An Agent-Based Model for Autonomic Coordination of Inter-Organizational Business Processes

Inter-Organizational Workflow (IOW) aims at supporting the collaboration between several autonomous and heterogeneous business processes, distributed over different enterprises or organizations. Coordination of these processes is a fundamental issue that has been mainly addressed in a static context, but it still remains open in a dynamic one such as the Internet in which IOW applications are more and more enacted nowadays. In such a context, Multi-Agent Systems (MAS) are known to be a natural solution for modeling IOW since they provide adequate abstractions and specific mediators to cope with IOW coordination. Consequently, this paper provides an agent-based model for coordinating business processes involved in a dynamic IOW. This model is a triplet (E, M, R). E is the set of coordinated entities. It corresponds to the different business processes that may be published, discovered or deployed by IOW partners. M is the media supporting coordination. It is a multi-agent architecture compliant with the Workflow Management Coalition architecture and integrating specific components devoted to coordination issues. Finally, R is the set of rules governing the coordination. In our context, R is described through an organizational model aiming at structuring the interaction among the coordinated entities and the different components of the architecture.