A graph transformation based approach for multi-agent systems reorganization

During the execution of a multi-agent system, a variety of events originating from the external environment and/or the internal elements of that system can lead to a state of failure. To overcome this kind of undesirable situations, the multi-agent system has to reorganize its behavior and structur e in order to continue fulfilling its overall objectives. Several approaches are proposed in the literature that investigate the problem of reorganization at the design time. Unfortunately, we cannot detect all the events at design time, because several among them occur at run-time. In this paper, first, we propose an approach for reorganizing the multi-agent systems at run-time. Second, we propose to formulate the reorganization using graph transformation. In particular, we define a type graph to represent the multi-agent system organization and a set of rules that define the different actions that can be performed so that reorganizing the system. We evaluated our approach on a case study related to a Conference Management System which is implemented as a multi-agent system.