Arbitration of Autonomous Multimedia Objects with a Multi

agents' organisation in which their behaviours are ruled by norms and ... Agents for Institution) dedicated to the organisation management and to the arbitration.
243KB taille 2 téléchargements 122 vues
Arbitration of Autonomous Multimedia Objects with a Multi-Agent System Benjamin Gâteau and Djamel Khadraoui CRP Henri Tudor – 29 Av. John F. Kennedy – L1855 Luxembourg, G.D. Luxembourg {benjamin.gateau, djamel.khadraoui} Extended Abstract The research work deals with the iTV domain and concerns the creation of multimedia contents with which a televiewer interacts. Multimedia objects involved in this creation are considered as autonomous entities able to adapt and modify their behaviours to the modification of the environment and its related scenario. They respect the definition of RAMO objects given in [1]. We consider here a “questions–answers” TV game show opposing a team of televiewers being at home and represented by Avatars and a real players’ team. Avatars are directly controlled by their correspondent televiewers. In order to define rights and duties of autonomous and generic entities assisting televiewers by means of unambiguous specifications, we think that the use of the Electronic Institutions principle [2] based on Multi-Agent System [3] is in concordance with our needs. Content Production

Rendering at End-User Level

RAMO Description Schemes’ Library RAM O D S

S ta te s D S

B eh avio rs D S S ta te s & B eh avior s R elatio ns

S pa tio T em p ora l D S

F eatu res D S

P roce sse s DS

F eatur es & S .- T . R e lation s D S


E ven ts & Im p acts D S P roce sse s & E ve nts R ela tion s D S


Inst Organisation Specification

Authoring/Edition SS



Conversion Tool Mapping tools XSL Transformation Code Generation

Scene Composition MM Content Description States & Behaviors Custom. Interactions - Events

FS Medial Libraries

MHP Java Code Xlets code MHP Meta-data

Figure 1: Global view of MOISEI Organisational Structure in a TV content production process NST

We define our Multi-Agent based Electronic Institution for Interactive Games as an autonomous agents’ organisation in which their behaviours are ruled by norms and controlled by an arbitration system. The role of this arbitration system consists in rewarding or punishing agents when they respect or not their agreements. Two kinds of agents are used: agents on which RAMO and so Avatars are built and an institutional multi-agent middleware called SYNAI (SYstem of Normative Agents for Institution) dedicated to the organisation management and to the arbitration. All agents are organised according to a model described with the MOISEI (Model of Organisation for multIagent SystEm) meta-model [4]. This way, agents are able to reason on the specification. They have the possibility to decide to take it into account or not. The institutional middleware reads this specification in order to supervise and control the agent-based Avatars. We integrate the definition of the organisation in the content production of the iTV content creation chain (see Figure 1). NST

MOISEI is composed of: (i) a Structural Specification (SS) defining roles that agents will play and relations between these roles; (ii) a Functional Specification (FS) defining all goals that have to be achieved in the system; (iii) a Contextual Specification (CS) defining the different contexts influencing the dynamic of the organisation as well as transitions between contexts; (iv) and a Normative Specification (NS) defining clearly rights and duties of roles on missions (set of goals) in specific contexts. These four specifications form the Organisational Specification and are detailled and illustrated with the iTV game show example in [4]. NST

An Organisational Specification is instantiated when agents adopt roles, commit on missions and achieve goals. The result is an Organisational Entity. It represents for instance in the iTV game domain a play. As agents are autonomous they can act on the organisation by not respecting its specification. If the OE is not consistent with the OS, the organisation is incoherent. To avoid that an arbitration system filters actions on the organisation to verify they respect the specification. It receives requests from agents, detects if they violate or not constraints and act on the organisation. For instance it verifies that an agent is authorized to commit on mission (answer a question) according to the role it is playing (Team's chief) in a special context (third round). Arbitration System represented by SYNAI is composed of agents organised in the same way that application agents. Our desire is to provide an Electronic Institution with the MOISEI meta-model allowing users to define organisation and a common arbitration system able to arbitrate all organisation. Thus the user doesn’t integrate the arbitration specification into his modeling. For our example, the user doesn't need to specify how the Avatars organisation is arbitrated. He only has to define the sanctions he want to applied. That is possible with the separation of “domain” specification and arbitration specification which means that the game rules are not hardcoded in the Avatars. NST

References [1] S. Renault, F. Meinkohn, D. Khadraoui and P. Blandin. Reactive and adaptive multimedia object approach for interactive and immersive applications. In International Conference on Information & Communication Technologies (ICTTA), Damascus - Syria, April 2004. [2] D. C. North. Institutions, Institutional Change and Economic Performance. Political Economy of Institutions and Decisions. Cambridge University Press, 1990. [3] M. Luck, P. McBurney, and C. Preist. Agent Technology: Enabling Next Generation Computing (A Roadmap for Agent Based Computing). AgentLink, 2003. [4] B. Gateau, O. Boissier, D. Khadraoui and E. Dubois. MOISEI : An organizational model for specifying rights and duties of autonomous agents. In 1st International Workshop on Coordination and Organisation (CoOrg), Namur - Belgium, April 2005. NST