Evaluating the performance of recombination operators with island parallel genetic algorithms
7th IFAC Conference on Manufacturing Modelling, Management, and Control, MIM 2013, Saint Petersburg, Rusya, 19 - 21 Haziran 2013, cilt.46, ss.748-753, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Cilt numarası: 46
- Doi Numarası: 10.3182/20130619-3-ru-3018.00260
- Basıldığı Şehir: Saint Petersburg
- Basıldığı Ülke: Rusya
- Sayfa Sayıları: ss.748-753
- TED Üniversitesi Adresli: Evet
Özet
Obtaining exact results for larger problem instances of the intractable problems is still an important issue. Genetic algorithms can provide good solutions for such problems by using appropriate recombination and mutation operators to search the solution space, provided that they are given a sufficient population size of individuals and enough time to iterate through generations. There are numerous proposed recombination operators that programmers use for this purpose. Most of the time, one of these operators are randomly selected to solve a problem without knowing its capabilities. In this study, we analyze the most used recombination operators and propose a new swap-based operator, Bubble Recombination Operator (BRX), for the intractable permutation problems. The proposed operator swaps iteratively and outperforms most of the well-known permutation recombination operators. BRX provides either the optimum or very close to the best available solutions reported so far for the problem instances of the Quadratic Assignment Problem. © IFAC.