Numerical Optimization of Computer Models

This book serves as an introduction to optimization methods giving an overview of the different types of optimization problems and their solutions. It begins with a clear description of the problems and methods of optimization and leads into optimization strategies based on hill climbing, direct search, and random search procedures. While these methods are well expounded elsewhere, the chapter containing the relatively new multimembered evolution strategy is not.

One’s overall impression is that much positive thought has gone into the writing of this book. This is borne out by the very clear explanations of the various methods and the well-presented step- by-step layout of the algorithms. The figures are also used to good effect as a further end to the explanation of algorithms, although my feeling is that more figures could have been employed. However, the number of references cited is staggering so there is plenty for the interested reader to follow up.

The two appendices are well worth while. In the first of these, a large number of test problems are considered to illustrate and compare the various methods described in the book. The second appendix gives well-documented lists of Fortran computer programs of evolution strategies, which would assist the potential users of such methods.

However, there does seem to be a serious draw- back to the book. The German author freely admits that most of the book was written in 1974, the remainder being completed by 1977. The original version of the book was written in German. This has since been translated into English and now published in 1981. This state of affairs is reflected in the fact that nearly all the references (vast in number though they are) date from pre-1975. One is left with the nagging doubt that much of the very recent literature has been overlooked.

Text: English, German (translation)