Leveraging human experts' knowledge to detect and publish compositions of Semantic Web services in a repository

摘要

Web services have added a considerable abstraction level to interact with applications regardless of their environment. Semantic Web services have augmented web services with rigorous models to describe web services' functionalities and how they could be chained with other web services to perform the desired task. Despite the explicit semantic models to describe Semantic Web services, there is no model that links the ability of UDDI to describe and host web services and the rigorousness of the description provided with Semantic Web services. In this paper we propose a publication and discovery approach that leverages human experts' knowledge to compose web services on-the-fly through expert systems that reason on what the web services repository contains and attempt to find composition patterns that can be provided by the published web services following expert-defined rules. We have implemented our approach using Java and JESS.