Agreeing on Institutional Goals for Multi-agent Societies
Lists
Only you see this
Abstract
We present an argumentation-based approach to the problem of finding a set of institutional goals for multi-agent systems. The behaviour of the autonomous agents we consider in this paper is goal-directed, driven by either individual or common goals. When multiple agents want to set up a collaboration framework (for themselves or others to use), they do so by forming an institution (or organisation). The goals of such institution must be agreed upon by the agents setting up the framework before it can be executed. We propose to use argumentation, and in particular assumption-based argumentation, to facilitate the negotiation of institutional goals. We first describe a centralised approach and then provide the rationale for and detail our preliminary efforts at de-centralising the problem. We propose to employ the argumentation system CaSAPI as a tool to reason about the collaborative goals of the institution. Our approach mitigates concerns about performance bottlenecks and vulnerability of the system while providing, to some extent, privacy to the individual members of the institution.
BibTeX
@inproceedings{gaertner2008agreeing,
author = {Dorian Gaertner and Juan A. Rodríguez-Aguilar and Francesca Toni},
title = {{Agreeing on Institutional Goals for Multi-agent Societies}},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2008},
url = {https://openreview.net/forum?id=TLvZHj0l6K},
}Loading related content...