【摘要】 工艺设计主要是应用生产实践中逐步积累起来的工艺知识来进行创造性的实践活动 ,因此 ,相当一部分知识是经验型的知识 ,用基于规则的知识表达方法以及基于框架的知识表达方法描述工艺设计领域中大量的数值型和字符型经验知识时 ,存在知识量大、表达效率低和专家推理效率低等缺陷 .为此 ,论述了工艺知识两种表示形式的特点 ,提出了采用人工神经网络的隐式表示与规则的显式表示相结合的工艺知识表示法 ,通过神经网络的BP算法的学习功能 ,把工艺知识的外部规则表示转化为内部网络权值表示形式 ,并利用该训练好的网络生成供加工方法推理选择的模糊规则矩阵 .更多还原

【Abstract】 Machining process knowledge is accumulated from machining practice, and is mainly used to create new machining process. As most machining knowledge is derived from experience, and the method to express most of numerical and character machining knowledge based on rules and frame has the drawbacks of low expression efficiency and low inference efficiency for a lot of knowledge. This paper discusses the features of two methods to express the machining process knowledge based on rules and ANN. The machining process knowledge is changed from the external rules expression to the internal net weight value description by the BP algorithm of ANN, and the trained net is used to provide the fuzzy rules matrix for the machining process inference.更多还原