A Research On Semantic Web Service Discovery Mechanism

Web service is a kind of new application program based on network environments and is the basic structural module of the distributed computing in the internet, it plays more and more important roles in E-commerce, Application integration development because of its characteristics of self-contained, self-described, loose coupling and modularity. Because of the number of published Web services increase sharply, the traditional search technology which based on string matching have more and more difficulty in capturing the right target service. The emergence of semantic Web service technology improves the problem highly and makes large task automation possible, such as automatic search, calls, synthesis, running monitoring and recovery. But the current semantic Web services discovery technology still has some problems, such as:①While publishing Web services, the definition of the QoS information contained in the services does not have uniform standard, which makes it difficult to reflect the user's QoS constraints during the semantic Web services discovery process;②While doing the functional matching between service requests and published services, the semantic matching algorithms is designed with defects , which caused the recall ratio and precision ratio of service finding is not satisfactory enough, this is mainly because the current semantic matching algorithms which are the core of the matching process only consider the included relation between concepts in ontologies except other semantic relations. These problems gradually become the bottlenecks of semantic Web services discovery technology developing that need to be solved urgently.The main research content of this dissertation is:①In the services description period, this dissertation adopts a description model contains function attributes and non-functional attributes, and defines each child characteristics of non-functional attributes , also the sub-properties calculation and evaluation algorithms are given, finally a sub attributes weighted QoS matching algorithm is proposed.②In semantic matching algorithm aspect, this dissertation proposes an improved semantic matching algorithms. The core idea is as follows: First the relationship between concepts in ontologies are divided into two kinds, namely included relations and other semantic relations, the other semantic relations are unified called the non-included relations here; Secondly, rebuild the traditional ontology, add the non-included relation into it, which changes the ontology from the classification tree to concepts' relation graph; According to the new building ontology, define different semantic distance between the concepts; Compute the new semantic similarity between concepts through the semantic distance ,and treated as semantic matching results;③Finally the main algorithm of the semantic Web services discovery is also given in the dissertation based on the QoS matching algorithm and semantic matching algorithm above.Finally, the simulation experiment is done to checkout the performance of the new algorithm, and the results show that the new algorithm enhances the result of services findings, the recall ratio and precision ratio is improved.