Improving probability selection based weights for satisfiability problems

Huimin Fu, Jun Liu, Guanfeng Wu, Yang Xu, Geoff Sutcliffe

Research output: Contribution to journalArticlepeer-review

Abstract

Boolean Satisfiability problem (SAT) plays a prominent role in many domains of computer science and artificial intelligence due to its significant importance in both theory and applications. Algorithms for solving SAT problems can be categorized into two main classes: complete algorithms and incomplete algorithms (typically stochastic local search (SLS) algorithms). SLS algorithms are among the most effective for solving uniform random SAT problems, while hybrid algorithms achieved great breakthroughs for solving hard random SAT (HRS) problem recently. However, there is a lack of algorithms that can effectively solve both uniform random SAT and HRS problems. In this paper, a new SLS algorithm named SelectNTS is proposed aiming at solving both uniform random SAT and HRS problem effectively. SelectNTS is essentially an improved probability selection based local search algorithm, the core of which includes new clause and variable selection heuristics: a new clause weighting scheme and a biased random walk strategy are utilized to select a clause, while a new probability selection strategy with the variation of configuration checking strategy is used to select a variable. Extensive experimental results show that SelectNTS outperforms the state-of-the-art random SAT algorithms and hybrid algorithms in solving both uniform random SAT and HRS problems effectively.

Original languageEnglish (US)
Article number108572
JournalKnowledge-Based Systems
Volume245
DOIs
StatePublished - Jun 7 2022
Externally publishedYes

Keywords

  • Boolean Satisfiability problem
  • Stochastic local search
  • Weights

ASJC Scopus subject areas

  • Management Information Systems
  • Software
  • Information Systems and Management
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Improving probability selection based weights for satisfiability problems'. Together they form a unique fingerprint.

Cite this