Conflict Resolution in Norm-Regulated Environments Via Unification and Constraints

Abstract

Abstract. We present a novel mechanism for the detection and resolution of conflicts within norm-regulated virtual environments, populated by agents whose behaviours are regulated by explicit obligations, permissions and prohibitions. A conflict between norms arises when an action is simultaneously prohibited and obliged or prohibited and permitted. In this paper, we use first-order unification and constraint satisfaction to detect and resolve such conflicts, introducing a concept of norm curtailment. A flexible and robust algorithm for norm adoption is presented and aspects of indirect conflicts and conflicts across delegation of actions between agents is discussed.

Detection and resolution of normative conflicts in multi-agent systems: a literature survey

“…3.2 Unification and constraints [Vasconcelos et al, 2009] The work presented in [Vasconcelos et al, 2009] integrates and extends the research presented in [Vasconcelos et al, 2007b;Kollingbaum et al, 2008a;Kollingbaum et al, 2008b;Vasconcelos et al, 2007a]. The definition of norm used was introduced in [Vasconcelos et al, 2007a], in which norms are associated with constraints.…”

“…The most common cases of conflict occur (i) between a norm that obliges and another one that prohibits the same behaviour; or (ii) between a norm that prohibits and another one that permits the same behaviour. When there is a normative conflict, whatever the agents do or refrain from doing may lead to a state that is not norm-compliant [Kollingbaum et al, 2008b;.…”

Normative conflict resolution in multi-agent systems

“…A simpler version of the conflict resolution mechanism was originally proposed in [41]. However, that work did not address norms with arbitrary constraints-these were introduced in [22], together with an extension of the mechanism to resolve conflicts making use of constraints. That work also introduced policies and indirect conflicts, incorporating these concepts into a norm adoption algorithm.…”

Déléguer la protection des données personnelles à des agents cognitifs

“…Dans [15], cette résolution se fonde sur une fusion des différents rôles joués par un agent, alors que dans [16] les conflits surviennent lorsque plusieurs sources de normes entrent en conflit, ce qui correspond mieux au cas du contexte d'exécution de l'agent PAw. Comme dans la plupart des travaux similaires [31], la notion de conflit correspond a l'inconsistance d'un système SDL, et sa résolution s'opère en fusionnant des ensembles de normes conflictuels entre eux pour aboutirà un ensemble non conflictuel, en désactivant un certain nombre de formules au profit des autres. Nous retiendrons le principe de cette méthode de fusion de normes avec préférences (correspondantà la recherche d'un sous-ensemble maximal consistant), la relation de préférence convenantà l'agent PAw restantà identifier de manière claire.…”