Consensus-finding that preserves mutually conflicting hypothetical information from a same agent

Article type: Research Article

Authors: Grégoire, Éric^{a; b; *} | Lagniez, Jean-Marie^a | Zhang, Du^b

Affiliations: [a] CRIL, Univ. Artois & CNRS, F-62300 Lens Cedex, France. E-mails: [email protected], [email protected] | [b] Faculty of Information Technology, Macau University of Science and Technology (MUST), Av. Wai Long, Taipa, Macau. E-mail: [email protected]

Correspondence: [*] Corresponding author. E-mail: [email protected].

Abstract: A consensus among logically conflicting agents is defined as a subset of all the information conveyed by the agents such that this subset does not conflict with the information of any of the agents. Hopefully, each agent might be inclined to endorse such a consensus since this latter one does not conflict with the agent’s own information. This form of consensus is investigated from a practical computational point of view within an extended Boolean logic-based knowledge representation framework. A method that is based on a single step discrete optimisation process to compute one maximal consensus is shown viable even for very large information sets. Interestingly, we allow each agent to require some additional hypothetical – possibly mutually conflicting – information to be point-wise compatible with any consensus.

Keywords: Artificial Intelligence, Boolean logic, consensus-finding, negotiation, multi-agents

DOI: 10.3233/AIC-170749

Journal: AI Communications, vol. 31, no. 3, pp. 303-317, 2018