Advances in Local Search for Satisfiability

Pham, Duc Nghia; Thornton, John; Gretton, Charles; Sattar, Abdul

doi:10.1007/978-3-540-76928-6_23

Cited by 13 publications

(22 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Again, referring to the SAT 2007 competition, the best individual local search algorithms (gNovelty + and AdaptG 2 WSAT) both employ the AdaptNovelty + self-tuning mechanismmaking this the state-of-the-art for online adaptation (at least within the SAT local search community). However, in relation to the current research, stagnation measures have not proved effective for tuning the clause weight decay parameter of any clause weighting algorithm except RSAPS (and RSAPS is known to be uncompetitive with gNovelty + or AdaptG 2 WSAT [8]). …”

Section: Parameter Tuning and Performance Predictionmentioning

confidence: 78%

“…Both AdaptNovelty + and AdaptG 2 WSAT are included for their class leading performance in the recent SAT competitions and RSAPS is included to compare iPAWS with another clause weighting adaptive algorithm. Finally, we included gNovelty + , the winner of the random satisfiable category of the 2007 SAT competition, and arguably the best general purpose SAT local search solver currently available [8].…”

Section: Tuning Paws Onlinementioning

confidence: 99%

“…Experimental observations have shown that gNovelty + 's performance can be significantly enhanced by treating this parameter as a binary switch that either allows weight to accumulate without reduction, or turns off clause weighting altogether [8]. To investigate the applicability of the iPAWS weight decay heuristic to other clause weighting algorithms, we decided to partially implement the iPAWS tuning process into gNovelty + .…”

Section: Tuning Paws Onlinementioning

confidence: 99%

“…While the best setting for this parameter varies widely from problem to problem, it can be effectively adapted during the search using a simple heuristic that measures the degree of search stagnation [7]. A similar heuristic was developed for the SAPS clause weighting algorithm [4] but this remains uncompetitive with the best WalkSAT techniques [8] Of the other current clause weighting algorithms, the pure additive weighting scheme (PAWS) is probably the best candidate for the development of a parameter adapting heuristic because its performance depends on a single M axInc parameter which controls the rate of decay of the clause weights [6]. However, despite considerable effort, no effective online method for adapting M axInc has been discovered.…”

Section: Introductionmentioning

confidence: 99%

“…However, despite considerable effort, no effective online method for adapting M axInc has been discovered. More recently, the gNovelty + algorithm has combined a WalkSAT-based heuristic with a clause weighting mechanism to produce the 2007 SAT competition random satisfiable category winner [8]. gNovelty + uses the WalkSAT heuristic to adapt noise and has a second parameter to control the rate of clause weight decay.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Using Cost Distributions to Guide Weight Decay in Local Search for SAT

Thornton

Pham

2008

PRICAI 2008: Trends in Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Abstract. Although clause weighting local search algorithms have produced some of the best results on a range of challenging satisfiability (SAT) benchmarks, this performance is dependent on the careful handtuning of sensitive parameters. When such hand-tuning is not possible, clause weighting algorithms are generally outperformed by self-tuning WalkSAT-based algorithms such as AdaptNovelty + and AdaptG 2 WSAT. In this paper we investigate tuning the weight decay parameter of two clause weighting algorithms using the statistical properties of cost distributions that are dynamically accumulated as the search progresses. This method selects a parameter setting both according to the speed of descent in the cost space and according to the shape of the accumulated cost distribution, where we take the shape to be a predictor of future performance. In a wide ranging empirical study we show that this automated approach to parameter tuning can outperform the default settings for two state-of-the-art algorithms that employ clause weighting (PAWS and gNovelty + ). We also show that these self-tuning algorithms are competitive with three of the best-known self-tuning SAT local search techniques: RSAPS, AdaptNovelty + and AdaptG 2 WSAT.

show abstract

Section: Parameter Tuning and Performance Predictionmentioning

confidence: 78%

Section: Tuning Paws Onlinementioning

confidence: 99%

Section: Tuning Paws Onlinementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Using Cost Distributions to Guide Weight Decay in Local Search for SAT

Thornton

Pham

2008

PRICAI 2008: Trends in Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

show abstract

Neighborhood Selection in Constraint-Based Local Search for Protein Structure Prediction

Shatabda

Newton

Sattar

2013

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Protein structure prediction (PSP) is a very challenging constraint optimization problem. Constraint-based local search approaches have obtained promising results in solving constraint models for PSP. However, the neighborhood exploration policies adopted in these approaches either remain exhaustive or are based on random decisions. In this paper, we propose heuristics to intelligently explore only the promising areas of the search neighborhood. On face centered cubic lattice using a realistic 20 × 20 energy model and standard benchmark proteins, we obtain structures with significantly lower energy and RMSD values than those obtained by the state-of-the-art algorithms.

show abstract

Switching among Non-Weighting, Clause Weighting, and Variable Weighting in Local Search for SAT

Wei

Zhang

2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. One way to design a local search algorithm that is effective on many types of instances is allowing this algorithm to switch among heuristics. In this paper, we refer to the way in which non-weighting algorithm adaptG 2 W SAT + selects a variable to flip, as heuristic adaptG 2 W SAT +, the way in which clause weighting algorithm RSAP S selects a variable to flip, as heuristic RSAP S, and the way in which variable weighting algorithm V W selects a variable to flip, as heuristic V W . We propose a new switching criterion: the evenness or unevenness of the distribution of clause weights. We apply this criterion, along with another switching criterion previously proposed, to heuristic adaptG 2 W SAT +, heuristic RSAP S, and heuristic V W . The resulting local search algorithm, which adaptively switches among these three heuristics in every search step according to these two criteria to intensify or diversify the search when necessary, is called N CV W (Non-, Clause, and Variable Weighting). Experimental results show that N CV W is generally effective on a wide range of instances while adaptG 2 W SAT +, RSAP S, V W , and gN ovelty+ and adaptG 2 W SAT 0, which won the gold and silver medals in the satisfiable random category in the SAT 2007 competition, respectively, are not.

show abstract

Advances in Local Search for Satisfiability

Cited by 13 publications

References 10 publications

Using Cost Distributions to Guide Weight Decay in Local Search for SAT

Using Cost Distributions to Guide Weight Decay in Local Search for SAT

Neighborhood Selection in Constraint-Based Local Search for Protein Structure Prediction

Switching among Non-Weighting, Clause Weighting, and Variable Weighting in Local Search for SAT

Contact Info

Product

Resources

About