider minimum entropy control problems. Presenting its connections with other optimal control problems, they show some advantages of entropy and find minimum entropy controllers. The book is organized as follows. Chapter 1 is an introduction. Chapter 2 (Preliminaries) deals with preliminary results, especially the operator theoretical description of discrete-time systems. In chapter 3 (Induced operator norms), the authors give characterizations for the induced norms of both discrete-time and hybrid linear systems in terms of operator Riccati equations and study their properties. In chapter 4 (Discrete-time entropy), the definition of entropy is given and its properties are studied. In chapter 5 (Connecting with related optimal control problems), the following topics are presented: The relationship with \(H_\infty\) control, with \(H_2\) control, average cost functions and risk-sensitive control. In chapter 6 (Minimum entropy control), the authors consider minimum entropy control problems for linear discrete-time systems. They show that minimum entropy controllers provide a simple means of trading off robustness and performance and find a solution via the separation principle. In chapter 7 (Continuous-time entropy), the authors give the definition of entropy for continuous-time systems and consider problems similar to those of the previous chapters.