lp2normal — A Normalization Tool for Extended Logic Programs

Answer set programming (ASP) features a rich rule-based modeling language for encoding search problems. While normal rules form the simplest rule type in the language, various forms of extended rules have been introduced in order to ease modeling of complex conditions and constraints. Normalization means replacing such extended rules with identically functioning sets of normal rules. In this system description, we present lp2normal, which is a state-of-the-art normalizer that acts as a filter on ground logic programs produced by grounders, such as gringo. It provides options to translate away choice rules, cardinality rules, and weight rules, and to rewrite optimization statements using comparable techniques. The produced logic programs are suitable inputs to tools that lack support for extended rules, in particular. We give an overview of the normalization techniques currently supported by the tool and summarize its features. Moreover, we discuss the typical application scenarios of normalization, such as when implementing the search for answer sets using a back-end solver without direct support for cardinality constraints or pseudo-Boolean constraints.