Galliwasp: A Goal-Directed Answer Set Solver

Abstract

Galliwasp is a goal-directed implementation of answer set programming. Unlike other answer set solvers, Galliwasp computes partial answer sets which are provably extensible to full answer sets. Galliwasp can execute arbitrary answer set programs in a top-down manner similar to SLD resolution. Galliwasp generates candidate answer sets by executing ordinary rules in a top-down, goal-directed manner using coinduction. Galliwasp next checks if the candidate answer sets are consistent with restrictions imposed by OLON rules. Those that are consistent are reported as solutions. Execution efficiency is significantly improved by performing the consistency check incrementally, i.e., as soon as an element of the candidate answer set is generated. We discuss the design of the Galliwasp system and its implementation. Galliwasp’s performance figures, which are comparable to other popular answer set solvers, are also presented.