The ARAL team is interested in modelling and implementing properties associated with structured software agents.
In the MultiAgent Systems (MAS) field, the agent term takes on very different meanings, going from space point agent to mobile intelligent agent. Our work concerns this second type of agents, i.e. agents provided with no commonplace contents and an internal architecture.
Although the studied properties essentially belong to the MAS corpus, the ARAL team relied on two other disciplines as well: Systemics of Complexity (SC) and Distributed Systems (DS). SC provided the conceptual framework to modelize the studied properties (autonomy and mobility), DS the technical tools to implement the models kept.
The most significant obtained results are:
- a taxonomy of agents properties (attributes/qualities),
- a model of artificial autonomy (autonomy with regard to an attribute),
- a generic model of software agent (Systemion model),
- a peer to peer platform based on mobile agents (RetroMob) used in a CSCL (Computer Supported Collaborative Learning) context.
software agent, autonomy, mobility, systemics, complexity