The traditional perception of isolated data systems is changing to a new one where the interest of a real and efficient interoperation among those data systems is recognized. However, many problems must be solved yet before a real interoperation becomes true. In order to overcome the existing problems, there is a considerable number of proposals that can be found in the specialized literature that promote the idea of semantic interoperability. A new framework to achieve semantic interoperability among agent based data systems at a communication level is proposed in this paper. This framework permits agents belonging to different data systems 1) to send each other suitable messages without requiring the establishment of a common communication pattern in advance; 2) to understand, completely or partially, those messages that are interchanged among themselves; and 3) to invoke web services offered by the data systems at a high level without needing to go into technical details. An ontology that captures the semantics of different communication acts is the key element for supporting the functionalities provided by the framework. Furthermore, the framework has been extended to support semantic descriptions of web services, which favor their automatic discovery. The usefulness of the presented framework is evaluated using two case study of interoperation among heterogeneous data systems; on one side through the agents of those systems, and on the other side, through the combination of an agent and a web service.
展开▼
