Abstract | ||
---|---|---|
Pseudo-Boolean solvers hold out the theoretical potential of exponential improvements over conflict-driven clause learning (CDCL) SAT solvers, but in practice perform very poorly if the input is given in the standard conjunctive normal form (CNF) format. We present a technique to remedy this problem by recovering cardinality constraints from CNF on the fly during search. This is done by collecting potential building blocks of cardinality constraints during propagation and combining these blocks during conflict analysis. Our implementation has a non-negligible but manageable overhead when detection is not successful, and yields significant gains for some SAT competition and crafted benchmarks for which pseudo-Boolean reasoning is stronger than CDCL. It also boosts performance for some native pseudo-Boolean formulas where this approach helps to improve learned constraints. |
Year | Venue | DocType |
---|---|---|
2020 | THIRTY-FOURTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THE THIRTY-SECOND INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE AND THE TENTH AAAI SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE | Conference |
Volume | ISSN | Citations |
34 | 2159-5399 | 0 |
PageRank | References | Authors |
0.34 | 0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jan Elffers | 1 | 0 | 1.01 |
Jakob Nordström | 2 | 177 | 21.76 |