A Tool Architecture to Verify Properties of Multiagent System at Runtime

This paper describes an architecture allowing to verify properties of a multiagent system during its execution. This architecture is the basis of our study whose goal is to check at runtime, if agents and more generally multiagent systems satisfy requirements. Considering that a correct system is a system verifying the properties specified by the designer, we are interested in the “property” notion. That is why we give here a definition of “property” and we present an architecture to validate them. The architecture, a multiagent system itself, is based on a set of agents which goal is to check at runtime the whole system’s properties. So after a brief description of the “property” notion, we describe our architecture and the way to check systems.