Ontology Generation And Evolution Framework Based On Heterogeneous Data Fusion

In today's manufacturing enterprises,the data heterogeneity problem commonly exists.Different divisions usually have their own individual systems to service for manufacturing.However,these systems are developed individually,so there exists large heterogeneity between divverent systems.Thus how to integrate heterogeneous data to form a complete view so as to manage enterprise data,and adjust the view according to data change so as to serve for enterprise production become a great problem.This paper uses the technology of ontology.Heterogeneous data from different data sources are integrated by generating a unified instance ontology,which is the complete view of data of the whole enterprise.Meanwhile,the ontology can evolute according to changes in data through ontology evolution so as to reflect the process adjustment of manufacturing process.Overall,this paper represents an ontology-based production support framework.The main contributions of this paper includes:(1)implementing a model which maps heterogeneous data to instancesThis paper focuses on the characteristics of relational database and designs a method to map heterogeneous data to unified instance model.Especially for the relation missing problem which is very commen in databases,this part implements the auto implementation of data relations so as to reflect the real situation of data and their relations through instance.(2)Proposing an ontology generation method based on instanceThis paper proposes an ontology generation method based on the instances mapped from different datasources.The process is devided into 5 steps: data property matching,concept matching,data property filtering,instance matching and object property matching.The method is fully based on property values instead of naming,so it can deal well with the heterogeneity brought by different data sources.(3)Introducing an ontology evolution strategy according to data changesThis paper introduces an ontology evolution strategy according to data changes.Data changes will cause changes in instances,which will lead to conflicts with restrictions.Thus adjusting the ontology following different rules for different types of conflicts can realize ontology evolution.Furthermore,the real application of ontology evolution is introduced according to core concepts of enterprise: process record,part,product and machine.(4)The prototype system and verificationA prototype system is designed and implemented according to the methods above.Experiments are conducted with the real data from a mould company so as to verify the effectiveness and adaptability to real life production.