DocumentCode :
588936
Title :
A Heuristic Restart Strategy to Speed Up the Solving of Satisfiability Problem
Author :
Ying Guo ; Bin Zhang ; Changsheng Zhang
Author_Institution :
Coll. of Inf. Sci. & Eng., Northeastern Univ., Shenyang, China
Volume :
2
fYear :
2012
fDate :
28-29 Oct. 2012
Firstpage :
423
Lastpage :
426
Abstract :
Restart strategies are widely used in today´s conflict-driven clause learning SAT solvers, such as fixed-interval policy, geometric policy, Luby´s policy, nested restart scheme, adaptive restart strategy, counter implication restart and so on. It has been demonstrated that appropriate use of restarts can improve the speed of SAT solvers tremendously. Growing numbers of studies have been beginning to pay attention to restart strategies, from triggering a restart to selecting the first decision after a restart. However, the number of restarts, an important state parameter of SAT solver, has never gotten any attention before. This paper introduces a novel heuristic method of decision-making after a restart by making full use of the number of each variable´s restart times, which is named VRT restart strategy in this paper. We integrated it into a modern SAT solver. Experimental results show that the solver with our strategy is considerably faster than the original one.
Keywords :
computability; computational complexity; decision making; learning (artificial intelligence); Luby policy; NP-complete problem; SAT solver speed; adaptive restart strategy; conflict-driven clause learning SAT solver; counter implication restart; decision making; decision selection; fixed-interval policy; geometric policy; heuristic restart strategy; nested restart scheme; satisfiability problem solving; state parameter; Algorithm design and analysis; Artificial intelligence; Computer hacking; Computer science; Educational institutions; Radiation detectors; Reactive power; SAT solver; Satisfiability Problem; restart strategy; variable´s restart times;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computational Intelligence and Design (ISCID), 2012 Fifth International Symposium on
Conference_Location :
Hangzhou
Print_ISBN :
978-1-4673-2646-9
Type :
conf
DOI :
10.1109/ISCID.2012.254
Filename :
6406029
Link To Document :
بازگشت