Lower Bounds for Width-Restricted Clause Learning on Small Width Formulas

@article{BenSasson2010LowerBF,
  title={Lower Bounds for Width-Restricted Clause Learning on Small Width Formulas},
  author={Eli Ben-Sasson and Jan Johannsen},
  journal={Electron. Colloquium Comput. Complex.},
  year={2010},
  volume={17},
  pages={85}
}
It has been observed empirically that clause learning does not significantly improve the performance of a SAT solver when restricted to learning clauses of small width only. This experience is supported by lower bound theorems. It is shown that lower bounds on the runtime of width-restricted clause learning follow from resolution width lower bounds. This yields the first lower bounds on width-restricted clause learning for formulas in 3-CNF. 
Clause-Learning Algorithms with Many Restarts and Bounded-Width Resolution
TLDR
It is proved that, although the solver is not explicitly designed for it, with high probability it ends up behaving as width-k resolution after no more than O(n2k+2) conflicts and restarts, where n is the number of variables.
Width-Restricted Clause Learning and k-DNF Resolution
The hierarchies of proof systems RTL(k) of Resolution trees with lemmas of width k and Res∗(k) of tree-like k-DNF Resolution lie between tree-like and dag-like Resolution in strength. We show that
Proof Complexity and SAT Solving
This chapter gives an overview of proof complexity and connections to SAT solving, focusing on proof systems such as resolution, Nullstellensatz, polynomial calculus, and cutting planes
Exponential Separations in a Hierarchy of Clause Learning Proof Systems
TLDR
A resolution-based propositional proof system that is related to the DPLL algorithm with clause learning and its fragments are related to clause learning algorithms where the width of learned clauses is bounded by k.
Improving the efficiency of learning CSP solvers
TLDR
The use of forgetting techniques from the literature is shown to significantly improve the performance of modern learning CSP solvers, contradicting some previous research.
Space Complexity in Polynomial Calculus ( Extended Abstract )
TLDR
It is proved that any k-CNF formula can be refuted in PC in simultaneous exponential size and linear space (which holds for resolution and thus for PCR, but was not known to be the case for PC).
Space Complexity in Polynomial Calculus
TLDR
It is proved that any k-CNF formula can be refuted in PC in simultaneous exponential size and linear space (which holds for resolution and thus for PCR, but was not known to be the case for PC).
On CDCL-Based Proof Systems with the Ordered Decision Strategy
We prove that CDCL SAT-solvers with the ordered decision strategy and the DECISION learning scheme are equivalent to ordered resolution. We also prove that, by replacing this learning scheme with its
An empirical study of learning and forgetting constraints
TLDR
A major empirical investigation into the overheads introduced by unbounded constraint learning in CSP is conducted and it is confirmed that forgetting can significantly improve the performance of modern learning CSP solvers, contradicting some previous research.
On the virtue of succinct proofs: amplifying communication complexity hardness to time-space trade-offs in proof complexity
TLDR
A family of modified pebbling formulas such as F_n, a new two-player communication complexity lower bound for composed search problems in terms of block sensitivity, are exhibited, a contribution that is believed to be of independent interest.
...
...

References

SHOWING 1-10 OF 28 REFERENCES
An Exponential Lower Bound for Width-Restricted Clause Learning
TLDR
It is shown by proving lower bounds on the proof length in a certain resolution proof system related to clause learning that the lower bound is of the same order of magnitude as a known lower bound for backtracking algorithms without any clause learning.
Clause-Learning Algorithms with Many Restarts and Bounded-Width Resolution
TLDR
It is proved that, although the solver is not explicitly designed for it, with high probability it ends up behaving as width-k resolution after no more than O(n2k+2) conflicts and restarts, where n is the number of variables.
Optimality of size-width tradeoffs for resolution
TLDR
It is shown that the recently proposed algorithm of Ben-Sasson & Wigderson for searching for proofs in resolution cannot give better than weakly exponential performance, and the optimality of the general relationship called size-width tradeoffs in Ben-sasson& WigDerson is shown.
A combinatorial characterization of resolution width
  • Albert Atserias, V. Dalmau
  • Mathematics, Computer Science
    18th IEEE Annual Conference on Computational Complexity, 2003. Proceedings.
  • 2003
TLDR
This work provides a characterization of the resolution width introduced in the context of propositional proof complexity in terms of the existential pebble game and a surprising proof that the minimum space of refuting a 3-CNF formula is always bounded from below by the minimum width of refuted it.
Short proofs are narrow—resolution made simple
We develop a general strategy for proving width lower bounds, which follows Haken's original proof technique but is now simple and clear. It reveals that large width is implied by certain natural
Clause Learning Can Effectively P-Simulate General Propositional Resolution
TLDR
This paper utilizes the idea of effective p-simulation, which is a new way of comparing clause learning with general resolution and other proof systems, and shows that pool proofs, a previously used characterization of clause learning, can effectively p-Simulate general resolution.
Efficient conflict driven learning in a Boolean satisfiability solver
TLDR
This paper generalizes various conflict driven learning strategies in terms of different partitioning schemes of the implication graph to re-examine the learning techniques used in various SAT solvers and propose an array of new learning schemes.
Switching lemma for small restrictions and lower bounds for k-DNF resolution
We prove a new switching lemma that works for restrictions that set only a small fraction of the variables and is applicable to DNFs with small conjunctions. We use this to prove lower bounds for the
On the Power of Clause-Learning SAT Solvers with Restarts
TLDR
It is proved that modern SAT solvers that learn asserting clauses indeed p-simulate general resolution without the need for any additional techniques.
Short resolution proofs for a sequence of tricky formulas
TLDR
This note proves, opposed to that conjecture, that there are proofs of polynomial lengths (or even linear in the lengths of the formulas) of those formulas.
...
...