Intelligent control. A stochastic optimization based adaptive fuzzy approach (Q1791999)

The book presents systematic designs of stable adaptive fuzzy logic controllers employing hybridizations of contemporary stochastic optimization techniques and the ``classic'' one, namely, either the Lyapunov strategy-based approach or the \(H^{\infty}\) strategy-based robust approach. Thus, the hybridization process attempts to combine strong points of local search methods and the stochastic optimization-based global search methods to evolve superior methods. Several novel variants of the hybridization have been presented in the book. The book begins with a Preface, which is quite informative; it lists the key features of the book, and briefly describes the contents of the book in chapters. Moreover, at the end of the Preface, various paths for reading a book are proposed depending on the interests of the reader. The Preface is followed by five parts of different sizes, two of which are Prologue and Epilogue. The latter in volume and capacity is consistent with the title, while the former is a rather voluminous introduction to the contents of the book. The first chapter briefly, but not superficially, paves the ``way to the modern control theory'' (this is the title of subsection 1.1); an impressive list of references is attached to it (191 points, from 1963 until the most recent publications). The second and third chapters briefly describe four stochastic optimization algorithms, as well as methods for constructing a fuzzy controller based on stochastic algorithms. These three chapters make up Part I: Prologue. The combination of stochastic optimization with classical methods of control theory is presented in the second and third parts of the book. The structure of these parts is similar. Both of them consist of two chapters, the first one presents the application of the corresponding approach to the construction of fuzzy controllers (of the Lyapunov method in Part II: Lyapunov strategy-based design methodologies, of \(H^{\infty}\) control theory in Part III: \(H^{\infty}\) strategy-based design methodologies), and the second one presents the methods by the authors. In both cases, three different variants of hybridizations are presented, namely, cascade model, concurrent model, and preferential model. Part IV: Applications (chapters 8, 9 and 10) is devoted to applications of independent interest: temperature control of an air heater system with transportation delay, vision-based navigation of mobile robots, control of robot manipulators. Part V: Epilogue (chapter 11) of this work pulls a lot of weight in just two pages, with authors managing to reference 36 sources and proposing multiple directions for developing the theory and ironing out problems. Each part is accompanied by its own list of references. The last in the book is Index. The book's dedication ``to our beloved students'' comes to mind multiple times throughout the reading: First, the book is well-structured and well-suited for familiarizing yourself with it. Second, the delivery is rigorous, clear and hands-on. Methods described are supported with algorithm listings, schemata and flowcharts, aiding greatly in understanding and comparison of said methods. Authors concisely present the necessary minimum, referring to additional reading for details, with book list at the end of every chapter. Third, throughout all chapters dedicated to algorithms and methods (chapters 3 to 7) authors illustrate their usage on two models, using helpful visual aids, going into detail on different cases, and laying out all the necessary parameters. One could end the book at that, since models do a great job at demonstrating described methods' efficiency; but after that, authors get into what they call ``the most interesting part''. Part IV, rather than focusing on illustrative examples, concerns itself with three different real-world applications, for each of them going the whole way from modeling the problem to applying obtained results to the real-world object. At the beginning of each chapter, the authors highlight ``the chief problem'' that justifies the use of the techniques proposed by the authors. The book appears to be fond of the ``rule of three'': three authors, three main body parts (excluding prologue and epilogue, but not for the lack of content), three methods, and three reviewed applications. Overall, this work is a great confluence of mathematical soundness, computer science and engineering applications, making it a good fit for either a mathematics undergraduate's library, a reseacher's desk, or a computer engineer's toolbox.