Review of the book „Genetic Algorithms Reference” written by Tomasz
Dominik Gwiazda.
This is surely an atypical book. It begins with a 4-page preface, and
the rest of this book (some 400 pages) is divided into nearly 180
sections (each some two page long). Every section discusses a given
recombination operator. This book should be evaluated from the point of
view of author’s intentions for he plans to publish a series of works
regarding various aspects of evolution algorithms (e.g. selection or
mutation). This is the very first book of entire cycle.
Scope and level against the background of
other books devoted to the same topics.
Such comparison is extremely difficult
because the book by Tomasz Gwiazda is very different from other
publications about genetic algorithms. Probably the most similar is the
Handbook of Evolutionary Computation edited by Baeck, T., Fogel, D., and
Michalewicz, Z., and published jointly by Oxford University Press and
Institute of Physics in 1997. The Handbook encompasses over 1000 pages
and presents all aspects of evolutionary methods in long and brief
sections of summary character. For sure, the book by Tomasz Gwiazda is
very unique on the market.
Modern way of topic presentation.
Also the way of topics presentation is
very unique. As I have stated in former paragraph, there is no similar
book available of the market right now. Moreover, its author agreed to
update all the volumes regularly via electronic means, and this should
make his offer even more attractive.
Is the book useful for students?
It is rather hard to say. Most probably,
this book will not be utilized as any basic text for evolutionary
algorithms course because (as the author admits) there is a plenty of
introductory texts available already. However, majority of the libraries
should definitely purchase this particular work as an important
supplementary material.
Potential readers.
This book may constitute an excellent
reference point for the students working on the projects, or doing
master’s degrees of even Ph.D.’s. It is also a valuable source for
researchers investigating this particular field.
Additional remarks.
There are two aspects of this review that
slightly overlap (and I have to admit I have had some troubles with
that). One aspect is the evaluation of a current version of this book;
the other one – estimation of its future Polish translation. Of course,
these are two different tasks, however they are linked together.
My general remark regarding the present
text is that: because this book constitutes the first volume of a cycle,
it’s a pity it does not include even the shortest introduction to
genetic algorithms (or to evolutionary algorithms). The author presumes that
the readers already know the specifics of this field. It may be true,
but on the other hand it definitely narrows the scope of the potential
readers. Author could consider an option of enhancement of his text when
preparing Polish translation of this analysis. I am also not very much
in favor of the adopted classification of the operators. Tomasz Gwiazda
admits that his division is a bit artificial. Personally, I prefer to
have a list of all the operators (without any classification
whatsoever). The main character of each operator could be marked with a
single sign in brackets. For instance, “injection crossover” listed on
page 202 in the table of contents would be presented as follows: ……………
Injection crossover
(R) ................................................................202
It would also make the notation much
simpler because some operators could have a double classification, for
example:
Random crossover (B,R)................................................................222
My other notes concern the possible
translation of this work from English into Polish.
I hope the above stated reviews will be
useful.
Zbigniew Michalewicz |