面向知识重用的制造知识图谱生产问题查询方法  

Production problem query of manufacturing knowledge graph for knowledge reuse

作　　者：邹帅 鲁博 余长江 钱涛 王肖丛 孙浩 ZOU Shuai;LU Bo;YU Chang-jiang;QIAN Tao;WANG Xiao-cong;SUN Hao(Research and Development Department,Xinjiang Information Industry Co.Ltd,Urumqi 830000,China;School of Microelectronics,Shandong University,Jinan 250000,China)

机构地区：[1]新疆信息产业有限责任公司研究开发部,新疆乌鲁木齐830000 [2]山东大学微电子学院,山东济南250000

出　　处：《计算机工程与设计》2024年第10期3193-3200,F0003,共9页Computer Engineering and Design

基　　金：山东省重点研发计划基金项目(2017GGX201009);山东省自然科学基金项目(ZR2020MF157);山东省自然科学基金联合基金项目(ZR2013FQ006)。

摘　　要：针对制造企业多样化需求对复杂生产设计的表示,提出一种面向知识重用的知识图谱生产问题查询方法。利用统一的知识过滤器和元知识模型,构建基于本体的制造知识图谱;结合结构化本体查询语言和时态查询模式,研究知识-问题驱动的导航策略,提出一种语义相似度匹配方法。经实验验证,作为改进企业制造实际案例,通过该方法提高了问题查询精确性,得出了由知识术语和实例描述组成的有效解决措施,相比现有查询方法的评价指标精度平均高出15%以上,为知识重用回答生产问题提供了可行方案。Aiming at the needs of manufacturing enterprise for representation of production design,a production problem query of manufacturing knowledge graph method for knowledge reuse was presented.An ontology based manufacturing knowledge graph was constructed using unified knowledge filter and meta knowledge model.Combining structured ontology query language and temporal query pattern,a knowledge driven and problem driven navigation strategy was proposed,and a semantic matching method was developed.Experimental results show that as a practical case for improving manufacturing enterprise production,the proposed method can not only has high accuracy performance in problem query,but provide effective solutions consisting of knowledge terms and instance descriptions,while it can obtain a general increase in the average accuracy of evaluating indicators by about 15%compared to the existing methods.A feasible solution for knowledge reuse is provided to answer production questions.

关 键 词：知识重用 制造知识图谱 时态查询 知识-问题驱动导航 语义匹配 本体 知识过滤器 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]