Results 1  10
of
20
Simplified and Improved Resolution Lower Bounds
 IN PROCEEDINGS OF THE 37TH IEEE FOCS
, 1996
"... We give simple new lower bounds on the lengths of Resolution proofs for the pigeonhole principle and for randomly generated formulas. For random formulas, our bounds significantly extend the range of formula sizes for which nontrivial lower bounds are known. For example, we show that with probabili ..."
Abstract

Cited by 102 (7 self)
 Add to MetaCart
We give simple new lower bounds on the lengths of Resolution proofs for the pigeonhole principle and for randomly generated formulas. For random formulas, our bounds significantly extend the range of formula sizes for which nontrivial lower bounds are known. For example, we show that with probability approaching 1, any Resolution refutation of a randomly chosen 3CNF formula with at most n 6=5\Gammaffl clauses requires exponential size. Previous bounds applied only when the number of clauses was at most linear in the number of variables. For the pigeonhole principle our bound is a small improvement over previous bounds. Our proofs are more elementary than previous arguments, and establish a connection between Resolution proof size and maximum clause size.
The efficiency of resolution and DavisPutnam procedures
 SIAM Journal on Computing
, 1999
"... We consider several problems related to the use of resolutionbased methods for determining whether a given boolean formula in conjunctive normal form is satisfiable. First, building on work of Clegg, Edmonds and Impagliazzo, we give an algorithm for satisfiability that when given an unsatisfiabl ..."
Abstract

Cited by 55 (1 self)
 Add to MetaCart
We consider several problems related to the use of resolutionbased methods for determining whether a given boolean formula in conjunctive normal form is satisfiable. First, building on work of Clegg, Edmonds and Impagliazzo, we give an algorithm for satisfiability that when given an unsatisfiable formula of F finds a resolution proof of F , and the runtime of our algorithm is nontrivial as a function of the size of the shortest resolution proof of F . Next we investigate a class of backtrack search algorithms, commonly known as DavisPutnam procedures and provide the first averagecase complexity analysis for their behavior on random formulas. In particular, for a simple algorithm in this class, called ordered DLL we prove that the running time of the algorithm on a randomly generated kCNF formula with n variables and m clauses is 2 Q(n(n/m) 1/(k2) ) with probability 1  o(1). Finally, we give new lower bounds on res(F), the size of the smallest resolution refutation ...
On the Complexity of Unsatisfiability Proofs for Random kCNF Formulas
 In 30th Annual ACM Symposium on the Theory of Computing
, 1997
"... We study the complexity of proving unsatisfiability for random kCNF formulas with clause density D = m=n where m is number of clauses and n is the number of variables. We prove the first nontrivial general upper bound, giving algorithms that, in particular, for k = 3 produce refutations almost cer ..."
Abstract

Cited by 51 (1 self)
 Add to MetaCart
We study the complexity of proving unsatisfiability for random kCNF formulas with clause density D = m=n where m is number of clauses and n is the number of variables. We prove the first nontrivial general upper bound, giving algorithms that, in particular, for k = 3 produce refutations almost certainly in time 2 O(n=D) . This is polynomial when m n 2 =logn. We show that our upper bounds are tight for certain natural classes of DavisPutnam algorithms. We show further that random 3CNF formulas of clause density D almost certainly have no resolution refutation of size smaller than 2 W(n=D 4+e ) , which implies the same lower bound on any DavisPutnam algorithm. We also give a much simpler argument based on a novel form of selfreduction that yields a slightly weaker 2 W(n=D 5+e ) lower bound. 1 Introduction The random kCNF model has been widely studied for several good reasons. First, it is an intrinsically natural model, analogous to the random graph model, that shed...
Stochastic Boolean Satisfiability
 Journal of Automated Reasoning
, 2000
"... . Satisfiability problems and probabilistic models are core topics of artificial intelligence and computer science. This paper looks at the rich intersection between these two areas, opening the door for the use of satisfiability approaches in probabilistic domains. The paper examines a generic stoc ..."
Abstract

Cited by 49 (2 self)
 Add to MetaCart
. Satisfiability problems and probabilistic models are core topics of artificial intelligence and computer science. This paper looks at the rich intersection between these two areas, opening the door for the use of satisfiability approaches in probabilistic domains. The paper examines a generic stochastic satisfiability problem, SSat, which can function for probabilistic domains as Sat does for deterministic domains. It shows the connection between SSat and well studied problems in belief network inference and planning under uncertainty, and defines algorithms, both systematic and stochastic, for solving SSat instances. These algorithms are validated on random SSat formulae generated under the fixedclause model. In spite of the large complexity gap between SSat (PSPACE) and Sat (NP), the paper suggests that much of what we've learned about Sat transfers to the probabilistic domain. 1. Introduction There has been a recent focus in artificial intelligence (AI) on solving problems exh...
Satisfiability Solvers
, 2008
"... The past few years have seen an enormous progress in the performance of Boolean satisfiability (SAT) solvers. Despite the worstcase exponential run time of all known algorithms, satisfiability solvers are increasingly leaving their mark as a generalpurpose tool in areas as diverse as software and h ..."
Abstract

Cited by 24 (0 self)
 Add to MetaCart
The past few years have seen an enormous progress in the performance of Boolean satisfiability (SAT) solvers. Despite the worstcase exponential run time of all known algorithms, satisfiability solvers are increasingly leaving their mark as a generalpurpose tool in areas as diverse as software and hardware verification [29–31, 228], automatic test pattern generation [138, 221], planning [129, 197], scheduling [103], and even challenging problems from algebra [238]. Annual SAT competitions have led to the development of dozens of clever implementations of such solvers [e.g. 13,
The Resolution Complexity of Random Graph kColorability
 In preparation
, 2003
"... We consider the resolution proof complexity of propositional formulas which encode random instances of graph kcolorability. We obtain a tradeoff between the graph density and the resolution proof complexity. For random graphs with linearly many edges we obtain linearexponential lower bounds on ..."
Abstract

Cited by 16 (6 self)
 Add to MetaCart
We consider the resolution proof complexity of propositional formulas which encode random instances of graph kcolorability. We obtain a tradeoff between the graph density and the resolution proof complexity. For random graphs with linearly many edges we obtain linearexponential lower bounds on the length of resolution refutations. For any # > 0, we obtain subexponential lower bounds of the form for some # > 0 for nonkcolorability proofs of graphs with n vertices and O(n # ) edges. We obtain sharper lower bounds for DavisPutnamDPLL proofs and for proofs in a system considered by McDiarmid.
Homogenization and the Polynomial Calculus
 Computational Complexity
, 1999
"... In standard implementations of the Grobner basis algorithm, the original polynomials are homogenized so that each term in a given polynomial has the same degree. In this paper, we study the effect of homogenization on the runtime of the Grobner basis algorithm for instances of satisfiability, and ..."
Abstract

Cited by 10 (4 self)
 Add to MetaCart
In standard implementations of the Grobner basis algorithm, the original polynomials are homogenized so that each term in a given polynomial has the same degree. In this paper, we study the effect of homogenization on the runtime of the Grobner basis algorithm for instances of satisfiability, and also on the degree of Polynomial Calculus (PC) and Nullstellensatz refutations of unsatisfiable formulas. We show that the Nullstellensatz degree is equivalent to the homogenized PC degree. Using this relationship, we prove nearly linear separations between Nullstellensatz and PC degree, for 3CNF formulas. Research partially supported by NSF grant CCR9457782 and a scholarship from the Arizona Chapter of the ARCS Foundation. + Research supported by NSF CCR9734911, Sloan Research Fellowship BR3311, and by a cooperative research grant INT9600919 /ME103 from NSF and the M SMT (Czech Republic), and USAIsraelBSF Grant 9700188 # Research supported by NSF grant CCR9457782 and USI...
The Complexity of Properly Learning Simple Concept Classes
, 2007
"... We consider the complexity of properly learning concept classes, i.e. when the learner must output a hypothesis of the same form as the unknown concept. We present the following new upper and lower bounds on wellknown concept classes: • We show that unless NP = RP, there is no polynomialtime PAC l ..."
Abstract

Cited by 9 (1 self)
 Add to MetaCart
We consider the complexity of properly learning concept classes, i.e. when the learner must output a hypothesis of the same form as the unknown concept. We present the following new upper and lower bounds on wellknown concept classes: • We show that unless NP = RP, there is no polynomialtime PAC learning algorithm for DNF formulas where the hypothesis is an ORofthresholds. Note that as special cases, we show that neither DNF nor ORofthresholds are properly learnable unless NP = RP. Previous hardness results have required strong restrictions on the size of the output DNF formula. We also prove that it is NPhard to learn the intersection of ℓ ≥ 2 halfspaces by the intersection of k halfspaces for any constant k ≥ 0. Previous work held for the case when k = ℓ. • Assuming that NP � ⊆ DTIME(2nɛ) for a certain constant ɛ < 1 we show that it is not possible to learn size s decision trees by size sk decision trees for any k ≥ 0. Previous hardness results for learning decision trees held for k ≤ 2. • We present the first nontrivial upper bounds on properly learning DNF formulas. More specifically, we show how to learn size s DNF by DNF in time 2 Õ( √ n log s). The hardness results for DNF formulas and intersections of halfspaces are obtained via specialized
Exponential lower bounds and Integrality Gaps for Treelike LovászSchrijver Procedures
, 2007
"... The matrix cuts of Lovász and Schrijver are methods for tightening linear relaxations of zeroone programs by the addition of new linear inequalities. We address the question of how many new inequalities are necessary to approximate certain combinatorial problems with strong guarantees, and to solve ..."
Abstract

Cited by 7 (1 self)
 Add to MetaCart
The matrix cuts of Lovász and Schrijver are methods for tightening linear relaxations of zeroone programs by the addition of new linear inequalities. We address the question of how many new inequalities are necessary to approximate certain combinatorial problems with strong guarantees, and to solve certain instances of Boolean satisfiability. We show that relaxations of linear programs, obtained by tightening via any subexponentialsize semidefinite LovászSchrijver derivation tree, cannot approximate maxkSAT to a factor better than 1+ 1 2k−1, maxkXOR to a factor better than 2 − ε, nor vertex cover to a factor better than 7/6. We prove exponential size lower bounds for treelike LovászSchrijver proofs of unsatisfiability for several prominent unsatisfiable CNFs, including random 3CNF formulas, random systems of linear equations, and the Tseitin graph formulas. Furthermore, we prove that treelike LS+ cannot polynomially simulate treelike cutting planes, and that treelike LS+ cannot polynomially simulate unrestricted resolution. All of our size lower bounds for derivation trees are based upon connections between the size and height of the derivation tree (its rank). The primary method is a treesize/rank tradeoff for LovászSchrijver refutations: Small tree size implies small rank. Surprisingly, this does not hold for derivations of arbitrary linear inequalities. We show that for LS0 and LS, there are examples with polynomialsize treelike derivations, but requiring linear rank.