Semantic Web-based Network Manufacturing Resources Intelligent Search Technology Research

Network manufacturing makes cooperation between enterprises more efficient, lower cost, resources more fully used. It increases the ability of enterprises to develop new products rapidly, low cost. However, on one side, web information is in various forms, unstructured or semi-structured, machines unreadable and difficult to fetch. On the other side, normal search engines use keyword matching only, without semantic information, which lead to low recall rate and low exact rate of the query results. Therefore, fetching and querying networks resources is the bottleneck which restricts the development of network manufacturing technology.In order to solve this problem, this thesis combines Semantic Web technology to network manufacturing. We design and implement a prototype system, MRISS, which includes ontology modeling, manufacturing resource acquisition, semantic annotation, intelligent searching and semantic browser module.First of all, this thesis studies the status quo both at home and aboard about the network manufacturing intelligent search related technologies. Then we propose MRISS overall structure and a four-level system model.Secondly, the thesis establishes an ontology structure for network manufacturing resources, using OWL modeling language, including concepts, properties of concepts and the relationship between concepts. We also create many axioms for reasoning. All of these construct a complex network structure and make the semantic base of the whole system.Thirdly, the thesis analyzes current information acquisition methods and proposes that a customized spider is a reasonable way to solve this problem. Based on the open source Spider, Heritrix, we realize the Focused Spider which fulfilled crawling for the target sites. We also complete the semi-automatic annotation for the target sites.Finally, the thesis proposes multi-level model of intelligent search and the overall structure of intelligent search module. Using Jena tool kits, we implement multi-level model of intelligent search, with the user-friendly Semantic Browser. We define a query tree to connect the user input with semantic query language SPARQL and study the algorithm to converse the query tree to SPARQL. At last, experiment proves the features of MRISS.