Optimum stable model search: algorithms and implementation

Abstract

Abstract Answer Set Programming (ASP) is a well-known declarative problem solving paradigm developed in the field of nonmonotonic reasoning and logic programming. The usual target of ASP is the solution of combinatorial search problems, nonetheless the language of ASP was extended with weak constraints for concise modelling of optimization problems. In the case of ASP programs with weak constraints, the main computational task of an ASP solver is optimum stable model search . In this article, we present and compare several algorithms for optimum stable model search. We consider solutions traditionally adopted by ASP solvers, and we introduce new solving strategies obtained by porting to the ASP setting some algorithms that were introduced for Maximum Satisfiability solving. The article also reports on the implementation of these algorithms in the ASP solver wasp . An empirical analysis highlights pros and cons of different strategies for computing optimum stable models.