Agent-Based Services for the Semantic Web
August 1, 2004
The semantic web suggests having data and information on the web defined and linked in a way that it is both human readable and machine understandable. Machine understandability means that data has been explicitly prepared for programs and softrware agents to reason about and reuse. However, most current software agents have not been designed to run in a web environment: the software agents and web components are working in different models. For example, most software agent frameworks and tools do not support reasoning over content expressed in a mixture of several ontologies. Web components, on the other hand, lack notions of autonomy and a rich message-oriented communication. This thesis suggests adding to a semantic web site a dedicated service agent, capable of understanding the ontologies used by the site and answering queries about the content found on the site. Semantic web tasks are accomplished by the cooperation of personal agents, service agents and semantic web servers. The semantic web language OWL is used as the agent content language for message passing and interaction. An OWL inference engine, F-OWL, has been developed to manage logical sentences explicitly stated in the web documents as well as those that can be inferred. Within this thesis, TAGA system, a Trading Agent System running in the open Agentcities platform, is used to support the hypothesis and demonstrate how the software agent and semantic web paradigms can be integrated.
We see two main contributions in our work. First, this thesis presents an enhanced semantic web vision involving software agents, semantic web and web services. Both human and software agents can directly access the semantic-marked web pages through web interface. An agent communicates with other agents by exchanging messages encoded in a semantically rich agent communication language. Both web services and agent services are described in a semantically rich language to improve the interoperability. The personal agents and service agents work together to understand the semantic web content, automate web services and better serve humans. Second, the TAGA system attests this semantic web vision and provides a flexible environment for simulating agent-based trading in dynamic markets.
UMBC, Department of Computer Science and Electrical Engineering
Downloads: 5738 downloads
Google Scholar Citations: 4 citations