基于模式化知识和人工神经网络的智能设计系统框架研究

阐述了将神经网络和模式识别引入智能设计系统的模型－集成转换模型，在该模型中，对智能系统与设计的实现系统的结合进行了探讨，提出了数值编译器与反编译器、图形／编码转换器等将神经网络引入设计系统来的工具和方法。

Pattern Knowledge and Artificial Neural Network Based Framework for Intelligent CAD System

Experience is an important source of knowledge for intelligent computer aided design. In this paper, an approach of effectively making use of process experience is proposed. A group of feature based code numbers is assigned to each sheet metal part. This number is sent to an Artificial Neural Network(ANN) that has been trained by forming process samples. The experience of part forming process has been gathered into sample library, so the ANN (BP network) trained by it will learn the knowledge of sheet metal forming. When a new group of feature code numbers of a part is sent to this ANN, it will map generate the feasible forming process. The weight matrix of the trained ANN is called pattern knowledge. Consisted of ANN subsystem, process parameter calculation subsystem and a feature based part modeling subsystem, a framework of intelligent system has been built up to work out the sheet metal part forming process.