Research And Application Of Personalized Qos Preferences In Service Selection

The research of this thesis comes from the topic “The Architecture System ofEnterprise Application Software for the Survey&Design Based on SOA Technology”,which is one sub-project of the topic “Research and Application for Key Technology ofthe Survey&Design Enterprise Information”, which is one of the Key Projects in theNational Science&Technology Pillar Program during the Eleventh Five-Year PlanPeriod.At present, when users submit the same query request to the service registrationcenter, without considering users’ personal preferences, the current service selectionsystem usually returns the same result, but not customized services to users. For thatreason, this paper studys the users’ personal preferences in the service selection withQoS attributes of Web services, which is user-centered and based on semantic webservice selection.Through the research, this paper mainly solves the QoS personal preferencecalculation and based on that, personalized service selection problems. Depending onwhether users directly give QoS preferences, QoS attributes can be divided into implicitand emplicit, which can reflect users’ QoS personal preference exactly; Semantic webservice selection based on QoS personal preferences can return service sequence setsthat can meet different users’ personal preferences.This paper mainly completed the following tasks:①It analyzed the Ontology, Semantic Web services, QoS and other relevanttechnology, studied the necessity of bringing the personalized preferences based QoSinto the service selection, as well as the challenges and Solution mentality.②It researched different conditions and reasons of QoS preferences in users’requests, divided the requests into implicit and explicit. It researched the theory of thetraditional AHP (Analytical Hierarchy Process) and QoS classification and hierarchicaldivision in the service selection, it converted the qualitative users’ QoS preferences intothe quantitative values with the improved AHP, which is to calculate the QoS weight ofexplicit preferences, finally the calculation example was shown.③It proposed computing method of implicit QoS preferences weight. It solvedthe preference weight computing not directly given by the users, using a similar usergroup, and the users’ preference historical records. ④It researched the application scope, advantages and dis advantages of thecurrent user models, built a user preference model, which can effectively describe user’spreferences and can update automaticly; It designed a comprehensive QoS personalizedpreference weight calculation method, calculated implicit and explicit QoS preferencein the user service request.⑤Based on the proposed preference weight calculation method, it designed aWeb service selection mechanism based on QoS personalized preferences, it canprovide customized web service selection results according to different userpreferences.⑥The experiment results show that the proposed calculation and selectionmechanism is feasible and effectively.