Service Composition Requirement Model Based On Environment Ontology And Petri Net And Its Key Technology Reaserch

The web content is not only designed for humans to read, but also for computer programs to utilize effectively, In order to make the web services understand the content, semantic web service is proposed. Recently, with the development and advancement on the new generation of large-scale Internet applications, such as grid computing and cloud computing, it demonstrates that the technologies of semantic web service play an important part in current main distributed applications. The capabilities of semantic web will be reflected fully, when the user builds a program to collect the services distributed on web, deal with the data and share the information with each other. The most important resource in all the collected web services is semantic web service which supplies the service. Then the application program will manage the services automatically to meet the requirement of user finally.With the number of network users increasing, the requirement from them is not only roaring, not also various mostly. So it becomes a hot target chasing by academic and industrial communities that the pre-existing web services on the Internet are integrated and combined reasonably to generate a new value-added service. As an important part in the research and application of semantic web service, the automatic combination of semantic web service includes discovering, selecting, invoking and combining of the required web service, as well as satisfying the requirement of users.In the traditional service-oriented architecture, the web service is waiting passively for discovering and invoking; while in the auto web service environment, the service could find the requirement and gather around it according to the ability by itself. As the technologies of Internet developing, the number of web service is expanding day by day, and then it results in that there may be several web services meeting the requirement. So it is another research problem that how to select the best and the most satisfied web service from the numerous web services similar in function.Above the two fields, it is mainly researched that the model of requirement of web services, the service composition and the selection of web service. So the works done in this paper are as follows:(1) The requirement model of web services is completed further. Described by Environment Ontology, with the relative theory of Petri Net, the requirement is built as Petri net. First the requirement is looked as intentions by Environment Ontology, then some relationship between the discrete intentions is defined as the compatible relationship; a task is a set of intentions which are compatible with each other. The tasks are viewed as transitions in Petri net theory; finally the requirement of service compositions is modeled as a Petri net.(2) We put forward two algorithms to generate reachable tree and to get the service composition respectively. The algorithm first transforms the Petri net system, the requirement of service composition model, to a reachable tree. Through traveling the tree, it is achieved that the service compositions with functional equivalence.(3) The service composition is selected in an active service environment. QoS is brought in service composition selection. First the order of magnitudes of QoS is unified to standardize the non-functional attribute of services. According to the structure of model, the QoS of the service compositions is calculated. Finally the optimal service composition is selected from the similar services with local optimal combination methods by QoS.(4) Finally we design a travel arrangement case to verify the above research effective and feasible. In order to prove the theory researched in this paper, the classic case, a travel arrangement case, is chosen to be the final case. In this case, the processing of modeling the requirement of travel arrangement is analyzed in detail, and then the reachable tree is built according to the model. All the service compositions could be gained through the reachable tree, finally the optimal service composition is figured up.In conclusion, this paper tries to do something useful on researching the requirement modeling and service selecting, and presents the material description of problems and concrete methods to solve them, which makes a contribution to the application theory of the semantic web service.