an:06916320
Zbl 06916320
Kochemazov, Stepan; Zaikin, Oleg
ALIAS: a modular tool for finding backdoors for SAT
EN
Beyersdorff, Olaf (ed.) et al., Theory and applications of satisfiability testing -- SAT 2018. 21st international conference, SAT 2018, held as part of the Federated Logic Conference, FloC 2018, Oxford, UK, July 9--12, 2018. Proceedings. Cham: Springer (ISBN 978-3-319-94143-1/pbk; 978-3-319-94144-8/ebook). Lecture Notes in Computer Science 10929, 419-427 (2018).
2018
a
68Q25 68T20
Summary: We present \textsc{ALIAS}, a modular tool aimed at finding backdoors for hard SAT instances. Here by a backdoor for a specific SAT solver and SAT formula we mean a set of its variables, all possible instantiations of which lead to construction of a family of subformulas with the total solving time less than that for an original formula. For a particular backdoor, the tool uses the Monte-Carlo algorithm to estimate the runtime of a solver when partitioning an original problem via said backdoor. Thus, the problem of finding a backdoor is viewed as a black-box optimization problem. The tool's modular structure allows to employ state-of-the-art SAT solvers and black-box optimization heuristics. In practice, for a number of hard SAT instances, the tool made it possible to solve them much faster than using state-of-the-art multithreaded SAT-solvers.
For the entire collection see [Zbl 1390.68015].