English   español  
Por favor, use este identificador para citar o enlazar a este item: http://hdl.handle.net/10261/139428
logo share SHARE logo core CORE   Add this article to your Mendeley library MendeleyBASE

Visualizar otros formatos: MARC | Dublin Core | RDF | ORE | MODS | METS | DIDL
Exportar a otros formatos:

Generating SAT instances with community structure

AutorGiraldez-Cru, Jesus; Levy, Jordi
Palabras claveSAT solver
SAT generator
Graph modularity
Fecha de publicación2016
CitaciónArtificial Intelligence 238: 119- 134 (2016)
ResumenNowadays, modern SAT solvers are able to efficiently solve many industrial, or real-world, SAT instances. However, the process of development and testing of new SAT solving techniques is conditioned to the finite and reduced number of known industrial benchmarks. Therefore, new models of random SAT instances generation that capture realistically the features of real-world problems can be beneficial to the SAT community. In many works, the structure of industrial instances has been analyzed representing them as graphs and studying some of their properties, like modularity. In this work, we use the notion of modularity to define a new model of generation of random SAT instances with community structure, called Community Attachment. For high values of modularity (i.e., clear community structure), we realistically model pseudo-industrial random SAT formulas. This model also generates SAT instances very similar to classical random formulas using a low value of modularity. We also prove that the phase transition point, if exists, is independent on the modularity. We evaluate the adequacy of this model to real industrial SAT problems in terms of SAT solvers performance, and show that modern solvers do actually exploit this community structure. Finally, we use this generator to observe the connections between the modularity of the instance and some components of the solver, such as the variable branching heuristics or the clause learning mechanism.
Identificadoresdoi: 10.1016/j.artint.2016.06.001
issn: 0004-3702
Aparece en las colecciones: (IIIA) Artículos
Ficheros en este ítem:
Fichero Descripción Tamaño Formato  
accesoRestringido.pdf15,38 kBAdobe PDFVista previa
Mostrar el registro completo

NOTA: Los ítems de Digital.CSIC están protegidos por copyright, con todos los derechos reservados, a menos que se indique lo contrario.