FRANCISCO GARCÍA-SÁNCHEZ, RAFAEL VALENCIA-GARCÍA, DAGOBERTO CASTELLANOS-NIEVES AND JESUALDO T. FERNÁNDEZ-BREIS
Department of Informatics and Systems University of Murcia Espinardo, 30100 Spain
Some controversial approaches have derived from the discussion about the scope of the well-established Multi-Agent Systems technology,and the functionality expected for the brand new Semantic Web Services technology. While some authors in the field of intelligent agents state that Semantic Web Services can be also viewed as services provided by agents distributed all over the Internet using semantic mark-up, others believe that the agent functionality can be integrated in the core of intelligent Web Services. We establish a setof parameters by means of which these technologies can be evaluated and compared. From the results of this study we conclude that both technologies are completely interoperable and the strength of their union depends on the way it is achieved.
Keywords: intelligent agents, multi-agent systems, web services, semantic web services, semantic web, ontologies
The IntelligentAgents (IA)  field has been broadly studied over the last 30 years and nowadays the topic under question is being revisited thanks to its relation to the Semantic Web (SW)  and the potential benefits that can be reached from their potential integration. Intelligent Agents emerged due to the promising benefits of having applications with a technology that allows applications to decide forthemselves what they need to do in order to satisfy their design objectives. Nevertheless, agents face the problems derived from the lack of structure in the information published on the current Web. On the other hand, Web Services (WS)  have arisen as the best solution for remote execution of functionality. This is due to some promising properties such as operating systems and programming languagesindependence, interoperability, ubiquity, loosely coupled
This paper was partially supported by xxxxxx.
GARCÍA-SÁNCHEZ F., VALENCIA-GARCÍA R., CASTELLANOS-NIEVES D. AND FERNÁNDEZ-BREIS J.T.
applications, etc. However, as the Web grows in size and diversity, there is an increased need to automate aspects of WS such as the discovery, execution, selection, composition andinteroperation. The problem is that current technology around WS (i.e., UDDI, WSDL, and SOAP) provides limited support for all what has been pointed out before . The emergence of the SW  may involve the overcome of the mentioned drawbacks for both IA and WS. SW technology aims at adding semantics to the data published on the Web (i.e., it attempts to establish the meaning of the data), sothat machines are able to process these data in a similar way a human can do. Therefore, IA can exploit the semantics to perform their assigned tasks in a better way. Furthermore, WS capability descriptions can be semantically annotated leading to what is called Semantic Web Services (SWS) . SWS allow for automatic discovery, composition, and invocation of WS. As a consequence, it can be thoughtthat the next obvious step towards more powerful Web applications should be the mixture of IA and semantic WS technologies. However, some controversial approaches have appeared at this point. While some researchers claim that services are merely a further technological option to implement agent systems, others state that agent behaviors can be implemented in WS wrappers thus obtaining intelligentWS. An additional approach that is arguably the most extended among researchers, is to consider that both technologies are complementary and can work together to make the Next Generation Web become real [3, 12, 13, 14, 16, 17, 27, 28]. The aim of this document is to clarify as much as possible whether IA and SWS are compatible or not. With this goal in mind, a comparison framework has been...