基于改进遗传编程的并行装配序列规划

摘要

According to the limitations that genetic algorithm can only generates linear assembly solutions,the genetic programming was adopted to solve the parallel assembly sequence planning problem.The traditional genetic programming algorithm was improved by modifying the genetic programming symbols set content and retaining the terminal set only,since there was only assembly action between each parts.The operators and parameters of genetic programming algorithm were designed and improved,so that it could adapt to dealing with the parallel assembly sequence planning problems.The improved method used tree structure solution instead of traditional string structure solution,and the efficiency and automatic assembly level in the assembly of parts was promoted with changing the assembly way.In addition,combined with the actual production requirement,a new assembly fitness measure factor—"total weight of assembly solution" was given,Combining with the other traditional fitness measure factors,the qualities of solutions of assembly was evaluated.%针对遗传算法只能生成线性装配解的局限,提出采用遗传编程来求解并行装配序列规划问题的方法.对传统的遗传编程算法进行改进,由于各零件间只有装配的动作,删除了遗传编程符号集内容,只保留了终端集；对遗传编程算法中有关算子及参数进行了设计与改进,使算法最终能够处理并行装配序列规划问题.该方法使用树型解结构代替串型解结构,转变了装配方式,提高了零件的装配效率和自动化装配水平.结合生产实际需求,给出了一种新的装配适应度衡量因子——装配总重,结合其他传统适应度衡量因子,共同评判装配解的优劣.