基于神经索引实例与知识推理的混合型智能CAPP策略
A Hybrid Intelligent CAPP System Based on Neural Indexing Case and Knowledge
-
摘要: 提出一种新的基于耦合神经网络实例与知识的混合推理策略.采用面向对象的方法表达实例和知识,将神经索引模型引入实例推理中,在此基础上实现了CAPP的变异设计.同时,建立了基于零件及其工艺数据知识的分层知识表达与层次式推理机制,从而实现了CAPP的创成式设计.在给出这种智能化的CAPP系统的总体结构之后,详细讨论了基于神经网络的实例索引模型及实例推理和知识基推理的实现过程.由于吸收了派生法的类比设计思想,又具有创成的功能,使整个CAPP系统的推理决策具有更高的效率和更好的质量.Abstract: A new hybrid reasoning strategy based on case and knowledge is proposed.With the object-oriented case and knowledge representation,variant process planning based on the neural indexing case-based reasoning is realized.Meanwhile,generative process planning based on hierarchical knowledge classification and successive reasoning is achieved.After the overall structure of the CAPP system is presented,building of the neural indexing-models for case-based reasoning and implementation of hybrid reasoning based on case and knowledge are discussed in detail.This system possesses higher efficiency and better quality due to its capability of carrying out both variant analogy and generative creation.
下载: