| 基于ART-2神经网络的设计知识需求研究 |
| |
| 引用本文： | 黄卫东,谢强,丁秋林. 基于ART-2神经网络的设计知识需求研究[J]. 机械科学与技术, 2008, 27(3): 399-404 |
| |
| 作者姓名： | 黄卫东 谢强 丁秋林 |
| |
| 作者单位： | 南京航空航天大学,南京,210016;南京邮电大学,南京,210003;南京航空航天大学,南京,210016 |
| |
| 摘 要： | 通过分析产品设计知识获取和知识共享研究中存在的问题,结合本体理论的运用,构建了产品设计中的概念语义,并利用OWL+RDF+XML形式化语言给出了案例知识的表达,在研究ART-2神经网络竞争学习和自稳机制的基础上,设计了案例的多级组织,实现了基于案例推理(CBR)的检索机制,提升了案例分类和检索的效率。同时,利用案例的聚类和学习机制,研究特征权值的自动获取和优化,并通过应用实验验证了该研究成果的实用性和有效性,从而较好地解决了产品设计中的知识需求。
|
| 关 键 词： | 知识需求 ART-2 CBR 本体 设计 |
| 文章编号： | 1003-8728(2008)03-0399-06 |
| 修稿时间： | 2006-10-18 |
On Acquisition of Product Design Knowledge by an ART-2 Neural Network |
| |
| Huang Weidong,Xie Qiang,Ding Qiulin. On Acquisition of Product Design Knowledge by an ART-2 Neural Network[J]. Mechanical Science and Technology for Aerospace Engineering, 2008, 27(3): 399-404 |
| |
| Authors: | Huang Weidong Xie Qiang Ding Qiulin |
| |
| Abstract: | We first analyzed the problems in acquisition of product design knowledge and knowledge sharing.Then we constructed the semantic concepts in product design based on ontology theory,and presented the representation of case knowledge by means of OWL+RDF+XML language.Utilizing the competitive learning capability and self-stabilization mechanism of an ART-2(adaptive resonance theory) neural network,we designed a case multistage organization,realized the retrival methodology in case-based reasoning(CBR),and raised the efficiency of case classification and retrival.Based on case clustering and learning mechanism,we also realized automatic acquisition and optimization of characteristic weight.Experiment results show that our method is practical and effective for solving the knowledge demand problem in product design. |
| |
| Keywords: | knowledge demand ART-2 CBR ontology product design |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
