Extending the bounds of the search space: a multi-population approach
Martín A Beck, Ian C. Parmee · 1999
This paper reports an investigation of GA based techniques that allow the search to move outside of its pre-defined limits. Using a hybrid coarsegrained / fine-grained GA, dynamic fitness sharing techniques are presented that encourage sub-populations to explore differing high performance areas of the search space. Such techniques when coupled with a migration typology provide a realistic mechanism for moving a GA based search to unspecified areas of the search space. 1