Abstract: To solve the multi-objective optimization problem of surface mounting, a multi-objective optimization model was established with feeders, placing sequence and changing times of the nozzles as sub-objectives. The concept of information entropy was introduced into the grey correlation analysis, and the grey entropy correlation analysis method was proposed. The original grey correlation degree was changed into grey entropy correlation degree. Based on grey entropy correlation analysis method, the grey entropy correlation degree functions of three sub-objectives were established and a multi-objective differential evolution algorithm with grey entropy correlation was presented. The coding method of real vector-location sequence was adopted to realize the application of differential evolution algorithm on surface mounting optimization. The experimental results indicated that at least two sub-objectives in Pareto optimal solutions of differential evolution algorithm were better than genetic algorithm, and the solution had better uniform distribution. Therefore, the multi-objective differential evolution algorithm with grey entropy correlation could solve the multi-objective optimization problem of surface mounting effectively.

Key words: surface mounting, multi-objective optimization, differential evolution algorithms, grey entropy correlation analysis

摘要： 为了解决表面贴装工艺优化问题，构建以喂料器、贴放顺序和吸嘴更换次数为子目标的表面贴装多目标优化模型。将信息熵理论引入灰关联分析，建立灰熵关联分析法，把普通灰关联度转化为灰熵关联度，进一步基于灰熵关联分析法建立表面贴装三个子目标的灰熵关联度函数，提出基于灰熵关联的多目标差分算法。采用实数向量-位置排序的编码方式实现差分算法在表面贴装工艺优化中的应用。实验结果表明，差分算法得到的Pareto最优解集中的三个子目标函数值至少有两个好于遗传算法，且解集分布更均匀，说明灰熵关联分析法能够有效地实现贴装工艺的多目标优化。

关键词: 表面贴装, 多目标优化, 差分算法, 灰熵关联分析