A Path in the Jungle of Logics for Multi-Agent Systems: on the Relation between General Game-Playing Logics and Seeing-To-It-That Logics

Abstract

In the recent years, several concurrent logical systems for reasoning about agency and social interaction and for representing game properties have been proposed. The aim of the present paper is to put some order in this 'jungle' of logics by studying the relationship between the dynamic logic of agency DLA and the game description language GDL. The former has been proposed as a variant of the logic of agency STIT by Belnap et al. in which agents' action are named, while the latter has been introduced in AI as a formal language for reasoning about general game-playing. The paper provides complexity results for the satisfiability problems of both DLALogic and GDL as well as a polynomial embedding of GDL into DLA.