Due to the lot of different Genetic Algorithm variants, encodings, and attacked problems, very little general theory is available to explain the internal functioning of Genetic Algorithms. Consequently it is very difficult for researchers to find a common language to document quality improvements of newly developed algorithms. In this paper the authors present a new Allele Meta-Model enabling a problem-independent description of the search process inside Genetic Algorithms. Based upon this meta-model new measurement values are introduced that can be used to measure genetic diversity, genetic flexibility, or optimization potential of an algorithm's population. On the one hand these values help Genetic Algorithm researchers to understand algorithms better and to illustrate newly developed techniques more clearly. On the other hand they are also meaningful for any GA user e.g. to tune parameters or to identify performance problems.
|Number of pages||10|
|Journal||Lecture Notes in Computer Science|
|Issue number||PART II|
|Publication status||Published - 2005|
|Event||First International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2005 - Las Palmas, Canary Islands, Spain|
Duration: 15 Jun 2005 → 18 Jun 2005
- Genetic Algorithms