|
Columnist |
Since
I last discussed fuzzy references, a slew of new fuzzy-logic books has been
published. Fuzzy books are coming out so quickly now, I am often reluctant to
look in the artificial-intelligence sections of local technical bookstores for
fear that I'll leave with another $300 charge to one of my credit cards. I just
saw two new ones but resisted the temptation to buy—maybe for Christmas.
Because
of the large number of new books, I can't discuss all that have been published,
and I won't try. I can't even review all the good ones. For example, because it
is not directly aimed at engineers, you may not find Earl D Cox's excellent
second book, Fuzzy Logic for Business and Industry (Reference
1) at technical bookstores, although its particulars are listed at
the end of the column. This column is limited to four books and a tutorial
article.
I lead off with the book I like best, Fuzzy Logic with Engineering Applications (ISBN 0-13-368465-2, $71.50) by Timothy J Ross (Reference 2). The first seven chapters present a detailed introduction to fuzzy sets and logic. The next seven chapters apply fuzzy logic to rule-based systems, nonlinear simulation, decision making, classification, pattern recognition, control systems, and other miscellaneous topics. The final chapter covers fuzzy measures as they are used in evidence systems and probability, possibility, and necessity measures.
This book has a classroom flavor; if you have volunteered to teach a fuzzy-logic course, this should be your text. The explanations and examples are good and are aimed at senior-level undergraduates or first-year graduate students. As with most teaching texts, it provides problems at the end of each chapter. These problems are not overly difficult, and working through them tends to flesh out a reader's understanding of fuzzy-logic theory and solution techniques.
Each chapter has loads of references—the author is definitely aware of what other researchers in the field are up to. The book also contains case studies, particularly for the chapter on control.
Overall, Ross' book is a complete rendering of fuzzy theory and application. Read it thoroughly, do a comfortable sampling of the examples and problems, and you will find yourself far down the road to applying many fuzzy-based methods to engineering problems.
The next book is Fuzzy Logic & NeuroFuzzy Applications (ISBN 0-13-368465-2, $39.95) by Constantin von Altrock (Reference 3). Von Altrock is an engineer with the German company, Inform (Aachen, Germany). He has been involved in a sizable number of fuzzy-development projects. If you have attended fuzzy presentations at technical conferences, you have probably heard von Altrock speak.
His book is a good, engineering-oriented introduction to fuzzy-rule-based design; case studies; and Inform's fuzzy-logic tool, fuzzyTECH. (An included disk provides a demo version.)
The first four chapters are introductory, including a primer, a discussion of development tools, and a short overview of neural nets and neuro-fuzzy systems. Chapter 5 is an excellent presentation of industrial-application case studies. The final two chapters discuss techniques for developing a fuzzy-rule-based system and an introduction to the fuzzyTECH demo.
I particularly like the case studies. There are approximately 15 case studies with detailed descriptions and 50 studies with short descriptions and pointers to supporting references. I also like the author. I have known von Altrock for about five years and have always enjoyed his solid engineering background and his experience with fuzzy-rule-based systems.
My third selection is Industrial Applications of Fuzzy Logic and Intelligent Systems (ISBN 0-7803-1-48-9, $64.95) edited by John Yen, Reza Langari, and Lotfi A Zadeh (Reference 4). Yen and Langari are professors at the Center for Fuzzy Logic and Intelligent Systems at Texas A&M University (College Station, TX). Since at least 1991, they have expended considerable energy to bring fuzzy-logic theory and concepts into industry. They are comfortable and conversant in both worlds. Zadeh, the father of fuzzy logic, has been similarly devoted for even longer. Together, the three have come up with a strong selection of application-oriented papers, written by researchers and practicing engineers alike.
The book is organized as five parts, containing a total of 15 chapters. The parts are a one-chapter introduction to fuzzy-rule-based control by Yen and Langari; five chapters that discuss fuzzy-logic applications in Japan; seven chapters on fuzzy-logic applications in the United States; a single chapter on fuzzy applications in Europe written by von Altrock; and a review of software- development tools for fuzzy control by Stephen Chiu of the Rockwell Science Center (Thousand Oaks, CA).
The chapters are well-written and convey design and implementation experiences from a diverse collection of fuzzy-control projects. My favorite chapter, however, is Chiu's review of development tools, in which he discusses seven tools. I have varying experience with five of these tools and have never attempted to write tool reviews in this column because of the time necessary to learn each well enough to make a credible comparison. My hat is off to Chiu, who has done just that. Technology moves quickly, however, and since the book was published, two of the vendors have closed their doors in the Unites States, and a major new player, The Mathworks (Natick, MA), has entered the market with the introduction of its Matlab Fuzzy Logic Toolkit.
Coming down the home stretch, my final book recommendation is Fuzzy Logic for Real World Design (ISBN 0-929392-24-8, $44.95) by Ted and Jill Heske, a husband-and-wife team (Reference 5). It is the most recently published of the four books discussed here. The first seven chapters deal with fuzzy-rule-based design, including basic theory and how that theory is applied to designing operational systems. Chapters 8 and 9 address using fuzzy rule bases in an object-oriented architecture. The final chapter is a primer on using FuzzyLab (the disk is included), which is described as an "interactive program featuring a fuzzy- control system and animated simulation, written with the enthusiastic beginner in mind."
My favorite part of this book is Chapter 5, entitled "Fuzzy Engineering II." It presents an excellent discussion of what goes into tuning a fuzzy-rule-based controller. This is material that individuals with a good number of fuzzy projects under their belts will have discovered (hit or miss) on their own but that I have seen presented nowhere else.
I
also highly recommend the glossary, especially to newcomers. When we first
become interested and excited about a new field, it is often the vocabulary,
typically familiar words with strange, specialized meanings, that either keeps
us from rapidly moving forward in our learning or completely turns off our
interest. Armed with the Heskes' glossary, you'll find any introductory book to
fuzzy systems much more readable.
The final entry is not a book, but an article: "Fuzzy Logic Systems for Engineering: A Tutorial," by Jerry M Mendel (Reference 6). A number of readers have asked that I include it in this review.
Mendel's article is an introduction to fuzzy theory, sets, operators, linguistic variables, hedges, fuzzy logic, and implication. It contains numerous reference and points to other, typically academic and research-oriented papers. The article is a good introduction, but keep in mind that the intended audience is students, academics, and researchers. It tends to be heavier on supporting theory than on application examples. A number of corrections, mostly of minor typographical and printing errors, appeared six months later (Reference 7).
That's it. All four books and Mendel's article are worthy additions to your Zadehan-logic library. I realize that, by not being able to discuss all new fuzzy books, I may have missed a favorite of yours. If so let me know. As always, all comments are welcome.
| EDN Access | feedback | subscribe to EDN! | Table of Contents |