基于父个体相似度的自适应遗传算法

标准遗传算法在产生后代个体时采用先交叉后变异的策略,一方面当父个体非常相似时,交叉操作很难产生新的个体,影响算法对新的解空间进行搜索,从而导致种群多样性的丧失;另一方面交叉产生的优秀个体再历经变异,极有可能遭破坏而影响算法的收敛性。该文根据染色体的相似性,给出了个体相似度的概念,并在此基础上提出了依据父个体相似度的大小自适应地选择遗传算子(交叉或变异)的遗传算法。仿真实验表明,与采用常规遗传策略的遗传算法相比,新算法能显著提高解的质量和收敛速度。

Adaptive Genetic Algorithm Based on Parents' Similarity

Simple Genetic Algorithm uses first crossover then mutation strategy while generating offspring.On one hand,it is difficult for new individuals to be generated during crossover operator when parents are very similar,resulting in the difficulty in the exploration of new solution space and the loss of population diversity.On the other hand,superior individuals generated by crossover operator have to undergo mutation and are very likely to turn into inferior individuals after mutation procedure,and then influence GA's convergence quality and convergence speed.Based on the similarity of chromosomes,individual similarity concept is proposed in this paper,and an adaptive genetic algorithm that perform crossover or mutation operator according to parents' similarity is put forward.Our experimental results indicate that this new genetic algorithm can find better solution and converge faster than the genetic algorithm employing general genetic strategy.