基于人工神经网络的制造执行系统软件构件库构件提取方法研究

Research on Artificial Neural Network Based Component Retrieval Methodology from MES Component Database

为了提高从软件构件库中搜索构件的速度和准确率 ,利用误差反向传播神经网络 (BP网络 )研究了构件的搜索问题。针对制造执行系统软件构件的特点 ,提出一个基于刻面的制造执行系统软件构件表示模型。在此基础上 ,建立了在构件刻面空间和构件标识之间进行匹配的BP网络模型 ,并详细研究了该BP网络模型的结构、算法、输入 /输出参数处理等问题。最后 ,通过一个构件库原型系统进行了验证。

To improve the speed and accuracy in component retrieval from software component database,BP network was introduced in the research of component searching. In line with the characteristics of Manufacturing Execution System (MES) software components,the paper proposed a facet-based MES software component model,on which it based,a BP network model was built to match the component facet and component identification. The structure,algorithm,and analysis of input/output data of BP network were investigated particularly. Finally a component database prototype was used to verify the method.