Agent-Oriented Modeling of Business Rules and Business Processes: The Case of Automatically Guided Transport Systems∗

Business rules are statements that express (certain parts of) a business policy , defining business terms and defining or constraining the operations of an enterprise, in a declarative manner. Since these rules define and constrain the interaction among business agents in the course of business processes, they have to refer to the components of their mental state, such as the knowl-edge/information and the commitments of an organization. We propose an agent-oriented approach to business rules and show how to represent and visualize business rules and business processes in Agent-Object-Relationship diagrams using the example of Automatically Guided Transport Systems. * Vorgetragen bei der Verbundtagung Verteilte Informationssysteme auf der Grundlage von Objekten, Komponenten und Agenten (vertIS2001), Bamberg, Oktober 2001, erscheint im Rundbrief der GI-Fachgruppen EMISA und MobIS. Agent-Orientation is emerging as a new paradigm in software and information systems engineering. It offers a range of high-level abstractions that facilitate the conceptual and technical integration of communication and interaction with established information system technology. Agent-Orientation is highly significant for business information systems since business processes are driven by and directed towards agents (or actors), and hence have to comply with the physical and social dynamics of interacting individuals and institutions. While today's enterprise information system technology is largely based on the metaphors of data management and data flow, and is under pressure to adopt concepts and techniques from the highly successful object-oriented programming paradigm, Agent-Orientation emphasizes the fundamental role of actors/agents 1 and their mental state, and of communication and interaction, for analyzing and designing organizations and organizational information systems. This turns out to be crucial for a proper understanding of business rules. Since these rules define and constrain the interactions among business agents, they have to refer to the components of their mental state, such as the knowledge/information and the commitments of an organization. We attempt to show that our agent-oriented approach, that is based on the Agent-Object-Relationship (AOR) metamodel proposed in [Wag01a, Wag01b], allows to capture more of the dynamic and deontic semantics of enterprise modeling than object-oriented modeling approaches, such as the UML, do. Taking into account that the main motivation for object-oriented modeling stems from software engineering and not from