Foundations of global genetic optimization. With contribution by Henryk Telega. (Q2644258)

This volume within the book series ``Studies in Computational Intelligence'' is devoted to the question whether and how genetic algorithms can be utilized to solve complex global contiunous optimization problems. The first author (Robert Schaefer) introduces into the field and focusses on the idea to model genetic algorithms as dynamic systems which transform the probabilistic sampling measure in a regular way (f.e. in density oriented clustering DC, followed by a local search in each cluster which is described in chapter 6 by the second author, Henryk Telega). This approach allows to find central parts of basins of attraction of the local minima rather than isolated minima itself. Of particular interest are the chapters about the analysis of asymptotic features and stopping rules, where the authors assume self-adapting genetic algorithms, which imply the uniform Markovian rule of state transition. A taxonomy and first short analysis of adaptive genetic search is given (in chapter 5) additionally. Applicability of global genetic search strategies seems to be restricted mainly to continuous global optimization problems where the focus is on recognition and approximation of basins of attraction and of level sets.