Diversification by Clauses Deletion Strategies in Portfolio Parallel SAT Solving

Guo, Long; Jabbour, Saïd; Lonlac, Jerry; Saïs, Lakhdar

doi:10.1109/ictai.2014.110

Cited by 6 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…At a given search state, using a relevant selection function based on progress saving (PSM), it activates the most promising learned clauses while freezing irrelevant ones. In [11], a new criterion to quantify the relevance of a clause using its backtrack level called BTL for BackTrack Level was proposed. From experiments, the authors observed that the learned clauses with small BTL values are used more often in the unit propagation process than those with higher BTL values.…”

Section: On the Learned Clauses Database Management Strategiesmentioning

confidence: 99%

See 1 more Smart Citation

Top-k Learned Clauses for Modern SAT Solvers

Lonlac¹,

Nguifo²

2023

Int. J. Artif. Intell. Tools

View full text Add to dashboard Cite

Clause Learning is one of the most important components of a conflict driven clause learning (CDCL) SAT solver that is effective on industrial SAT instances. Since the number of learned clauses is proved to be exponential in the worst case, it is necessary to identify the most relevant clauses to maintain and delete the irrelevant ones. As reported in the literature, several learned clauses deletion strategies have been proposed. However the diversity in both the number of clauses to be removed at each step of reduction and the results obtained with each strategy increase the difficulty to determine which criterion is better. Thus, the problem to select which learned clauses are to be removed during the search step remains very challenging. In this paper, we propose a novel approach to identify the most relevant learned clauses without favoring or excluding any of the proposed measures, but by adopting the notion of dominance relationship among those measures. Our approach bypasses the problem of difference in the results obtained with different measures and reaches a compromise between the measures assessments. Furthermore, the proposed approach also reduces the non-trivial problem of choosing the number of clauses to delete at each reduction of the learned clause database.

show abstract

Section: On the Learned Clauses Database Management Strategiesmentioning

confidence: 99%

“…Managing the learned clauses database was the subject of several studies [1,8,2,9,10,11]. These strategies were proposed with the objective to maintain a learned clause database of reasonable size by eliminating clauses deemed irrelevant to the subsequent search.…”

mentioning

confidence: 99%

Top-k Learned Clauses for Modern SAT Solvers

Lonlac¹,

Nguifo²

2023

Int. J. Artif. Intell. Tools

View full text Add to dashboard Cite

show abstract

“…At a given search state, using a relevant selection function based on progress saving (PSM), it activates the most promising learned clauses while freezing irrelevant ones. In [14], a new criterion to quantify the relevance of a clause using its backtrack level called BTL for BackTrack Level was proposed. From experiments, the authors observed that the learned clauses with small BTL values are used more often in the unit propagation process than those with higher BTL values.…”

Section: On the Learned Clauses Database Management Strategiesmentioning

confidence: 99%

“…Managing the learned clauses database was the subject of several studies [17,19,11,2,3,14]. These strategies were proposed with the objective to maintain a learned clause database of reasonable size by eliminating clauses deemed irrelevant to the subsequent search.…”

Section: Introductionmentioning

confidence: 99%

Towards Learned Clauses Database Reduction Strategies Based on Dominance Relationship

Lonlac,

Nguifo

2017

Preprint

Self Cite

View full text Add to dashboard Cite

Clause Learning is one of the most important components of a conflict driven clause learning (CDCL) SAT solver that is effective on industrial instances. Since the number of learned clauses is proved to be exponential in the worse case, it is necessary to identify the most relevant clauses to maintain and delete the irrelevant ones. As reported in the literature, several learned clauses deletion strategies have been proposed. However the diversity in both the number of clauses to be removed at each step of reduction and the results obtained with each strategy creates confusion to determine which criterion is better. Thus, the problem to select which learned clauses are to be removed during the search step remains very challenging. In this paper, we propose a novel approach to identify the most relevant learned clauses without favoring or excluding any of the proposed measures, but by adopting the notion of dominance relationship among those measures. Our approach bypasses the problem of the diversity of results and reaches a compromise between the assessments of these measures. Furthermore, the proposed approach also avoids another non-trivial problem which is the amount of clauses to be deleted at each reduction of the learned clause database.

show abstract

“…On the other hand, keeping a lot of learned clauses slows down the propagation process, and may even cause the solver to run out of memory. Therefore, solvers clean their learned clause database regularly and delete some of the learned clauses [19,111,126,169].…”

Section: Clause Deletion Strategiesmentioning

confidence: 99%

SAT and CP - Parallelisation and Applications

Ehlers¹

View full text Add to dashboard Cite

In this thesis, we consider the parallelisation and application of SAT and CP solvers. In the first chapter, we consider SAT, the decision problem of propositional logic. We discuss details of the implementations of SAT solvers, and show how to improve upon existing sequential solvers. Furthermore, we discuss the parallelisation of these solvers with a focus on the communication of intermediate results within a parallel solver. The second chapter is concerned with Contraint Programing (CP) with learning. Contrary to classical Constraint Programming techniques, this incorporates learning mechanisms as they are used in the field of SAT solving. We present results from parallelising CHUFFED, a learning CP solver. In the final chapter, we discuss Sorting Networks, which are data oblivious sorting algorithms. Their independence of the input data lends them to parallel implementation. We consider the question how many parallel sorting steps are needed to sort some inputs, and present both lower and upper bounds for several cases.

show abstract

Diversification by Clauses Deletion Strategies in Portfolio Parallel SAT Solving

Cited by 6 publications

References 11 publications

Top-k Learned Clauses for Modern SAT Solvers

Top-k Learned Clauses for Modern SAT Solvers

Towards Learned Clauses Database Reduction Strategies Based on Dominance Relationship

SAT and CP - Parallelisation and Applications

Contact Info

Product

Resources

About