A methodology for plan revision under norm and outcome compliance

Abstract

Scholars understand an agent as a system acting in an environment. Such an environment is usually governed by norms, and the agent has to obey to such norms when pursuing her objectives. We report a non-monotonic modal logic able to describe the environment, the norms, and the agent's capabilities as well as her mental attitudes (e.g., desires, intentions). First, we show how such a logic is expressive enough to determine when the agent is compliant with respect to norms and objectives by extending it with a formal characterisation of the concepts of norm and outcome compliance. Then, in the case the agent violates some norms or does not achieve all her objectives, we propose a preliminary analysis of methodologies to revise the theory and restore compliance. © 2013 Springer-Verlag.