Research On Semantic-Based Web Service Discovery

Web services discovery is to find out services which are the most similiar to services description information provided by requesters.At present ,semantic web services discovery has become a issue in rearch field,but many of them only considered the matchmaking in services capabilities,not in basic information and QoS(Qos of Service)—unfunctional properties.On the other hand,the results of matchmaking is represented by match degree.All of these resulted in some limits.This thesis made a deep reseach in sevices description and services matchmaking.Firstly, introduced the relevant background knowledge and research status of Semantic Web Servece and Web Servece discovery; secondly, analyzed several models of demain ontology description and put forward our domain ontology model OCModel and semantic web service description model SWSDM.In addition ,to support the match in QoS, we expanded OWL-S to enable to describe the QoS ,also gave a simple model of QoS;thirdly, introduced a multi-layer matchmaking method called "five levels and three stages ",which adopt similarity measurement to matchmake services. The results of qualitative and quantitative analysis on performance of the matchmaking method showed that it would improve the precision and the recall in semantic web services matchmaking.Then discussed every layer of match in detail;lastly,on the base of proposed service discovery engine,realised the original model and decribed the detail of the multi-layer matchmaking process using instances analysis.