Research On Process Matching Method Based On OWL-S Service Model

As a new standard accessing to Web application, Web services is a good solutionfor the integration and sharing of highly heterogeneous data and application. In the fieldof Web services, service discovery is one of the key issues being studied and it is thebasis of services composition，service invocation and execution. However, with therapid development of computer applications, it has become increasingly difficultto,accurately and efficiently, select the service to meet the demand from the mass ofWeb services.Traditional Web service discovery method based on keyword matching has a lowerrecall rate and precision lacking of semantic information. To solve this problem, theontology technology is introduced into the field of Web services. Meanwhile theresearchers propose service discovery method based on semantic Web services. Thesemethods perform the match between two Web services by means of ontologytechnology and precision and recall rate is greatly improved. However, the abovemethods are based on the function information matching i.e. OWL-S ServiceProfileinformation matching and lack of considering the service interaction details. For thecomposite services, the two services might have the same input and output parameters,but has a different interactive behavior. The above services discovery methods can notdistinguish between the two different services. Therefore, the service discovered by theabove methods might not perform service interactions correctly.Aiming at the problem, this paper has a in-depth research on OWL-SServiceModel which descript the details of the service process and proposes a discoverymethod based on OWL-S ServiceModel match. Firstly, it defines the transforation rulesbetween the process graph and eight control structures the ServiceModel contains andthen in accordance with the rules，ServiceModel is equivalent formed into processgraph. Secondly we strike a process graph matching algorithm based on process graphsimilarity. Finally, we obtained the ServiceModel similarity of the two services bycalculating the process graph similarity to address the ServiceModel matching problem.Thus, the service discovery precision and recall rate is improved further.The basis of the theory, this paper designs the process matching model, conductedin-depth analysis of the modules, and code to achieve each module. Finally, we test thesystem respectively from ServiceModel transformation, matching as well as precision,recall rate and response time. The experimental results show that this methodcan convert ServiceModel correctly and match accurately, and has a higher precisionand recall ration. It verifies the feasibility and effectiveness of the proposed method.