In: 06, 2006: Proceedings of the Fourth International Workshop on Modelling of Objects, Components and Agents, MOCA'06, Bericht 272, FBI-HH-B-272, pages 145-160. June 06.
Abstract: Both process mining and multi agent simulations are relatively new research aeras, both evolving rapidly. In multi agent simulations, techniques are being developed to make agents more intelligent and more adaptive to their environment. In process mining, the focus has shifted from discovering a complete process model from an execution log, to the development of all kind of analysis techniques related to processes. In this paper, we combine the two research areas by showing that process mining techniques, applied to the communication logs of multi-agent simulations, can help in understanding the behaviour of agents. Furthermore, we show that process mining could actually help agents to adapt to their environment, by enabling them to discover why another agent makes certain decisions instead of just reacting on the fact that the other agent made a decision.
Keywords: Process Mining; Multi-Agent Simulation; Auctioning Systems.