........ published in NEWSLETTER # 69
[ PCO Home Page ] [ Table of Contents of NEWSLETTER # 69 ]
........ published in NEWSLETTER # 69
COMPUTATIONAL INTELLIGENCE: SOFT COMPUTING AND FUZZY-NEURO INTEGRATION WITH APPLICATIONS
by Professor O. Kaynak, Bogazici University, Istanbul (Turkey)
Soft computing is a consortium of computing methodologies which provide a foundation for the conception, design, and deployment of intelligent systems and aims at the formalization of the remarkable human ability to make rational decisions in an environment of uncertainty and imprecision. It provides a better and more natural foundation for intelligent systems. The traditional "hard computing" (HC), based on binary logic, crisp systems, numerical analysis, and crisp software, has the characteristics of precision and categoricity, while soft computing has those of approximation and dispositionality. Although in the former, imprecision and uncertainty are undesirable properties, in the latter the toleranmce for imprecision and uncertainty is exploited to achieve tractability, lower cost, high machine intelligence quotient (MIQ), and economy of communication.
This book (NATO ASI SERIES F162) consists of papers that address the above issues. It is divided into six main parts, namely:
(i) computational intelligence, (ii) foundations of fuzzy theory, (iii) fuzzy systems, (iv) neural networks, (v) data analysis, and (vi) applications.
Each part has a number of papers authored by leading experts in the field. The first part starts with a paper by Prof. Zadeh himself in which he expresses his view on the roles of soft computing and fuzzy logic in the conception, design, and deployment of information/intelligent systems. The second paper is by Prof. Bezdek, addressing the controversial issue of what intelligence is, under the title "Computatinal Intelligence Defined - By Everyone!".
In the later parts of the book, the principal constituents of soft computing, namely fuzzy logic 8FL), neurocomputing (NC), genetic computing (GC), and probabilistic reasoning (PR), the relation between them and their fusion in industrial applications are discussed. A number of papers concentrate on the connections between neural networks and fuzzy logic. Use of vectoral neural networks in soft computing and modelling fuzzy reasoning by neural networks are among the topics of discussion. On the fuzzy logic side, fuzzy data analysis and fuzzy decision support systems are discussed in depth.
The last part of the book is devoted to the software and hardware architectures for soft computing and industrial applications. Computer vision is a major topic of discussion as an application area. A paper by Prof. Toshio Fukuda discusses the use of computational intelligence in robotics with emphasis on adaptation, learning and evolution, others discuss the state of art in manufacturing and fuzzy decision making. The final contribution to the book is on a product development system developed in NASA.
The conclusion that can be drawn from the papers in this book is that, using Zadeh's words, "soft computing is likely to play an important role in science and engineering, but eventually its influence may extend much further. In many ways, soft computing represents a significant paradigm shift in the aims of computing-a shift which reflects the fact that the human mind, unlike present day computers, possesses a remarkable ability to store and process information which is pervasively imprecise, uncertain and lacking in categoricity".
Reference books: C177, D38, F11, F136, F144, F162, F167, F168, G17