The Multi-engine ASP Solver ME-ASP: Progress Report
This work addresses efficiency improvements in ASP solving for researchers and practitioners, but it is incremental as it builds on existing multi-engine and algorithm selection techniques.
The paper tackles the problem of solving ground and non-ground Answer Set Programming (ASP) programs by introducing an optimized multi-engine solver (ME-ASP) that uses algorithm selection to choose among heterogeneous ASP solvers, resulting in a substantially faster implementation for ground programs and some benefits for non-ground programs.
MEASP is a multi-engine solver for ground ASP programs. It exploits algorithm selection techniques based on classification to select one among a set of out-of-the-box heterogeneous ASP solvers used as black-box engines. In this paper we report on (i) a new optimized implementation of MEASP; and (ii) an attempt of applying algorithm selection to non-ground programs. An experimental analysis reported in the paper shows that (i) the new implementation of \measp is substantially faster than the previous version; and (ii) the multi-engine recipe can be applied to the evaluation of non-ground programs with some benefits.