Narrow proofs may be spacious: separating space and width in resolution
Proceedings of the thirty-eighth annual ACM symposium on Theory of computing
Clause learning can effectively P-simulate general propositional resolution
AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 1
Towards understanding and harnessing the potential of clause learning
Journal of Artificial Intelligence Research
Formalizing dangerous SAT encodings
SAT'07 Proceedings of the 10th international conference on Theory and applications of satisfiability testing
Solving satisfiability problems with preferences
Constraints
Complexity issues related to propagation completeness
Artificial Intelligence
| Hi-index | 0.00 |
Resolution is the most widely studied approach to propositionaltheorem proving. In developing efficient resolution-basedalgorithms, dozens of variants and refinements of resolutionhave been studied from both the empirical and analyticsides. The most prominent of these refinements are: DP(ordered), DLL (tree), semantic, negative, linear and regularresolution. In this paper, we characterize and studythese six refinements of resolution. We give a nearly completecharacterization of the relative complexities of all sixrefinements. While many of the important separations andsimulations were already known, many new ones are presentedin this paper; in particular, we give the first separationof semantic resolution from general resolution. Asa special case, we obtain the first exponential separationof negative resolution from general resolution. We also attemptto present a unifying framework for studying all ofthese refinements.