Foundations of Global Genetic Optimization

Schaefer R.978-3-540-69431-1

This book is devoted to the application of genetic algorithms in continuous global optimization. Some of their properties and behavior are highlighted and formally justified. Various optimization techniques and their taxonomy are the background for detailed discussion. The nature of continuous genetic search is explained by studying the dynamics of probabilistic measure, which is utilized to create subsequent populations. This approach shows that genetic algorithms can be used to extract some areas of the search domain more effectively than to find isolated local minima. The biological metaphor of such behavior is the whole population surviving by rapid exploration of new regions of feeding rather than caring for a single individual. One group of strategies that can make use of this property are two-phase global optimization methods. In the first phase the central parts of the basins of attraction are distinguished by genetic population analysis. Afterwards, the minimizers are found by convex optimization methods executed in parallel.

---

Table of contents :
front-matter.pdf……Page 1
chapter1.pdf……Page 11
chapter2.pdf……Page 17
chapter3.pdf……Page 41
chapter4.pdf……Page 64
chapter5.pdf……Page 123
chapter6.pdf……Page 161
chapter7.pdf……Page 206
back-matter.pdf……Page 209