dc.contributor.author | Newsham, Zackary | |
dc.date.accessioned | 2015-01-21 14:15:55 (GMT) | |
dc.date.available | 2015-05-22 05:30:11 (GMT) | |
dc.date.issued | 2015-01-21 | |
dc.date.submitted | 2015 | |
dc.identifier.uri | http://hdl.handle.net/10012/9110 | |
dc.description.abstract | Despite enormous improvements in Boolean SATisfiability solver performance
over the last decade, it is still unclear why specific input formula are slow
to solve, when other similarly specified formula execute more quickly. This
work explores the relationship between the community structure of a SAT
formula and its execution time on several state-of-the-art solvers. We explore
the analysis of this data from a number of directions; first, we explore the
relationship between the well known clause-variable ratio result, and com-
munity structure in randomly generated instances. Second, we perform a
standard linear regression on data obtained from the 2013 SAT competition.
Third, we present a visualisation tool and data repository for viewing the
structure of a SAT formula. Fourth, we explore the effect of hardware con-
straints on the solution time of instances across various machines. Finally,
we explore survival analysis, a technique that is new to the field of Boolean
SATisfiability. By collating the results from each of these experiments, we
have determined that the community structure is critical in determining the
solution time of a SAT formula, more important than the clause-variable
ratio of the formula. While this work is not a complete explanation of the
varying solution time of SAT formulae, it has provided us with significant
insight for further research to answer the question: why different similarly
specified formula have such different solution times? | en |
dc.language.iso | en | en |
dc.publisher | University of Waterloo | en |
dc.subject | community | en |
dc.subject | survival | en |
dc.subject | SAT | en |
dc.subject | solver | en |
dc.subject | boolean satisfiability | en |
dc.subject | linear regression | en |
dc.title | An Analysis of the Effect of Community Structure on SAT Solver Performance | en |
dc.type | Master Thesis | en |
dc.pending | false | |
dc.subject.program | Electrical and Computer Engineering | en |
dc.description.embargoterms | 4 months | en |
uws-etd.degree.department | Electrical and Computer Engineering | en |
uws-etd.degree | Master of Applied Science | en |
uws.typeOfResource | Text | en |
uws.peerReviewStatus | Unreviewed | en |
uws.scholarLevel | Graduate | en |