Fuzzy and Neural Approaches in Engineering

by Tsoukalas, Lefteri H.; Uhrig, Robert E.; Zadeh, Lotfi A.

• ISBN13: 9780471160038
• ISBN10: 0471160032
• Edition: 1st
• Format: Hardcover
• Copyright: 1997-02-05
• Publisher: Wiley-Interscience

Summary

Neural networks and fuzzy systems represent two distinct technologies that deal with uncertainty. This definitive book presents the fundamentals of both technologies, and demonstrates how to combine the unique capabilities of these two technologies for the greatest advantage. Steering clear of unnecessary mathematics, the book highlights a wide range of dynamic possibilities and offers numerous examples to illuminate key concepts. It also explores the value of relating genetic algorithms and expert systems to fuzzy and neural technologies.

Author Biography

LEFTERI H. TSOUKALAS, PhD, is on the faculty of the School of Nuclear Engineering at Purdue University and is an active industrial consultant and speaker. ROBERT E. UHRIG, PhD, holds a joint appointment as Distinguished Professor in the Nuclear Engineering Department at the University of Tennessee and Distinguished Scientist in the Instrumentation and Control Division at the Oak Ridge National Laboratory. He is the author of Random Noise Techniques in Nuclear Reactor Systems.

Table of Contents

Foreword	xiii
Preface	xvii
Introduction to Hybrid Artificial Intelligence Systems	1	(10)
Introduction	1	(1)
Neural Networks and Fuzzy Logic Systems	2	(1)
The Progress in Soft Computing	3	(2)
Intelligent Management of Large Complex Systems	5	(2)
Structure of this Book	7	(1)
Problems and Programs Available on the Internet	8	(3)
References	9	(2)
I FUZZY SYSTEMS: CONCEPTS AND FUNDAMENTALS	11	(178)
Foundations of Fuzzy Approaches	13	(36)
From Crisp to Fuzzy Sets	13	(2)
Fuzzy Sets	15	(2)
Basic Terms and Operations	17	(11)
Properties of Fuzzy Sets	28	(2)
The Extension Principle	30	(4)
Alpha-Cuts	34	(3)
The Resolution Principle	37	(1)
Possibility Theory and Fuzzy Probabilities	38	(11)
References	45	(1)
Problems	46	(3)
Fuzzy Relations	49	(28)
Introduction	49	(3)
Fuzzy Relations	52	(5)
Properties of Relations	57	(3)
Basic Operations with Fuzzy Relations	60	(5)
Composition of Fuzzy Relations	65	(12)
References	74	(1)
Problems	75	(2)
Fuzzy Numbers	77	(28)
Introduction	77	(2)
Representing Fuzzy Numbers	79	(5)
Addition	84	(6)
Subtraction	90	(5)
Multiplication	95	(4)
Division	99	(2)
Minimum and Maximum	101	(4)
References	102	(1)
Problems	103	(2)
Linguistic Descriptions and Their Analytical Forms	105	(40)
Fuzzy Linguistic Descriptions	105	(8)
Linguistic Variables and Values	113	(7)
Implication Relations	120	(5)
Fuzzy Inference and Composition	125	(11)
Fuzzy Algorithms	136	(9)
References	141	(1)
Problems	142	(3)
Fuzzy Control	145	(44)
Introduction	145	(6)
Fuzzy Linguistic Controllers	151	(12)
Defuzzification Methods	163	(13)
Issues Involved in Designing Fuzzy Controllers	176	(13)
References	185	(2)
Problems	187	(2)
II NEURAL NETWORKS: CONCEPTS AND FUNDAMENTALS	189	(218)
Fundamentals of Neural Networks	191	(38)
Introduction	191	(1)
Biological Basis of Neural Networks	192	(1)
Artificial Neurons	193	(3)
Artificial Neural Networks	196	(7)
Learning and Recall	203	(8)
Features of Artificial Neural Networks	211	(2)
Historical Development of Neural Networks	213	(8)
Separation of Nonlinearly Separable Variables	221	(8)
References	227	(1)
Problems	227	(2)
Backpropagation and Related Training Algorithms	229	(60)
Backpropagation Training	229	(5)
Widrow-Hoff Delta Learning Rule	234	(4)
Backpropagation Training for a Multilayer Neural Network	238	(10)
Factors That Influence Backpropagation Training	248	(7)
Sensitivity Analysis in a Backpropagation Neural Network	255	(2)
Autoassociative Neural Networks	257	(9)
An Alternate Approach to Neural Network Training	266	(4)
Modular Neural Networks	270	(4)
Recirculation Neural Networks	274	(5)
Functional Links	279	(1)
Cascade-Correlation Neural Networks	280	(1)
Recurrent Neural Networks	281	(8)
References	285	(2)
Problems	287	(2)
Competitive, Associative, and Other Special Neural Networks	289	(44)
Hebbian Learning	289	(1)
Cohen-Grossberg Learning	290	(6)
Associative Memories	296	(10)
Competitive Learning: Kohonen Self-Organizing Systems	306	(9)
Counterpropagation Networks	315	(4)
Probabilistic Neural Networks	319	(6)
Radial Basis Function Network	325	(1)
Generalized Regression Neural Network	326	(2)
Adaptive Resonance Theory (ART-1) Neural Networks	328	(5)
References	331	(1)
Problems	332	(1)
Dynamic Systems and Neural Control	333	(52)
Introduction	333	(1)
Linear Systems Theory	333	(8)
Adaptive Signal Processing	341	(4)
Adaptive Processors and Neural Networks	345	(8)
Neural Network Control	353	(10)
System Identification	363	(5)
Implementation of Neural Control Systems	368	(6)
Applications of Neural Networks in Noise Analysis	374	(6)
Time-Series Prediction	380	(5)
References	382	(1)
Problems	383	(2)
Practical Aspects of Using Neural Networks	385	(22)
Selection of Neural Networks for Solution to a Problem	385	(1)
Design of the Neural Network	386	(9)
Data Sources and Processing for Neural Networks	395
Data Representation	391	(4)
Scaling, Normalization, and the Absolute Magnitude of Data	395	(4)
Data Selection for Training and Testing	399	(2)
Training Neural Networks	401	(6)
References	405	(2)
III INTEGRATED NEURAL-FUZZY TECHNOLOGY	407	(114)
Fuzzy Methods in Neural Networks	409	(36)
Introduction	409	(1)
From Crisp to Fuzzy Neurons	410	(4)
Generalized Fuzzy Neuron and Networks	414	(2)
Aggregation and Activation Functions in Fuzzy Neurons	416	(2)
AND and OR Fuzzy Neurons	418	(3)
Multilayer Fuzzy Neural Networks	421	(2)
Learning and Adaptation in Fuzzy Neural Networks	423	(8)
Fuzzy ARTMAP	431	(3)
Fuzzy-Neural Hybrid Data Representation	434	(3)
Survey of Engineering Applications	437	(8)
References	440	(2)
Problems	442	(3)
Neural Methods in Fuzzy Systems	445	(26)
Introducing the Synergism	445	(2)
Fuzzy-Neural Hybrids	447	(3)
Neural Networks for Determining Membership Functions	450	(5)
Neural-Network-Driven Fuzzy Reasoning	455	(6)
Learning and Adaptation in Fuzzy Systems via Neural Methods	461	(5)
Adaptive Network-Based Fuzzy Inference Systems	466	(5)
References	468	(2)
Problems	470	(1)
Selected Hybrid Neurofuzzy Applications	471	(22)
Introduction	471	(1)
Neurofuzzy Interpolation	472	(2)
General Neurofuzzy Methodological Developments	474	(2)
Engineering Applications	476	(1)
Diagnostics in Complex Systems	477	(1)
Neurofuzzy Control Systems	478	(3)
Neurofuzzy Control in Robotics	481	(1)
Pattern Recognition and Image Enhancement	482	(1)
Medical and Environmental Imaging Using Neurofuzzy Methodologies	483	(1)
Transportation Control	484	(1)
Adaptive Fuzzy Systems	485	(1)
Inspection Using Neurofuzzy Methods	486	(1)
Neurofuzzy Methods in Financial Engineering	486	(1)
Commercial Neurofuzzy System Software	487	(6)
References	488	(5)
Dynamic Hybrid Neurofuzzy Systems	493	(28)
Introduction	493	(2)
Fuzzy-Neural Diagnosis for Vibration Monitoring	495	(5)
Decision Fusion by Fuzzy Set Operations	500	(4)
Hybrid Neurofuzzy Methodology for Virtual Measurements	504	(6)
Neurofuzzy Approaches to Anticipatory Control	510	(11)
References	516	(5)
IV OTHER ARTIFICIAL INTELLIGENCE SYSTEMS	521	(46)
Expert Systems in Neurofuzzy Systems	523	(16)
Introduction	523	(1)
Characteristics of Expert Systems	524	(1)
Components of an Expert System	525	(2)
Knowledge Representation and Inference	527	(2)
Uncertainty Management	529	(2)
State of the Art of Expert Systems	531	(1)
Use of Expert Systems	532	(2)
Expert Systems Used with Neural Networks and Fuzzy Systems	534	(1)
Potential Implementation Issues for Expert Systems	535	(4)
References	537	(1)
Problems	538	(1)
Genetic Algorithms	539	(22)
Introduction	539	(1)
Basic Concepts of Genetic Algorithms	540	(2)
Binary and Real-Value Representations of Chromosomes	542	(2)
Implementation of Genetic Algorithm Optimization	544	(2)
Fitness Functions	546	(6)
Application of Genetic Algorithms to Neural Networks	552	(2)
Fuzzy Genetic Modeling	554	(2)
Use of Genetic Algorithms in the Design of Neural Networks	556	(5)
References	557	(2)
Problems	559	(2)
Epilogue	561	(6)
Introduction	561	(1)
Is Artificial Intelligence Really Intelligent?	562	(1)
The Role of Neurofuzzy Technology	563	(1)
Last Thoughts	564	(3)
References	565	(2)
Appendix: T Norms and S Norms	567	(8)
Index	575