asprin: Customizing Answer Set Preferences without a Headache

International audienceIn this paper we describe asprin, a general, flexible, and extensibleframework for handling preferences among the stablemodels of a logic program. We show how complex preferencerelations can be specified through user-defined preferencetypes and their arguments. We describe how preferencespecifications are handled internally by so-called preferenceprograms, which are used for dominance testing. Wealso give algorithms for computing one, or all, optimal stablemodels of a logic program. Notably, our algorithms dependon the complexity of the dominance tests and make use ofmulti-shot answer set solving technology