Knowledge Management And Initiative Recommendation For Business-level And Larger-granularity Service Composition

The Service-Oriented computing technology can integrate applications with standards and loose-coupling manner, which has provided a new chance to participate in constructing applications for the end-users on their own initiative. But how to make full use of services resource and construct personalized application by composing existing services on demand in the dynamic and open environment is a challenge issue. To support End-User-Programming and reuse composition knowledge, this paper discusses the key problems of how to extract, manage and reuse the composition knowledge in the Service-Oriented applications, and presents a service composition template based approach to larger-granularity business-level composition knowledge management and initiative recommendation-TryBest, The main research work and contribution of this dissertation is described as follows:(1) Proposed a novel approach to service composition knowledge management and recommendation- TryBestWe introduce the thinking of software reuse into composition knowledge extraction and use; propose an approach to service composition knowledge extraction by using make use of the technology of case based rule and VINCA process. With the approach and the composition logic, we define a service composition knowledge template, which using pattern, quality of service and service resource can configure flexibly. End-users can construct application by configuring the parameter according to their own individuation requirement and realize the service optimal selection by initiative recommendation at runtime.(2) Proposed domain ontology based template selection algorithm and content based service recommendation algorithmWe introduce domain ontology into composition template, provide the semantic description with domain ontology, and facilitate the management of composition template. We propose domain ontology based template selection algorithm, which overcomes the shortages of existing selection algorithm based keyword lack of semantic description. Furthermore, propose content based service recommendation algorithm, which shields the realization detail of composition for end-users and facilitates them using.(3) Proposed QoS model of service composition template and service recommendation algorithm based on global QoS constraintThe most service selection arithmetic is aimed at the function and non-function property of one service, our approach is aimed at the global non-function property of composition process and relationship between services, which turn service selection into initiative recommendation...