Abstract
A social norm describes a standard of behavior expected to
be followed by agents in a given society, and failure to comply results in sanctioning and loss of utility for the violating agent. Agents need to take existing norms into account when generating plans to achieve their goals, to either forestall potential violations if an agent wants to be fully norm compliant, or understanding the effects on its utility if violations are acceptable. In this paper we model a normative system in terms of classical planning, and develop two norm formalizations: the first concerns actions in a given context; while the second concerns sequences of states and is based on Linear Temporal Logic. We use these norm formalizations to develop different planning approaches that take into account such norms, and empirically evaluate the algorithm’s performance.
be followed by agents in a given society, and failure to comply results in sanctioning and loss of utility for the violating agent. Agents need to take existing norms into account when generating plans to achieve their goals, to either forestall potential violations if an agent wants to be fully norm compliant, or understanding the effects on its utility if violations are acceptable. In this paper we model a normative system in terms of classical planning, and develop two norm formalizations: the first concerns actions in a given context; while the second concerns sequences of states and is based on Linear Temporal Logic. We use these norm formalizations to develop different planning approaches that take into account such norms, and empirically evaluate the algorithm’s performance.
Original language | English |
---|---|
Title of host publication | 2017 International Workshop on Coordination, Organisations, Institutions and Norms |
Number of pages | 13 |
Publication status | Published - 2017 |