An application of answer set programming: Superoptimisation (A preliminary report)
T Crick, Martin Brain, Marina De Vos, John Fitch, Jürgen Dix, Aaron Hunter · 2006
Answer set programming (ASP) is a declarative problem-solving technique that uses the computation of answer set semantics to provide solutions. Despite comprehensive implementations and a strong theoretical basis, ASP has yet to be used for more than a handful of large-scale applications. This paper describes such a large-scale application and presents some preliminary results. The TOAST (Total Optimisation using Answer Set Technology) project seeks to generate optimal machine code for simple, acyclic functions using a technique known as superoptimisation. ASP is used as a scalable computational engine for conducting searches over complex, non-regular domains. The experimental results suggest this is a viable approach to the optimisation problem and demonstrates the value of using parallel answer set solvers. This paper was presented at and accepted for publication in the informal proceedings of the 11th International Workshop on Non-Monotonic Reasoning (NMR’06), Lake District, UK. Brain, Crick, De Vos and Fitch / An Application of Answer Set Programming: Superoptimisation