Utilisation de normes et de réputations pour détecter et sanctionner les contradictions - Contribution au contrôle social des interactions dans les systèmes multi-agents ouverts et décentralisés

Abstract : Open and Decentralized Multi-Agent Systems (ODMAS) are particularly
vulnerable to the introduction of badly designed or malevolent
agents. It is therefore necessary to control such systems.

In this thesis, we propose the LIAR model, which enables agents to
control their peers' interactions, thanks to a reputation model.

Agents equiped with the LIAR model can first, represent interactions
they perceive with the help of a social commitment model. They can
also model the rules that each agent should follow thanks to a model
of social norms. By comparing observed behaviours with the norms they
know, agents are able to evaluate their peers and to estimate a
reputation level to associate to each of them. Agents are then able to
make a decision about the sanctions they wish to apply to their peers,
based on these levels of reputation.

Thanks to the complete integration of both steps: evaluation of the
perceived behaviours and decision of the sanctions to apply, the
LIAR model allows the agents to establish a fuly automatic social
control of agents' interactions.

Various experimentations have been conducted with this model in a
peer-to-peer context in order to show how agents where able control
their peers' interactions.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00134755
Contributor : Guillaume Muller <>
Submitted on : Monday, March 5, 2007 - 12:06:59 PM
Last modification on : Tuesday, October 23, 2018 - 2:36:03 PM
Long-term archiving on : Friday, September 21, 2012 - 12:25:38 PM

Identifiers

  • HAL Id : tel-00134755, version 1

Citation

Guillaume Muller. Utilisation de normes et de réputations pour détecter et sanctionner les contradictions - Contribution au contrôle social des interactions dans les systèmes multi-agents ouverts et décentralisés. Autre [cs.OH]. Ecole Nationale Supérieure des Mines de Saint-Etienne, 2006. Français. ⟨tel-00134755⟩

Share

Metrics

Record views

648

Files downloads

263