一种基于实数编码的自适应多亲遗传算法

A Multi-Parent Self-Adapted Genetic Algorithm Based on Real Coding

为了克服基于二进制编码的遗传算法在求解连续参数优化问题时的缺陷,采用实数编码,定义1种度量多样性的指标,来自适应地调整基于实数编码的多亲遗传算法中交叉概率和变异概率,提出1种基于实数编码的自适应多亲遗传算法.该算法能自适应地调整其参数,且在求解优化问题的过程中,能克服早熟收敛的现象,提高搜索能力,加速收敛速率.最后对该算法进行了理论分析.

To break the limitation of binary coded genetic algorithm in solving the optimizing problem with continuous parameter, an index, which reflected the diversity of population, was used to adjust the probabilities of crossover and mutation in multi-parent genetic algorithm. Thus, a self adapted multi-parent genetic algorithm based on real coding（AMGARC） was proposed. AMGARC could adjust automatically its main parameters. And it could also overcome premature convergence, improve the optimizing ability and accelerate the convergence speed. Finally, its theoretical analysis was addressed.