基于知识图谱和犹豫模糊理论的复杂产品设计知识检索系统

Complex product design knowledge retrieval system based on knowledge graph and hesitant fuzzy theory

针对复杂产品设计知识检索系统的检索精度低、人员使用要求高、匹配知识粒度大等问题,本文开发了一种基于知识图谱(KG)和模糊相似度的设计知识检索系统。结合设计知识需求及历史设计数据,开发知识表示模型以建立设计知识网络;基于知识图谱的三元组存储形式,构建设计知识检索框架;设计基于犹豫模糊理论的句子检索算法,从词和句子层面融合句子的特征属性,并给出不同属性的隶属度计算公式,对比验证了该算法的平均准确率优于其他相关算法。在上述方法和技术的基础上,以柴油发动机设计知识为例开发设计知识检索系统,通过两类设计知识检索实例,验证了所开发系统的有效性和实用性。

Aiming at the problems of complex product design knowledge retrieval system such as low precision,high requirements for users,large knowledge granularity,a design knowledge retrieval system based on knowledge graph(KG)and fuzzy similarity is developed.Combining reuse requirements and historical design data,a knowledge representation model is designed to establish a design knowledge network.Based on the three-tuple storage form of knowledge graph,a design knowledge retrieval framework is constructed.In addition,a sentence similarity algorithm based on the hesitant fuzzy theory is developed,which mainly fuses the characteristic attributes from the word and sentence.Further,the membership degree functions of attributes are given,respectively.And the algorithm proposed is compared and verified to be better than other retrieval algorithms.Based on the above methods,a design knowledge retrieval system is developed taking diesel engine design knowledge as an example,and the effectiveness and practicability of the system are verified through two types of design knowledge retrieval cases.