Skip to Main content Skip to Navigation

Jugement éthique pour la décision et la coopération dans les systèmes multi-agents

Abstract : The increasing use of multiagent systems in various fields raises the need of autonomous agents able to take into account such ethical principles in their decisions. More and more propositions are published, but they are often agent-centered and they don’t consider the issues raised by the interactions between artificial agents and possibly humans, potentially using another ethics. Our goal is to give the agents the ability to reason on ethics to enable an ethics-based cooperation in multiagent systems. This work presents a model of ethical judgment for artificial autonomous agents in multiagent systems both useful to influence their decisions and behaviors, and describes an ethics-based cooperation framework. This model distinguishes the morality (or theory of the good), describing the goodness of actions in a context regarding a set of moral values and moral rules, and ethics (or theory of the right), describing the rightness of an action regarding a set of ethical principles. The use of this model in the decision process generates a conform behavior regarding the chosen theories of good and right. An agent may also use this model to judge the observed behavior of the other agents and employ this judgment to adapt its own behavior towards the judged agents. The detailed presentation of this model is followed by some experimentations to show the use of this model in a realistic application based on an ethical asset management scenario. The results show how the behaviors of the agents might be impacted and the efficiency of this model to discriminate the behaviors of the others.
Document type :
Complete list of metadatas

Cited literature [195 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Monday, July 30, 2018 - 11:31:06 AM
Last modification on : Tuesday, May 7, 2019 - 4:05:48 PM
Long-term archiving on: : Wednesday, October 31, 2018 - 12:51:09 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01851485, version 1



Nicolas Cointe. Jugement éthique pour la décision et la coopération dans les systèmes multi-agents. Autre. Université de Lyon, 2017. Français. ⟨NNT : 2017LYSEM043⟩. ⟨tel-01851485⟩



Record views


Files downloads