Research On Method Of Semantic Based Web Service Discovery And Composition And Its Application In Medical Service Integration

Because of its interoperability and reusability, web service has become the core support technology of Service Oriented Architecture. In order to deal with changeable business, how to utilize numerous web services to build loosely-coupled applications has become a research hotspot.Current researches on web service composition problem can be divided into two directions: process driven web service composition and semantic driven web service composition. Based on workflow techniques, process driven web service composition method has the advantage of convenient implementation but lack of flexibility and dynamic characteristic. Semantic driven web service composition focuses on web service’s self-description characteristic, by adding semantics to web services automatically generates the composite service through AI planning etc techniques. Comparing with process driven methods, the semantic driven methods improve the flexibility but cannot guarantee the availability. Aiming at the enterprise application integration of medical service domain, this thesis mainly researches the architecture of semantic web service composition, service matchmaking and selection mechanism, automatic service composition problems etc., detailed as follows:(1) Based on the characteristic analysis of enterprise service integration domain, aiming at how to solve the dynamic adaptability problem when the traditional process driven service composition method faces the flexible and mutable business process in complex network environment, this thesis proposes a new semi-automatic service composition method which describes the dynamic sub-process with abstract service, and replaces it during the runtime using semantic service composition method.(2) Aiming at the problem of service discovery and selection at the service runtime, this thesis discusses two kinds of logic-based semantic web service matchmaking scheme and proposes a hybrid matchmaking algorithm meets the abstract service’s functional requirement. And a QoS-based service selection algorithm is proposed to meet the abstract service’s non-functional requirement.(3) When there is no single web service can meet the requirement of abstract service during the runtime, this thesis proposes a semantic-backward tree based web service composition algorithm. Firstly, the algorithm uses a semantic binding matrix to save the semantic relations between web services, then generates the semantic-backward trees according to the abstract service, and finally selects and composes web service based on the service’s QoS profile.(4) Based on the methods and techniques mentioned above, this thesis designs and implements the framework of a semi-automatic web service composition. And a complete medical treatment process is shown in this framework which tests and verifies the robustness and flexibility of the composed system as well.