Architecture of Systems Problem Solving

536 Pages·1985·12.534 MB·English

by  George J. Klir (auth.)

ARCH ITECTURE
OF SYSTEMS
PROBLEM SOLVING
ARCH ITECTURE
OF SYSTEMS
PROBLEM SOLVING
GEORGE J. KLIR
State University of New York at Binghamton
Binghamton, New York
SPRINGER SCIENCE+BUSINESS MEDIA, LLC
Library of Congress Cataloging in Publicat ion Data
Klir, George J., 1932-.
Architecture of systems problem solving.
Bibliography: p.
Inlcudes indexes.
1. System theory. J. Title.
Q295.K55 1985 003 85-9283
ISBN 978-1-4757-1170-7 ISBN 978-1-4757-1168-4 (eBook)
DOI 10.1007/978-1-4757-1168-4
© 1985 Springer Science+Business Media New York
Originally published by Plenum Press, New York in 1985
Softcover reprint ofthe hardcover Ist edition 1985
A Division of Plenum Publishing Corporation
233 Spring Street, New York, N.Y. 10013
Ali rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
To the Memory of
W. ROSS ASHBY
What is the hardest thing
of all to see: that which
lies before your eyes.
-GOETHE
PREFACE
One criterion for classifying books is whether they are written for a single purpose or
for multiple purposes. This book belongs to the category of multipurpose books, but
one of its roles is predominant-it is primarily a textbook. As such, it can be used for a
variety of courses at the first-year graduate or upper-division undergraduate level. A
common characteristic of these courses is that they cover fundamental systems
concepts, major categories of systems problems, and some selected methods for
dealing with these problems at a rather general level.
A unique feature of the book is that the concepts, problems, and methods are
introduced in the context of an architectural formulation of an expert system
referred to as the general systems problem solver or GSPS-whose aim is to provide
users of all kinds with computer-based systems knowledge and methodology. The
GSPS architecture, which is developed throughout the book, facilitates a framework
that is conducive to a coherent, comprehensive, and pragmatic coverage of systems
fundamentals--concepts, problems, and methods.
A course that covers systems fundamentals is now offered not only in systems
~cience, information science, or systems engineering programs, but in many programs
in other disciplines as well. Although the level of coverage for systems science or
engineering students is surely different from that used for students in other disciplines,
this book is designed to serve both of these needs.
A course in systems science or engineering programs would normally cover the
whole text, including the various appendices. This material will provide the students
with a broad base for further studies. When equipped with such a base, the student will
maintain an overall perspective during his studies of more advanced and specialized
topics. He will be able to recognize the role of each topic within the overall GSPS
architecture, to see from this larger perspective how the various topics are inter
related, and to use this knowledge in developing a meaningful program of study for
himself.
When used in a course offered to students in the various traditional disciplines,
some parts of the text need not be covered. Specific passages which can be skipped
without jeopardizing the intelligibility of subsequent parts of the book are marked by
symbols ~ and .... at the beginning and end, respectively. They include some instances
of a mathematical presentation of material that for general understanding is
adequately covered at the conceptual level and sufficiently illustrated by examples.
Also marked are passages in which a specific methodological alternative is described
that is not essential for a general comprehension of systems fundamentals. Depending
on the course objectives (or study objectives of an individual reader), the marked
vii
viii PREFACE
passages can be either totally excluded from study, or only omitted during a
preliminary reading of the book.
In addition to its primary function as a text, the book is also intended for
practicing scientists and professionals in various subject areas. An increasing number
of them are becoming interested in learning about modern developments in systems
science, which may be utilized in their own work. The book is obviously of particular
significance to those specialists who are involved in multidisciplinary team projects.
It is also expected that the book will serve as a useful reference for researchers as
well as practitioners in systems science and related fields on one side, and the area of
expert systems on the other side. Systems science researchers will find in the book a rich
source of underdeveloped research areas. Practitioners, on the other hand, will find in
it some general methodological tools of considerably broad applicability.
While most expert systems described in the literature are designed to provide the
user with expertise in a traditional discipline (such as a specific subject area of medicine,
geology, chemistry, or law), the role of the GSPS is to assist the user in dealing with
systems problems. Its expertise is thus systems knowledge and methodology and,
consequently, its utility transcends boundaries between the traditional disciplines. In
this sense, the book should be a useful reference for designers of expert systems and, in
fact, also for computer systems architects-it is the computer architecture that ought
to reflect the underlying systems problem-solving architecture and not the other way
around.
Prerequisite dependencies between individual chapters and sections of this book
are well defined and are expressed by the diagram in Figure P.I. Since they do not form
a linear ordering, there are several alternative ways of studying the material. Chapter I,
which represents an overall introduction, must always be read first. Chapters 2-5 form
a core of the book and are dependent on each other as shown in the diagram. All
fundamental types of systems and key categories of systems problems are introduced in
these chapters. One way of studying the material is to read all these core chapters
before proceeding to the remaining chapters. Another alternative is to follow the
prerequisite dependencies and proceed to relevant sections in Chapter 7 (Goal
Oriented Systems) and Chapter 8 (Systems Similarity) after completing the study of
each of the core chapters. Chapter 6, which is devoted to systems complexity, can be
read in virtually any order. The last chapter, Chapter 9, which overviews the whole
GSPS, should also be the last one read.
Mathematical prerequisites are restricted to the material covered normally in a
one-semester course in finite mathematics. Some knowledge of calculus is useful, but it
is not necessary. Special mathematical concepts, such as the concepts of the Shannon
entropy, fuzzy measure, or metric distance, are introduced in the book before they are
used. For a quick reference, lists of relevant mathematical symbols and a glossary of all
mathematical terms employed in the book are given in Appendices A and B,
respecti vel y.
In order to minimize interruptions in the main text, almost all bibliographical,
historical, terminological, and other remarks are included in the Notes that
PREFACE ix
~--------------~!~~----~
Figure P. I. Prerequisite dependencies of this book.
accompany individual chapters. They are placed at the end of each chapter, numbered,
and occasionally referred to in the main text.
One additional feature of this book should be mentioned. Each of its chapters and
sections is introduced by a quote that depicts the essence of the material covered in it.
The aim of these carefully selected quotes is to appeal to the right hemisphere of the
readers's brain to communicate the key ideas advanced in this book. I expect that in
some instances the quotes will help the reader to understand the material presented
under it, while in other instances the studied material will help him to properly
understand and appreciate the quote. In any case, I am confident that the quotes will
reinforce the learning process and will make it a little more enjoyable.
GEORGE J. KLIR
Binghamton, New York
NOTE TO THE READER
Passages that are set off by symbols ~ and... at the beginning and at the end,
respectively, can be omitted without jeopardizing the intelligibility of the remaining
text.
References to literature are denoted by brackets and contain the first two letters of
the author's last name and the reference number. For instance, [ASl] means the first
reference to Ashby found in the References at the end of the book.
When confused with mathematical notation, the reader should consult the List of
Symbols in Appendix A. When uncertain about the meaning of a mathematical term,
he should consult the Glossary of Relevant Mathematical Concepts in Appendix B.
CONTENTS
Preface vii
Note to the Reader x
CHAPTER 1: INTRODUCTION 1
1.1. Systems Science 3
1.2. Systems Problem Solving 8
1.3. Hierarchy of Epistemological Levels of Systems 13
1.4. The Role of Mathematics 16
1.5. The Role of Computer Technology 19
1.6. Architecture of Systems Problem Solving 23
Notes 27
CHAPTER 2: SOURCE AND DATA SYSTEMS 31
2.1. Objects and Object Systems 33
2.2. Variables and Supports 38
2.3. Methodological Distinctions 44
2.4. Discrete Versus Continuous 51
2.5. Image Systems and Source Systems 54
2.6. Data Systems 65
Notes 78
Exercises 79
CHAPTER 3: GENERATIVE SYSTEMS 81
3.1. Empirical Investigation 83
3.2. Behavior Systems 87
3.3. Methodological Distinctions 96
3.4. From Data Systems to Behavior Systems 102
3.5. Measures of Uncertainty 112
3.6. Search for Admissible Behavior Systems 125
3.7. State-Transition Systems 133
3.8. Generative Systems 148
3.9. Simplification of Generative Systems 149
3.10. Systems Inquiry and Systems Design 160
Notes 164
Exercises 168
xi
xii CONTENTS
CHAPTER 4: STRUCTURE SYSTEMS 175
4.1. Wholes and Parts 177
4.2. Systems, Subsystems, Supersystems 183
4.3. Structure Source Systems and Structure Data Systems 186
4.4. Structure Behavior Systems 194
4.5. Problems of Systems Design 204
4.6. Identification Problem 212
4.7. Reconstruction Problem 227
4.8. Reconstructability Analysis 260
4.9. Simulation Experiments 267
4.10. Inductive Reasoning 274
4.11. Inconsistent Structure Systems 281
Notes 284
Exercises 289
CHAPTER 5: MET ASYSTEMS 295
5.1. Change versus Invariance 297
5.2. Primary and Secondary Systems Traits 300
5.3. Metasystems 303
5.4. Metasystems versus Structure Systems 312
5.5. Multilevel Metasystems 315
5.6. Identification of Change 317
Notes 322
Exercises 322
CHAPTER 6: COMPLEXITY 323
6.1. Complexity in Systems Problem Solving 325
6.2. Three Ranges of Complexity 328
6.3. Measures of Systems Complexity 334
6.4. Bremermann's Limit 336
6.5. Computational Complexity 342
6.6. Complexity Within GSPS 349
Notes 351
Exercises 353
CttAPTER 7: GOAL-ORIENTED SYSTEMS 355
7.1. Primitive, Basic, and Supplementary Concepts 357
7.2. Goal and Performance 358
7.3. Goal-Oriented Systems 360
7.4. Structure Systems as Paradigms of Goal-Oriented Behavior
Systems 363
7.5. Design of Goal-Oriented Systems 367
7.6. Adaptive Systems 370
7.7. Autopoietic Systems 377
Notes 381
Exercises 382