We anticipate that most readers of this book will use it in one of three ways, which represent differing levels of use:
1.as an introduction to nonlinear estimation for non-specialists
2. to learn/use a particular tool
3. to build their own nonlinear parameter estimation system or to integrate our methods with another (statistical or database) package.
--From chapter 1
This item is a hybrid--the 5.25” floppy in the endpaper pocket is at least half the reason to purchase. It contains a nonlinear optimization system, formatted for IBM PCs and compatibles with Microsoft BASIC (if you have patience or a grad student or other free labor, all the code is also printed in the book). Why such a horrible choice (from a purist’s point of view)? Because just about every micro out there has BASIC, and only a fraction have any other language you would care to name. These are the realities of the marketplace as it exists today; the intensity of your shudder is a measure of the depth of your failure (individually or collectively) to make the world a reasonable place. The book itself was produced from camera-ready copy created with a dot-matrix printer. It seems to meet its stated intentions well.
In this book, the approach to parameter estimation is the minimization of loss functions. For this reason, a better title would have been Nonlinear optimization and parameter estimation: while parameters do get estimated, the emphasis is on how to find minima well. The system looks like it could at least take a crack at any performance measure expressible as the result of a BASIC subroutine.
The chapter headings describe fairly well what is covered:
:9N(1)Overview and Purpose
(2)Examples of Nonlinear Parameter Estimation Problems
(3)Exploring the Problem
(4)Direct Search Methods
(5)The Hooke and Jeeves Method
(6)The Nelder-Mead Polytope Method
(7)Minimization Using the Gradient
(8)The Conjugate Gradients Method
(9)The Truncated Newton Method
(10)The Variable Metric Method
(11)Methods for Sums of Squared Functions
(12)Choosing an Approach and Method
(13)Measures and Consequences of Nonlinearity
(14)Difficult Problems
(15)Implementation Details
(16)Examples from Chemical Kinetics
(17)Applications in Economic Modeling and Forecasting
(18)Test Problems and Other Examples
Appendix (A)Program Code for Test and Example Problems
Appendix (B)Data Dictionary
Appendix (C)Line Number Conventions
Appendix (D)Exceptions from Minimal BASIC in Code Presented
Appendix (E)Computing Environment.
Author and subject indices are also provided. To quote from the preface,
In our presentation, we have tried to provide enough background to the methods to allow for at least an intuitive understanding of their working. However, we have avoided mathematical detail and theorem proving. This is a book about mathematical software and about problem solving rather than mathematics. In partial compensation for this deliberate restriction in the scope of our treatment, we include a quite extensive bibliography.
At nearly 400 items, the authors do seem to have given an adequate window into the relevant literature.
As for the material on the disk, I found it to be both outstanding and extremely easy to use. The programming standards were obviously first-rate, and the documentation is excellent.
