Computational Intelligence PC Tools
by Russell C. Eberhart, Roy W. Dobbins, and Patrick K. Simpson
Copyright 1996 Academic Press Professional
Goals and Target Audience
Computational Intelligence PC Tools is targeted for senior undergraduate
and first-year graduate students in engineering and computer science.
It is also written for engineers, computer scientists, and others who
want to learn about computational intelligence and its practical
applications. At the end of most chapters, exercises are presented that
test the understanding of the material presented, and that give the
opportunity to expand knowledge of computational intelligence tools
beyond the chapter conte nts. Executable code for major paradigm
implementations is provided on a diskette with the book. Source code
(written in C/C++) and additional executable code is available from the
Table of Contents
Presents the goals and objectives of the text, and describes its
intended audience. Presents an outline of the book, with a brief
description of each part, chapter and appendix. Discusses use of the
text in a course.
Chapter 1: Background
Discusses background of neural networks, fuzzy logic and evolutionary
computation, and how they form the basis for the unified field called
computational intelligence (CI).
Chapter 2: History
Reviews history of CI component technologies, and how they have been
combined to form computational intelligence. The focus is on people,
rather than just on theory or technology.
Chapter 3: Neural Network Theory and Paradigms
Reviews the concepts associated with neural networks. Included are
terminology and symbology, and discussions of the three features of any
neural network: architecture, processing element transfer functions, and
learning algorithms. A number of network paradigms are discussed and
compared. Preprocessing and postprocessing are examined.
Chapter 4: Neural Network Implementations
Presents implementations of back-propagation, learning vector
quantization, and radial basis function networks. (Software is on the
Chapter 5: Evolutionary Computation Theory and Paradigms
Reviews the four basic methodologies of evolutionary computation.
Included are genetic algorithms, evolutionary programming, evolution
strategies and genetic programming.
Chapter 6: Evolutionary Computation Implementations
Includes implementations of a Genesis-like genetic algorithm and a
particle swarm optimizer. (Software is on the diskette.)
Chapter 7: Fuzzy Logic Theory and Paradigms
Discusses fuzzy reasoning, the basics of fuzzy set theory, and fuzzy
membership functions. Presents examples of fuzzy reasoning systems
using both Mamdani and Takagi-Sugeno methods. Examines measures of
Chapter 8: Fuzzy Logic Implementation
Presents an implementation of a fuzzy expert system. The implementation
has both linear and nonlinear membership functions, and various
fuzzification and defuzzification techniques. (Software is on the
Chapter 9: Computational Intelligence Theory and Concepts
Discusses unified field of computational intelligence. Offers
definitions. Discusses computational intelligence in context with other
intelligence types. Compares adaptation with learning, and
stochasticity with chaos.
Chapter 10: Computational Intelligence Implementations
Presents an implementation of a fuzzy min-max neural network. Also
included is an evolutionary fuzzy expert system. (Software is on the
Chapter 11: Performance Metrics
Examines issues related to measuring how well a CI implementation is
performing. Performance measures discussed include percent correct,
average sum-squared error, normalized error, evolutionary algorithm
effectiveness measures, receiver operating characteristic curves
measurements, confusion matrices, cost functions, and chi-square
Chapter 12: Analysis and Explanation
Presents analysis and explanation tools that are used to explain how CI
systems do what they do. Included are sensitivity analysis, Hinton
diagrams, and explanation facilities. Application of CI tools to
explanation facilities is reviewed.
Chapter 13: Case Study Summaries
Several case studies are summarized in this chapter. Each summary cites
references where more information can be obtained, and discusses what
makes the case study significant. Included are summaries describing
electroencephalogram (EEG) spike detection, battery state of charge
determination, schedule optimization, and control system design.
Presents a comprehensive glossary of terms relevant to neural networks,
evolutionary computing, and fuzzy logic.
A unified reference list for the book, comprising over 230 entries.
APPENDIX A - This appendix briefly describes the Computational
Intelligence Toolkit (CIT) software, which includes PC-executable code
for all paradigms to which an entire chapter was devoted. Also included
on the diskette is a User's Manual for all the software. The source
code for this software, written in C/C++ is available on diskette from the
authors for a modest price (less than $100). Included in this appendix
(and on the diskette as an ASCII text file) is a one-page order form.
APPENDIX B - This appendix reviews additional resources that can be used
to continue learning about computational intelligence. Included are the
major technical societies around the world, the major journals, and a
sampling of data bases and computer bulletin boards.
A diskette in the back of the book contains software for all of the
implementations discussed in the book. Executable programs and data
files are included. An ASCII file containing an order form for the C/C++
source code is also included.
Information on the authors:
Russ Eberhart phone: 317-278-0255
Purdue School of Engineering fax: 317-274-4567
799 West Michigan Street email: firstname.lastname@example.org
Indianapolis, IN 46202
Pat Simpson phone: 907-345-7347
P. O. Box 242065 email: email@example.com
Anchorage, AK 99524 fax: 907-345-9769
Roy Dobbins phone: 602-981-0783
Software Frontiers Systems, Inc. email: firstname.lastname@example.org
P. O. Box 8524
Mesa, AZ 85214-8524