Heidelberg

Science
Library
Heidelberg
Science
Library
DeliaKoo
Elements of Springer Science+

Optimization
Business Media, LLC

With Applications in
Economics and Business
Delia Koo
Eastern Michigan University
Ypsilanti, MI 48197
USA

AMS Subject Classifications: 49-xx, 90Cxx

Library of Congress Cataloging in Publication Data

Koo, Delia.
Elements of optimization, with applications in
economics and business.
(Heidelberg science library)
Bibliography: p.
Includes index.
l. Mathematical optimization. 2. Economics, Mathematical. 3. Business
mathematics. I. Title. II. Series.
QA402.5.K66 519.4 66-7629

All rights reserved.

No part of this book may be translated or reproduced in any form without written
permission from Springer-Verlag.

© 1977 by Springer Science+Business Media New York

Originally published by Springer-Verlag New York in 1977

98765432 I

ISBN 978-0-387-90263-0 ISBN 978-1-4612-6358-6 (eBook)

DOI 10.1007/978-1-4612-6358-6
Preface

This book attempts to present the concepts which underlie the various
optimization procedures which are commonly used. It is written primarily
for those scientists such as economists, operations researchers, and en-
gineers whose main tools of analysis involve optimization techniques and
who possess a (not very sharp) knowledge of one or one-and-a-half year's
calculus through partial differentiation and Taylor's theorem and some
acquaintance with elementary vector and matrix terminology. Such a
scientist is frequently confronted with expressions such as Lagrange multi-
pliers, first- and second-order conditions, linear programming and activity
analysis, duality, the Kuhn-Tucker conditions, and, more recently, dy-
namic programming and optimal control. He or she uses or needs to use
these optimization techniques, and would like to feel more comfortable
with them through better understanding of their underlying mathematical
concepts, but has no immediate use for a formal theorem-proof treatment
which quickly abstracts to a general case of n variables and uses a style
and terminology that are discouraging to people who are not mathematics
majors. The emphasis of this book is on clarity and plausibility. Through
examples which are worked out step by step in detail, I hope to illustrate
some tools which will be useful to scientists when they apply optimization
techniques to their problems.
Most of the chapters may be read independently of each other-with
the exception of Chapter 6, which depends on Chapter 5. For instance, the
reader will find little or no difficulty in reading Chapter 8 without having
read the previous chapters.
I am indebted to Alan Baquet, Luke Chan, David Cheng, Norman
Obst, W. Allen Spivey, M. B. Suryanarayana, and P. K. Wong for having
read all or portions of the manuscript and given me comments. Norman
Obst has been most generous with his time and helpful suggestions and
deserves special thanks. Anthony Koo initiated the idea of writing this
book and supplied me with all the economic literature with which I was
vi Preface

not familiar, and I thank him for his consideration and restraint in
demanding my time during the course of writing. I am most grateful for
the generous support of the Mathematics Department at Michigan State
University, where I spent my sabbatical year to finish this book, and for
the very kind and able assistance of Glendora Milligan, Mary Reynolds,
and Leota Steadman.

Delia Koo
Contents

Notation ix

1 Extrema of a Function of One Variable

1.1 Extreme points 1
1.2 Necessary and sufficient conditions 2
1.3 Taylor's formula for one variable 3
1.4 Necessary condition for an extremum 4
1.5 Sufficient condition for an extremum 5
1.6 Necessary and sufficient condition for an extremum 10
1.7 Remarks 11
Exercises II

2 Extrema of a Function of Two or More Variables

(without Constraint)
2.1 Necessary condition 13
2.2 Taylor's formula for a function of two variables-
necessary condition 13
2.3 Sufficient condition for a function of two variables 15
2.4 Sufficient condition for a functionf(x 1,x2 , ••• ,xn) of n variables 21
Exercises 27

3 Functions of Two or More Variables (with Constraint)

3.1 Preliminary 29
3.2 Necessary condition 30
3.3 Sufficient condition 32
3.4 Examples 37
Exercises 49

4 Simultaneous Maxima of Several Functions

4.1 Statement of the problem 51
4.2 Graphic interpretation 53
4.3 Necessary conditions for Pareto optimality in production,
consumption, and the economy 56
Vlll Contents

4.4 Pareto optimality in general 60

Exercises 62

5 Linear Programming
5.1 General nature of linear programming problems 64
5.2 Some geometric and algebraic interpretations 69
5.3 The simplex method 81
5.4 Complications and adjustments 88
5.5 Solving a minimization problem 97
5.6 Assumptions of linear programming 99
Exercises 100

6 Linear Programming-Duality and

Sensitivity Analysis
6.1 Duality 102
6.2 Proofs of some theorems ll2
6.3 Economic interpretation of duality ll4
6.4 Dual simplex method 117
6.5 Sensitivity analysis 118
Exercises 127

7 Nonlinear Programming
7.1 General nonlinear programming problems 128
7.2 The Kuhn-Tucker conditions 131
7.3 Further look at the Kuhn-Tucker conditions 134
7.4 Quadratic programming 142
7.5 Separable functions 152
7.6 Some economic applications of the Kuhn-Tucker conditions 159
Exercises 165

8 Optimal Control
8.1 The control problem and some terminology 167
8.2 The classical calculus of variations 171
8.3 The maximum principle (modem calculus of variations) 183
8.4 Maximum principle-the costate variables and constraints 191
8.5 Dynamic programming 193
8.6 Dynamic programming and the calculus of variations 201
8.7 Stochastic and adaptive controls 204
Exercises 204

Appendix I
Quadratic Forms and Characteristic Roots 208

Appendix II
Convexity and Quasiconvexity 210
Bibliography 212
Index 217
Notation

Throughout this book we use lower case boldface letters such as x, y, u to

denote vectors. Column vectors are normally used, but, to save space, they
are often written as transposes of row vectors, such as x=(x 1, ... ,xnl or
xT =(x 1, ... ,xn), where T means transpose.
Matrices are denoted by capital letters such as A,B,C, and we some-
times write
A= [ aii J.
The determinant of a matrix A is written as det[A] or IAI.
The derivative of a function f ( x) evaluated at x = x 0 is written

dfl
dx x=xo or df(x 0 ) / dx.

Subscripts and superscripts are explained at the places where they

appear. In Chapters 2 and 3, subscripts such as f 1 or f 2 are used to refer to
partial derivatives. Thus f)= of(x),xz)jox), fu = aj"(x),xz)joxf, f12
= oj"(x 1,,x2)/ox 1ox2 • Subscripts, however, are also used in other places to
label variables, etc., but they are clearly specified. Superscripts are gener-
ally used to label constraints. Again, they are explained in the text.
When there may be doubt as to where a theorem or proof or example
ends, the symbol D is used to indicate the end.
The formulations, tables, and figures are numbered according to
chapter and section. Thus (2.3.1) means the first equation in Section 3 of
Chapter 2. Table (5.3.1) means the first table in Section 3 of Chapter 5.
Some commonly used symbols are:
E: is an element of;
n
~ a;=a 1 +a2 + ... +an;
i= I
Vf(x) is the gradient vector (offox 1,offox2, ... ,offoxnl;
[a, b] is a closed interval. For example, t E [ t0, t1) means t0 .;;; t.;;; t1.