On the cavity method for decimated random constraint satisfaction problems and the analysis of belief propagation guided decimation algorithms - Condensed Matter > Disordered Systems and Neural NetworksReportar como inadecuado




On the cavity method for decimated random constraint satisfaction problems and the analysis of belief propagation guided decimation algorithms - Condensed Matter > Disordered Systems and Neural Networks - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: We introduce a version of the cavity method for diluted mean-field spinmodels that allows the computation of thermodynamic quantities similar to theFranz-Parisi quenched potential in sparse random graph models. This method isdeveloped in the particular case of partially decimated random constraintsatisfaction problems. This allows to develop a theoretical understanding of aclass of algorithms for solving constraint satisfaction problems, in whichelementary degrees of freedom are sequentially assigned according to theresults of a message passing procedure belief-propagation. We confront thistheoretical analysis to the results of extensive numerical simulations.



Autor: Federico Ricci-Tersenghi, Guilhem Semerjian

Fuente: https://arxiv.org/







Documentos relacionados