• 论文 •    

基于反向传播神经网络的变型设计变异点快速定位

赵秀燕,赵婷婷,张强,陈微微,魏小鹏   

  1. 大连大学 信息科学与工程省重点实验室,辽宁大连116622
  • 出版日期:2009-08-15 发布日期:2009-08-25

Fast location of aberrant point for variant design based on the BP neural network

ZHAO Xiu-yan, ZHAO Ting-ting, ZHANG Qiang, CHEN Wei-wei, WEI Xiao-peng   

  1. Liaoning Key Lab of Intelligent Information Processing, Dalian University, Dalian 116622, China
  • Online:2009-08-15 Published:2009-08-25

摘要: 为了对产品进行有效变型设计,针对产品模型的最小模块,提出了变异点概念,并在分析产品结构的基础上,将反向传播神经网络引入到变型设计中,建立了快速定位变异点的反向传播人工神经网络模型,较好地解决了人为确定变异点的杂乱性与欠准确性问题。最后进行了实证分析,结果表明,采用反向传播人工神经网络快速定位变异点是可行的。

关键词: 反向传播神经网络, 模块划分, 变型设计

Abstract: Aiming at the efficiency of variant design, module partition was often carried out till the minimum module after establishing the product model. According to the minimum module, Aberrant Point was defined and the BP neural network model for its fast location was established, which solved well the problem of the disorder and inaccuracy in the earlier man-made location method. Finally, a typical case was given to evaluate the performance of the proposed approach, which demonstrated that the approach was obviously helpful for quick of design variant product.

Key words: back propagation neural network, module partition, variant design, aberrant point

中图分类号: