Abstract: Aiming at the selective assembly problem for mechanical products with multi-objective and multi-quality assembly function,a method based on Density-based Multi-objective Evolutionary Algorithm(DMOEA) was proposed.Considering the relationship between dimension chains and dimensions,a selective assembly model under the requirement of multi-objective and multi-quality was established with the success rate of selective assembly and assembly accuracy which were the optimization objectives.Using natural number coding,a random sequence with the dimension index of a part was used as a paired coded gene.The fitness function was defined by individual aggregation density,and the selective assembly problem with multi-objective was solved by DMOEA,which realized the optimization of multi-objective and multi-quality assembly function.A part selective-assembly system based on DMOEA was developed,and was applied to a company's products,the feasibility and effectiveness of the proposed method were verified.

Key words: selective assembly, multi-objective evolution algorithm, assembly quality, individual aggregation density, mechanical products

摘要： 针对机械产品多目标多质量要求下的选择装配问题,提出一种基于密度的多目标进化算法(DMOEA)的选择装配方法。考虑零件尺寸链与尺寸的关联关系,以选配成功率和装配精度为优化目标构建多目标多质量要求下的选配模型。使用自然数编码,以一个零件尺寸值序号的随机序列作为配对编码的基因,并采用个体聚集密度定义适应度函数,利用DMOEA对多目标选配问题进行求解,实现了多目标多质量要求下的选配优化。开发出基于DMOEA的零件选配系统,应用于某公司产品,验证了所提方法的可行性与有效性。

关键词: 选择装配, 多目标进化算法, 装配质量, 个体聚集密度, 机械产品