Randomized algorithms for analysis and control of uncertain systems. With applications (Q2276433)

The book focuses on the fundamentals of probabilistic and randomized methods for analysis and design of uncertain systems. It introduces a new approach to control problems that presents a combination of the classical stochastic treatment and the modern worst-case paradigm. In this setting, the worst-case aspects of the uncertainty are described as random variables with multidimensional probability distributions. Consequently, the algorithms derived in this context are based on the uncertainty randomization are called randomized algorithms. The book pays a considerable attention to efficiency of these algorithms. The book includes 19 chapters, an appendix, a thorough and very helpful list of symbols, an extended bibliography (a few hundreds of references, and a rather short index. Basic elements of probability theory are summarized in Section 2. The main topics of the book are the following: uncertain systems (Chapters 3--5), probabilistic methods (Chapters 6 and 7), statistical learning theory (Chapters 8 and 9), randomized algorithms for design (Chapters 10--13), multivariate random generation, and applications of randomized algorithms (Chapter 19). Chapter 19 discusses many applications: multiagent systems and PageRank computation; systems biology; aerospace control; control of hard disk drives; congestion control of high-speed networks; stability of quantized and switched systems; reachability analysis; model predictive control, fault detection and iterative learning control; electric circuits and embedded systems; structural analysis and design; linear parameter varying systems; automotive and driver assistance systems; and quantum systems and control. Six of the applications are considered in the corresponding sections of the chapter in detail. The book covers a wide range of the topics related to randomized algorithms and includes a number of rigorous results, in particular, on stability and design of linear uncertain systems, linear robust control design, and statistical approaches including statistical theories of random vectors and matrices. The new paradigm proposed in the book is well explained. The book can be recommended to mathematicians working in pure mathematics or application of mathematics to uncertain systems.