|
The idea for this second edition came from Janusz Kacprzyk on April 29,
2005, who kindly invited me to his new Springer series, Studies in Computational
Intelligence. The initial plan was to correct the usual typos and mistakes
but leave the book unchanged, as Janusz thought (and I agreed with
him) that it was the proper moment for a second edition. But then there was
the problem of the new format and I had to reformat and proofread everything
again. And I just thought that while I was at it, I might as well change
some things in the book to make it more enjoyable and interesting. Foremost
in my thoughts was the restructuring of chapter 4, The Basic GEA in Problem
Solving. In that chapter, buried together with a wide variety of problems,
were several important new algorithms that I wanted to bring to the
forefront. These algorithms include: the GEP-RNC algorithm (the cornerstone
of several new other algorithms); automatically defined functions; polynomial
induction; and parameter optimization. So I removed all these materials
from chapter 4 and gave them the deserved attention by writing four
new chapters (chapter 5 Numerical Constants and the GEP-RNC Algorithm,
chapter 6 Automatically Defined Functions in Problem Solving, chapter 7
Polynomial Induction and Time Series Prediction, and chapter 8 Parameter
Optimization). Then this new structure just begged for me to include one of
the last additions to the GEP technique – decision trees – that I was regretfully
unable to include in the first edition (I implemented decision trees in
August of 2002, two months before sending the manuscript to the printer).
So, chapter 9, Decision Tree Induction, is totally new to this second edition
and an interesting addition both to the book and to GEP. The last three chapters,
chapter 10 Design of Neural Networks, chapter 11 Combinatorial Optimization,
and chapter 12 Evolutionary Studies, remain basically unchanged. |