Norm Consistency in Practical Reasoning Agents
Martin J. Kollingbaum, Timothy J. Norman · 2003
Agents based on reactive planning architectures use pre-specified plans as behaviour specifications. Typically, such agents are motivated by desires to achieve or maintain states of the world. In this paper, however, we discuss a reactive planning agent architecture – NoA – where the behaviour of an agent is motivated and governed by norms. Normative agents are motivated by obligations to act, prohibitions for acting in particular ways, and permissions to act. Within this context, we focus on the problem of the adoption of a norm: under what circumstances is it appropriate for an agent to adopt a new obligation, prohibition or permission and what effect does it have on the agent’s normative state? In particular, the issue of the consistency of a new norm with the norms that currently influence the agent’s behaviour is considered. This provides useful input to the decision of whether to adopt a norm. This problem is investigated with respect to the NoA agent architecture, but, it is argued, provides useful insight to the problem of norm consistency in general and particularly where practical reasoning agents are concerned.