The Research Of Service Composition Based On User Context

With the increasing prosperity of the online Web service resources, users slide into the plight of "information overload". How to accurately and quickly find out the Web services required by the user and provide effective recommendation has become an great challenge. Current research and Web service application status has the following limits:(1) the existing service recommendation method are generally based on user’s functional requirements and quality preferences, ignoring the fact that the preferred service of a same user may differ under different contexts;(2) the existing Web service recommendation method merely focused on online functional properties FQoS and the quality attributes QoS, which is far from enough.The020(Online To Offline) mode is prevalent recently, which brings offline business opportunity and the Internet together and makes Internet the entry of offline transactions. So customers tend to be more concerned about the functional properties and quality attributes of the offline service than online service. In addition, with the development of the Internet economy and modern services, meta Web service is no longer able to meet customers’demand, so combining the existing Web services together to provide high value-added services becomes an effective solution.To address the issues mentioned above, we did much analysis on existing research work about user scenarios and Web services technology, then the paper explored a context-based composition service recommendation method, from the view of personalized and collaborative management. The main work includes:(1)Designed the user context classification structure system, introduced the user scenario modeling method, the original contextual information processing method, and user scenarios management method considering user context information storage and update.(2)Proposed an online and offline service description model based on users’ comprehensive need for the both kind of services, designed FQoS and QoS index system, defined QoS attributes formula, and built a ontology meta model. All these work provided solid theoretical basis for service discovery and composition recommendation.(3)Designed user context-based three-level service resources discovery mechanisms,①service discovery based on functional requirements;②service discovery based on user context reasoning, built a reasoning rule between service ontology model and user context ontology, on the basis of relationship between the two sides③service discovery based on user scenarios clustering, since users under similar context are likely to choose the same service, so it is possible to select the most frequently used service among the user cluster,and recommend it to the target user. (4) Proposed service composition recommendation algorithm aiming at optimizing global QoS under cost constrains. In the case study, this paper took the composition of online travel services as an example, and proved that the proposed method is effective.