|
Computer Sciences Seminar Enabling Agents in E-Business
Yannis Labrou Abstract This presentation outlines two building blocks for engineering agent systems in general, and for e-business applications in particular: (a) a language for agents to communicate knowledge and to enable complex agent interactions and (b) a language for expressing knowledge relevant to e-business. An Agent Communication Language (ACL) is a foundational component of complex agent systems. We will briefly outline contributions to the specification, semantics and protocols of Knowledge Query & Manipulation Language (KQML), the first ACL. KQML has enjoyed widespread use and has influenced the work of the Foundation of Intelligent Physical Agents (FIPA), which is an international standardization organization tasked with the specification of the standards necessary for the development of complex agent applications. An ACL provides the framework for supporting interactions between agents but we also need to capture the knowledge that communicating agents may exchange is e-business scenarios. We thus target the representation of business rules for expressing contracts, product & service offerings, business processes, etc. with the goal of supporting shared understanding and interoperability at both the representational and implementational layers. The proposed knowledge representation formalism is a generalized version of Courteous Logic Programs (CLP), which expressively extends declarative ordinary logic programs (OLP) to include prioritized conflict handling. The XML encoding of CLP, called Business Rules Markup Language (BRML) is suitable for interchange between heterogeneous commercial rule languages and for integration into XML-encoded ACL messages and eventually into the larger "web of meaning" ("semantic web") of interconnected information and knowledge sources. We also present examples of using those building blocks for agent applications in supply chain scenarios. The presentation will conclude with some insights on the implementational challenges of building agent systems in the "real" world of enterprise software and e-business. Bio |