An Argumentation Inspired Heuristic for Resolving Normative Conflict

In multi-agent systems, norms provide a means for regulating agent behaviour at a system or society level rather than by constraining agent behaviour directly. In order to cope effectively with scenarios in which norms conflict, agents must be able to reason about norms and the consequences of compliance and violation. In particular, in this paper we argue that, if an agent must violate a norm (because norms conflict, for example) then it should determine which norm to violate in such a way that enables it to otherwise maximise its compliance with the remaining set of applicable norms. This concept of maximising compliance and minimising violation or conflict is similar to the notion of preferred extensions from argument theory, which provides a potentially valuable way to analyse sets of norms to determine which norms to violate. In this paper, therefore, we map normative structures to argument theory, and show how some resulting heuristics may be applied to minimising normative conflict.