摘要： 通过对遗传算法早熟现象的认识,受高等动物进化能够保持可以产生众多的变异个体的启发,提出了一个具有性别特征的遗传算法,该算法的交叉算子要求性别不同和双亲不同才能实施,通过实验函数的测试表明该算法可以有效地保持种群的多样性,避免近亲繁殖,有摆脱早熟的能力。特别是和小生境技术结合求解多峰函数质量明显提高。

关键词: 遗传算法, 早熟现象, 结构编码, 人工生命, GA欺骗问题

Abstract: Understanding the premature phenomenon, this paper introduces a genetic algorithm with sex character enlightened by superior court evolution maintaining many mutation individual. The algorithm requires crossover operators to choose different sexes and different parents to bring into effect. The test of experimental function makes clear that it could effectively maintain population diversity to avoid close relative propagating, and it is able to get rid of premature. Especially it joins niche technology to solve multi model function optimization, the solving quality is improved in evidence.

中图分类号: 

• TP18