1. AOT Lab Dipartimento di Ingegneria dell’Informazione Università degli Studi di Parma Unifying MAS Meta-Models ADELFE, Gaia & PASSI Carole Bernon, Massimo Cossentino, Marie-Pierre Gleizes, Paola Turci and Franco Zambonelli

2. 2 TFG AOSE, Rome, July 2004 Agent-Oriented vs. Object-Oriented  Design a system  Instantiate a meta-model  OO context - design rely on a common denominator Universally accepted concept of object and related meta ‑ model of object-oriented systems  AO context - to date, no common denominator Each methodology has its own concepts and system structure In the agent world the meta-model is the critical element when applying the “method engineering” paradigm

3. 3 TFG AOSE, Rome, July 2004 Reasons to Unify  Meta-model = means to unify concepts  Unique meta-model, a way of unifying the different concepts  Helping designer to build a meta-model depending on the system he needs  Choice of useful elements From the “omni-comprehensive” MAS meta-model  Composition of process “Method fragment” selection phase  Model coherence checking by CASE tools  Unification eases model transformation

4. 4 TFG AOSE, Rome, July 2004 MAS Meta-Model for ADELFE no predefined organization local goal cooperative detect and remove NCS cognitive and behavioural representation

5. 5 TFG AOSE, Rome, July 2004 Gaia Methodology  First version  Designed to handle small-scale, closed agent-based systems  Modelled agents, roles, interactions  Missed in modelling explicitly the social aspects of a MAS  Official extension of Gaia  Thought for open agent systems  Focused on the social organization of the system

6. 6 TFG AOSE, Rome, July 2004 MAS Meta-Model for Gaia

7. 7 TFG AOSE, Rome, July 2004 MAS Meta-Model for PASSI

8. 8 TFG AOSE, Rome, July 2004 Towards a Unifying MAS Meta-Model  In order to define a unifying meta-model we will consider the following specific aspects:  Agent structure  Agent interactions  Agent society and organizational structure  Agent implementation

9. 9 TFG AOSE, Rome, July 2004 Agent Structure  ADELFE: cooperative agents  Gaia and PASSI: composition of roles

10. 10 TFG AOSE, Rome, July 2004 Agent Structure - Considerations  None imposes a specific classical agent architecture  Seen as some kind of low level architectures that can be adopted during MAS implementation  None explicitly deals with goals and plans  ADELFE Notion of goal is only used to determine skills Plans are built at run-time by the system  Gaia The concept of “goal” is implicit in roles Plans play no explicit role  PASSI Goals are considered as non functional requirements Plans are modelled as algorithms (activity diagrams)

11. 11 TFG AOSE, Rome, July 2004 Agent Interactions  No relevant differences

12. 12 TFG AOSE, Rome, July 2004 Agent Interactions  No relevant differences  ADELFE: agents can communicate indirectly using environment  Gaia: communication mediated by the environment seen as a side effect

13. 13 TFG AOSE, Rome, July 2004 Agent Interactions  No relevant differences  ADELFE: ontologies have not to be modelled; agents are able to adapt to the environment and other agents  ADELFE: agents can communicate indirectly using environment  Gaia: communication mediated by the environment seen as a side effect

14. 14 TFG AOSE, Rome, July 2004 Agent Society and Organizational Structure  ADELFE: no predefined organization

15. 15 TFG AOSE, Rome, July 2004 Agent Society and Organizational Structure  ADELFE: no predefined organization  Gaia: organization = primary abstraction  PASSI: services and scenarios

16. 16 TFG AOSE, Rome, July 2004 Agent Implementation  ADELFE  The problem of the system implementation has not been treated yet; no platform has been imposed  Gaia  Totally abstract from implementation  PASSI  FIPA compliant systems Direct map among the most important elements of the model and their implementation

17. 17 TFG AOSE, Rome, July 2004 Proposal for a Unifying Meta-model

18. 18 TFG AOSE, Rome, July 2004 Conclusion  Societies with or without predefined organizations (open systems)  Generic agent (cooperative or not) Follows organizational or cooperation rules  Possibility of identifying three domains  Improvements: an agent has a representation of the environment expressed via an ontological model  Step towards a unique meta-model  A unifying framework for the systems produced with different approaches

19. 19 TFG AOSE, Rome, July 2004 Open Questions  Can a unique meta-model exist?  What description level has to be reached?  How may designers choose meta-model elements?  What kind of tools may ease their work?