Utilizing permission norms in BDI practical normative reasoning

Abstract

Norms have been used in multi-agent systems as a standard description of agents' behaviors. A lot of effort has been put into for­malizing norms and utilizing them in agent decision making. Such work focuses mostly on two types of norms: prohibitions and obligations; with the unstated assumption that agents are completely aware of all norms. However, agents may have incomplete knowledge about norms in a system for several reasons such as deficient norm identification or because norms are not fixed. In this work we argue that, by assuming that agents do not have complete knowledge of the norms within a system, permission norms are fundamental for modeling unknown normative states. Using Event Calculus (EC), we propose a formal representation of permission norms and we show how to use it in agent normative practical reasoning. We implement a simple mineral mining scenario to demonstrate our work.