计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (18): 239-242.

• 工程与应用 • 上一篇    下一篇

产品设计模块聚类的多目标优化及其应用研究

梁贤烨,王永娟,方  峻   

  1. 南京理工大学 机械工程学院,南京 210094
  • 出版日期:2015-09-15 发布日期:2015-10-13

Module product design clustering multi-objective optimization and its application research

LIANG Xianye, WANG Yongjuan, FANG Jun   

  1. School of Mechanical Engineering, Nanjing University of Science & Technology, Nanjing 210094, China
  • Online:2015-09-15 Published:2015-10-13

摘要: 在产品的模块化协同设计中,设计模块的划分是一个关键问题。为了满足实际应用对模块划分的多方面要求,采用多目标优化方法对设计结构矩阵进行聚类划分,该方法通过粒子群算法建立非支配解集,从中获得理想的划分方案。在聚类优化算法中,为了方便设置粒子的速度,提出了一种简易的编码方式对设计结构矩阵的聚类进行编码,在适应度函数中计算出了每个聚类的联系信息流量,在建立非支配解集时采用拥挤距离方法将多余的解删除。以轻武器模块化协同设计的例子对该方法进行了验证。

关键词: 协同设计, 设计结构矩阵, 粒子群算法, 多目标优化

Abstract: Clustering design module is a crucial step in module product design. In order to fulfill various demands in practice applications, this paper uses multi-objective optimization to cluster DSM. This method uses PSO to establish non dominated solution set, and then selects the best solution from the set. In order to adjust particle velocity in the optimization algorithm, this paper proposes a new easy encoded mode to encode DSM, then calculates relation information flow of each cluster in fitness function, and utilizes crowding distance to delete redundant solution in non dominated solution set. This paper uses a simple case to verify the method.

Key words: collaborative design, design structure matrix, particle swarm optimization, multi-objective optimization