Industrial control and supervision tasks have many features that are not adequately addressed by traditional methods of automated system development, such as investigation of previously unstructured problems; processing of integrated symbolic and mathematical information; use of expert knowledge to achieve efficient decision making; an ability to process incomplete, uncertain, or imprecise data; an ability to learn and adapt to new situations; and an ability to structure and use (“understand”) audiovisual and natural language information. One approach to providing these features has been intelligent systems. The intelligent system methodologies illustrated in this book are neural networks, fuzzy logic, rule-based systems, pattern recognition, various learning algorithms, and some hybrid forms.

The book is divided into two parts. Part 1 includes the first three chapters and addresses the fundamentals and methodologies of intelligent systems. The first chapter gives a general introduction to modern control technologies; the system environment and uncertainty; and intelligent systems. Chapter 2 introduces the concepts, architectures, and algorithms of fuzzy logic, neural networks, and rule-based systems. Chapter 3 focuses on the learning strategies and algorithms.

Part 2 includes five chapters that cover various applications of intelligent systems in industrial control. Chapter 4 introduces math-model-free (but not math-free) techniques of process modeling and estimation. Chapter 5 investigates dynamic controls based on neural networks, fuzzy logic, and rules. Expert optimization control and production scheduling are covered in chapter 6. Chapter 7 introduces quality control and statistical process control with multivariate statistics. The last chapter investigates the applications of fuzzy logic and neural networks in fault detection and diagnosis. A number of industrial applications are investigated in each chapter of Part 2.

The book creates two important connections: the link between the industrial control and intelligent system approaches, and the link between the resulting theory and industrial practice. The architectures, models, and algorithms of industrial control systems are integrated with those resulting from intelligent systems. This combination is not always easy to read. To make the book more practical and applicable, many industrial examples from oil refineries, steel mills, and other applications are described in detail. The book includes a bibliography and author and subject indexes.

An important topic--safety and quality of intelligent control systems--seems to be missing from the book. Reliability of intelligent methodologies is not a straightforward issue, and it should be given full attention in industrial systems, which are often subject to major financial and environmental hazards. A few typos, such as “1.5 Introduction and knowledge” in the contents, instead of “Introduction to intelligent control,” do not seriously affect the book’s readability.

The book requires a background in industrial control and mathematics, especially differential equations and algebra. It can be used as a reference by graduate students and faculty at engineering schools, or as a textbook for practicing engineers, researchers, and consultants in the control industry. It may also give intelligent systems specialists valuable ideas for combining math and non-math approaches, and for practical applications of theory.