Some extensions of biased random-key genetic algorithms
M.L. Lucena, C.E. Andrade, M.G.C. Resende, and F.K. Miyazawa
Proceedings of the XLVI Symposium of the Brazilian Operational Research Society, Salvador, Bahia, Brazil, September 16-19, 2014.
ABSTRACT In this paper, we propose two new variants of the Biased Random-Key Genetic Algorithm. In the first, the algorithm assigns a gender to each chromosome and only allows crossover between chromosomes of different gender. In the second variant, more than two parents are used to generate a new offspring. Computational experiments are conducted on instances of set covering and set packing problems. We empirically show that the multi-parent modification is able to overcome the original biased random-key genetic algorithm although it requires slightly more time and iterations to converge.
PDF file of full paper
Mauricio G.C. Resende's Home Page Last modified: 13 October 2014