Abstract:
“Software as a service” approach has become a reality since efforts of both industry and research focused on service integration on the web. It would be possible to see the true potential of such a distributed infrastructure when those services can be joined together as parts of a workflow in order to collectively achieve combined functionality. The collective composition requires adaptive, dynamic systems where all parties are subject to unexpected changes. It also requires intelligent systems that are able to compose services as building blocks to generate a single service that meets requirements. Web Services, in our context, are mainly information services. These services produce information rather than changing the world. The other type of services are called world altering services. These services change environmental view of the world. Such services introduce state complexity into composition process and make it harder to determine next actions of composition process. In this research we present an agent based system that guides service discovery and service composition by integrating graph theory into the Web Services domain. As the momentum around Web Services grows, there is an increasing need for effective mechanisms of Web Services interaction. In order to test our approach we developed an agent based system that generates composition plans by using techniques inherited from network theory. The proposed system is applied on a prototype that depicts Semantic Web innovations and agent interactions while targeting to generate effective Web Services composition plans in an efficient manner.