基于双螺旋染色体和分层结构的遗传算法

Research on genetic algorithm based on double helix chromosome & multi-layers population structure

为满足制造业相关项目中寻找最优化解的需求,提出了一种基于双螺旋染色体和分层种群结构的遗传算法模型。在此模型中,解染色体数据结构采用的是类似DNA的双螺旋对称结构,一个解染色体中包含2份完全相同的解信息,但在遗传运算中各自承担不同的角色;种群采用优、中、差3类解按一定比例组成的分层结构,使得新生的解种群能够涵盖更广的求解范围;遗传运算参数中的交叉和变异概率随种群结构特征动态变化。基于此模型的遗传算法,能够保护优秀解,避免算法早熟,同时还具有收敛速度快、稳定性强的特点。

To deal with the problem of seeking optimal solution for mechanical projects in manufacturing, a new model of genetic algorithms based on double helix chromosome and multi-layer population structure was proposed. The gene data structure of chromosome of the proposed model representing the real solution was based on double symmetrical chains structure,just like the double helix spiral structure of DNA in nature. There were two kinds of completely identical solution information in every solution chromosome but played different roles in genetic algorithm. The population of chromosomes was proportionally made up of excellent, medium and bad layers so that each layer could be prorated by the user. And the probability of crossing and aberrance would be dynamically changed with the structure of population. This model could keep the excellent solution genes and avoid the precocity of computing process with fast astringency and better stability.