Improving Genetic Algorithms by Using New Operators

Based on natural selection in environments and natural genetics in biology, the genetic algorithm evolves according to the principle of survival of the fittest (natural selection). The evolution of such a random search algorithm always encounters uncertainties, which limit the search efficiency of the algorithm. To enhance the genetic algorithm, many researchers lay great emphasis on the improving of the genetic operators or the parameter adjustments. In this paper, we propose a new operator called gene extraction to extract the good and bad genes of the individuals. These good and bad genes can be used to sieve out the patterns of good individuals. We also propose two advanced methods for population initialization, which force the initial individuals to locate at specific regions of the search space. These methods are well integrated with the simple genetic algorithm and exhibit good performances.

並列關鍵字

Genetic algorithms ； gene extraction ； uniform initialization ； unbiased initialization

國際替代計量

全文下載

主題瀏覽