Normative conflict resolution in multi-agent systems

Abstract

Norms (permissions, obligations and prohibitions) offer a useful and powerful abstraction with which to capture social constraints in multi-agent systems. Norms should exclude disruptive or antisocial behaviour without prescribing the design of individual agents or restricting their autonomy. An important challenge, however, in the design and management of systems governed by norms is that norms may, at times, conflict with one another; e.g, an action may be simultaneously prohibited and obliged for a particular agent. In such circumstances, agents no longer have the option of complying with these norms; whatever they do or refrain from doing will lead to a social constraint being broken. In this paper, we present mechanisms for the detection and resolution of normative conflicts. These mechanisms, based on first-order unification and constraint solving techniques, are the building blocks of more sophisticated algorithms we present for the management of normative positions, that is, the adoption and removal of permissions, obligations and prohibitions in societies of agents. We capture both direct and indirect conflicts between norms, formalise a practical concept of authority, and model conflicts that may arise as a result of delegation. We are able to formally define classic ways for resolving conflicts such as lex superior and lex posterior.

Modeling conflicts between legal rules

“…An interesting model of lex superior... is presented in [26], where the authors treat the level of authority which establishes a norm as a root of preference between norms. Unfortunately, such a conception does not fit the Polish legal system in which the hierarchy of legal acts is not strictly based on authority hierarchy, but is established by legal theory.…”

A coherence maximisation process for solving normative inconsistencies

“…These experiments compare our proposal against existing proposals and demonstrate that it allows agents to solve normative inconsistencies by adapting the solution in response to a dynamic, uncertain and non-deterministic environment. In particular, we have compared three inconsistency resolution strategies: (i) maximising coherence, which is the method proposed in this paper; (ii) following the static order according to salience, which is the method used in the majority of previous proposals [4,7,21] (see Section 7 for a discussion of previous work); and (iii) following a conditional order that determines which instance prevails according to a condition [41].…”

“…In a more recent proposal, Vasconcelos et al [41] propose to avoid inconsistencies among instances by curtailing the scope of norms that may cause inconsistencies when they are instantiated. To determine which instance prevails in case of inconsistency, the authors use orders, or policies, that determine, given a pair of instances, which one is to be curtailed.…”

“…Existing work concerning normative agents typically proposes the use of a static order based on norm salience (i.e., importance of norms) to determine which instance prevails in the case of inconsistency (i.e., the instance created out of the most salient norm prevails) [7,12,4,41]. However, a significant drawback of such work is the fact that the order is specified off-line and is hard-wired into agents.…”

OWL-POLAR: Semantic Policies for Agent Reasoning

“…If we can determine possible logical conflicts while designing policies, we can create better policies that are less likely to raise conflicts at run time. Furthermore, we can use various conflict resolution strategies such as setting a priority ordering between the policies to solve conflicts [11,21,22], once we determine that two policies may conflict.…”