Declarative Strategies for Agents with Incomplete Knowledge

Definite Agent Logic Programs (definite ALPs), recently introduced by Drescher, Schiffel and Thielscher, provide an interesting alternative to imperative languages like GOLOG for specifying agent behavior. The main advantages of ALPs is that they are fully declarative and independent of the underlying action theory. In this paper we extend the expressiveness of ALPs by introducing nonmonotonic negation in rule bodies. This allows us to handle incomplete knowledge, and to represent preferences among different agent behaviors. The semantics of ALPs with negation is based on a variant of answer set semantics.