Abstract: To improve the auto-body assembly dimensional quality by assembly technology optimization, aiming at the geometric feasible assembly sequences for the auto-body, the precedence relationships and assembly control characteristics quantity between parts were described by multi-attribute directed graph to eliminate the unfeasible engineering assembly sequences. With assembly dimensional quality as the objective function, the hybrid particle swarm optimization and genetic algorithm was proposed to optimize the assembly operations between parts. The optimal assembly sequence was obtained through assembly variation propagating based on linear assembly variation analysis model. The optimization of assembly control characteristics was illustrated by auto-body side assembly. The results indicated that the selection of assembly control characteristics was affected by different assembly sequences, thus the final product assembly variation was affected.

Key words: particle swarm optimization algorithm, genetic algorithms, auto-body, assembly operation, optimization

摘要： 为通过装配工艺优化提高车身装配尺寸质量，针对车身众多几何可行装配顺序，应用多属性有向图描述零件间的优先关系和装配控制特征数量,来去除非工程可行装配顺序。以装配尺寸质量为目标函数，提出粒子群—遗传混合算法优化零件间装配操作，通过线性装配偏差分析模型进行装配偏差累积运算，获得了最优装配顺序。通过车身侧围装配体阐述了装配控制特征的优化过程，结果表明，不同的装配顺序将影响装配控制特征的选择，从而影响最终的产品装配偏差。

关键词: 粒子群优化算法, 遗传算法, 车身, 装配操作, 优化