Quality Evaluation and Selection Framework of Service Composition Based on Distributed Agents

Abstract

In this paper, we propose a integrate framework for evaluation and selection of service composition based on distributed agent. Meanwhile, a distributed method, which is used for acquiring the user-based quality data of Web services (WSs) from the agents scattered in the network, is introduced. In the framework, the location and time factor, concerning quality of Web service composition, is considered. The procedure for processing the quality of Web service extends from single location to multiple locations. And the service is invoked by end-user under different times. By studying service quality of a lot of historic data in the evaluation framework, the LOA Algorithm, which considers location and time factor, is developed to optimize quality of service composition. From the view of the end-user, the distribution agents with different physical locations can improve quality of composite service. In the last part of the paper, pros and cons of framework and complexity of algorithm are discussed in detail.