Premium
Parallelization of genetic algorithm with sexual selection
Author(s) -
Omori Kiyohiro,
Maekawa Satoshi,
Tamaki Hisashi,
Kitamura Shinzo
Publication year - 2004
Publication title -
electrical engineering in japan
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.136
H-Index - 28
eISSN - 1520-6416
pISSN - 0424-7760
DOI - 10.1002/eej.20029
Subject(s) - selection (genetic algorithm) , population , genetic algorithm , computer science , process (computing) , algorithm , sexual selection , mutation , population based incremental learning , mutation rate , population size , sexual discrimination , artificial intelligence , machine learning , biology , evolutionary biology , demography , genetics , sociology , gene , operating system , gender studies
We propose a parallel genetic algorithm with sexual selection. In genetic algorithms with sexual selection with one population, females keep their traits around local optima by using a lower mutation rate than that of the males, while the males change their traits actively. When a runaway process takes place, the transitions of the males' traits are biased toward a certain direction which depends on the bias of the females' preferences. If the population size is large, the search converges quickly. However, the large population size causes a decrease in search performance. In the proposed method with parallelization, the population size of each subpopulation is maintained adequately, and each subpopulation seeks its own direction of evolution independently. As a result, the proposed method makes a search converge quickly because the runaway process which leads to intermittent evolution tends to take place more quickly than one population model. We applied the proposed method to some test problems. In these problems, while the performance of conventional genetic algorithms was decreased by parallelization, the proposed method exhibited better performance with parallelization. Moreover, the performance of the proposed method is better than that of conventional methods. This capability of parallelization is a remarkable characteristic of sexual selection. © 2004 Wiley Periodicals, Inc. Electr Eng Jpn, 150(1): 42–49, 2005; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/eej.20029