Automatic Composition and Mediation on Multiple-Language Semantic Web Services

Abstract

Web services have been attracting many attentions in both research and industry communities due to their potential of being invoked automatically based on users requests. In order to make this potential practical, there are two tasks needed to be fulfilled including (i) automatic discovery and composition of appropriate services and (ii) automatic invocation of sequenced services to achieve the desired goal. However, the lack of a mechanism of semantic representation in typical services descriptions has hindered this vision significantly. The recently introduced Semantic Web services are promising to overcome this obstacle by employing ontologies to describe the service operations, therefore the semantics of the provided services can be understood and reasoned automatically by computer programs. Nevertheless, since the standard for ontological knowledge representation over the Semantic Web services has not been completely established yet, there are many description languages have been introduced. Automatic composition and mediation of multiple-language Semantic Web services thus become a challenging issue.

In this paper we introduce an approach to tackle this problem. Based on extended plug-in technique for service discovery, we develop a framework for composition of multiple services and make use of concept similarity evaluated over ontological concepts to handle the problem of multiple service description languages. We also discuss automatic mediation of multiple services to have the involved services automatically invoked necessarily. To show the practical aspect of our approach, we have applied the composition framework on the most two popular description languages, namely OWL-S and WSMO. The initial results showed that our approach has achieved reasonable performance in terms of accuracy.