Research For Web Services Selection In Web Composition With Mixed QoS Model

The research subject of this thesis comes from the topic "The Architecture System of Application Software for the Survey & Design Enterprise Based on SOA Technology", which is one sub-topic of the topic "Research and Application for Key Technology of the Survey & Design Enterprise Infomationization", which is one of the Key Projects in the National Science & Technology Pillar Program during the Eleventh Five-Year Plan Period.Web service is open, interoperable, integratable, encapsulated and loose coupled, etc., which makes it become the main technology to implement the applications based on Service-Oriented Architecture (SOA). When the Web services deployed in the network cannot fully meet the needs of the service requester, it is often solved by compositing some existing Web services selectively, which makes services selection be the key to services composition. Currently, the global optimal candidate services selection is one of the difficulties in the research for services composition.In the exsiting researches, the global optimal candidate services selection is mostly based on the determinate values of QoS attributes and weight factors. However, the values of some QoS attributes and the weight factors of all QoS attributes can not be exactly described in practice. Therefore, this thesis first proposes an approach for services selection to support mixed QoS attributes and uncertain user preferences in services composition, then analyzes and improves the process of the proposed approach, and finally verify the effectiveness and correctness of the proposed approach and its improvement by the simulations and the analysis for experiments.In this thesis, the jobs are completed as following:①It analyzed the relevant technologies for Web services composition and selection, which had included the relevant contents about traditional Web service and semantic Web services, the generation process and basic atomic structures of workflow model for composited service, the relevant contents about Quality of Service (QoS) and Intuitionistic Fuzzy Sets.②It researched the relvent content about the Web services ontology and the QoS ontology associated with Web services, which had researched the concepts of QoS attributes in upper and middle QoS ontologies, and classified the concept of QoS attributes in middle QoS ontology. ③It proposed the description and conversion methods for two fuzzy QoS attributes, the interval and the linguistic, in services composition. It proposed the method to describe the uncertain user preference and convert it into the weight factor for the corresponding QoS attribute. It researched and proposed the aggregation and computation method for the integrated QoS attributes by the basic atomic structures of workflow model for composited service.④It proposed the approach for services selection to support mixed QoS attributes and uncertain user preferences in services composition and gave the improvement strategy for splitting and optimization based on the workflow model for the proposed apporach.⑤Through the simulations and the analysis of experiments, it verified the effectiveness and correctness of the proposed approach for services selection and its improvement strategy for splitting and optimization, and verified that there are different impacts on the results of the proposed approach for services selection caused by the fuzzy QoS attributes and uncertain user preferences.