Research Of Inference Engine Oriented Workflow Based On Ontology

Tim Berners-Lee brought forward the concept of Semantic Web in 1998.and set the goal as make computer process information automatically by adding metadata to traditional web data to make the latter machineunderstandable. The clearer meanigs of data and domain thoery (Ontology) are supposed to provide high quality services for Web. One major job of making the web data machine-understandable is to let computers understand the connotation of data. Computers'compreh- ensive capabilities rely mostly on reasoning techniques. So the reasoning tachniques play a key role in semantic Web.On the other hand, to make the job working orderly, efficiently, high-rhythmically, we bring in the conception of overflow management technology. The main feature of overflow management technology is realizing automation in the process of people and computer interacting event. The main related work of overflow are: the whole treatment process of the task; the text file exchanged between the workgroup members by a set of defined, integrative rules and the estabished public target; all sorts of media information or task related information. Overflow management system is the software system applied in distributed environment, implementing the coordination among task processes and the cooperative processing. Because of these features of overflow management technology, we need to combine semantic and overflow model urgently. As a result, overflow management system will provide better abilities in semantic understanding, analyzing and reasoning. So, the introduction of ontology theory to overflow management system improves the efficiency and veracity of the system enormously.The integration of the workflow management system and the brainpower ontology is belong to Business Peocess Integration (BPI), BPI is a technology that has been evolving dramatically and gaining more and more acceptance in the real-world application.Research on that subject is important for"isolated island"problem, semantic search and brainpower reasoning problem, the integration of workflow process model, and the realization of flexible and high-efficiency business process. This article comes from a project cooperation of Jilin University & West Virginia University University.Thepurpose of the project design a new workflow system that the integration of the traditional workflow system and the semantic reasoning system on a base of the ontology theory. By introducing ontology base and realizing the ontology definition of flow service, domain knowledge and user interface, the system is able to incorparate collaborative knowledge dynamically and thus is flexible enough to apply in other domains which have the same application patterns.The extension and upgrade of the system can be achieved by updating and adding the definition of ontology, which is one of the distinct features of the system. Using this system, we construct an intelligent teaching system, which is built by college teaching knowledge, provides knowledge agent attributes, and has the ability of reasoning and self-learning. This workflow system is provided with better capability of correctness validation and semantic search.First of all, this paper studies on the basic conceptions and technology framework of overflow system and ontology, and brings forward the correctness standard of overflow model. Then, this paper analyzes the status of ontology in the overflow engine in collaborative work, through applying the operations of storage, query, verification and reasoning to the overflow model based on ontology, gives strongly support to the upper overflow engine, provides the overflow system based on ontology with the abilities of intelligent search and self-learning, thereby, improves the efficiency and veracity of the overflow engine enormously.The center of gravity for the article emphasize particularly on two aspects. (1) Base ontology of workflow file is modeling memory and query. (2) Base ontology of workflow modeling is validation and reasoning. The first part of the article implements pasing RDF(S)/XML based on SAX parser and mapping RDF(S)/OWL into EkSarva_ODM, and storage, inference and query of ontologies in relational databases. And this part implements the query subsystem based on RDQL.RDQL is query language with RDF, which inherits SquishQL and rdfDB. Although RDQL isn't formal standard, but RDQL is implemented far and wide by the RDF frame. RDQL allows expressing the complicated query concisely, the query engine implements burdensome work which accesses data model. The second part of the article implements simple reasoning.A reasoning based on Ontology has various applications: EkSarva_ODM describes a structural inference engine which performs TBOXand ABOX reasoning based on RDBMS and provides Pattern-like query mechanism to promote data mining. Furthermore, the paper discusses tha error that may be in the workflow model, and it gives different methods to different kinds of errors. The paper designs a method based on workflow net and the reachability tree of Petri net to examine the complicated errors. This method can find out those complicated errors: the dead lock,the dead loop,the redundant node and so on. Using this method, the modeling tool can create accurate workflow model. And finally, the paper designs the predigestion of the model from Ontology to graph and the transformation of the model from directed graph to Petri net.The paper researches the permanent storage of the workflow modeling based on Ontology and the reasoning based on Ontology and model verifying using Petri net, and realizes a workflow system based on Ontology and the modeling tool with examining errors. And the tool reaches the requirement of Electronic Government Affairs.