New Inference Rules for Max-SAT

Thumbnail
View/Open
Issue date
2007
Author
Li, Chu-Min
Manyà Serres, Felip
Planes Cid, Jordi
Suggested citation
Li, Chu-Min; Manyà Serres, Felip; Planes Cid, Jordi; . (2007) . New Inference Rules for Max-SAT. Journal of Artificial Intelligence Research, 2007, vol. 30, p. 321-359. https://doi.org/10.1613/jair.2215.
Impact

Google Scholar logo  Google Scholar
Share
Export to Mendeley
Metadata
Show full item record
Abstract
Exact Max-SAT solvers, compared with SAT solvers, apply little inference at each node of the proof tree. Commonly used SAT inference rules like unit propagation produce a simplified formula that preserves satisfiability but, unfortunately, solving the Max-SAT problem for the simplified formula is
not equivalent to solving it for the original formula. In this paper, we define a number of original inference rules that, besides being applied efficiently, transform Max-SAT instances into equivalent Max-SAT instances which are easier to solve. The soundness of the rules, that can be seen as refinements of unit resolution adapted to Max-SAT, are proved in a novel and simple way via an integer programming transformation. With the aim of finding out how powerful the inference rules are in practice, we have developed a new Max-SAT solver, called MaxSatz, which incorporates those rules, and performed an experimental investigation. The results provide empirical evidence that MaxSatz is very competitive, at least, on random Max-2SAT, random Max-3SAT, MaxCut, and Graph 3-coloring instances, as well as on the benchmarks from the Max-SAT Evaluation 2006.
URI
http://hdl.handle.net/10459.1/57413
DOI
https://doi.org/10.1613/jair.2215
Is part of
Journal of Artificial Intelligence Research, 2007, vol. 30, p. 321-359
Collections