Towards a multi-agent system for the supervision of dynamic systems

The supervision of dynamic systems is essential in industrial applications. The decentralization of such systems makes their supervision more difficult. In this article, we tackle the supervision problem by using a temporal scenario recognition approach. The supervision is performed by a society of agents, where an agent is considered as a watching process that is responsible for a subset of possible scenarios of the functioning of the system. The cooperation among agents is based on interaction protocols as well as dependence networks.