English

On the cavity method for decimated random constraint satisfaction problems and the analysis of belief propagation guided decimation algorithms

Disordered Systems and Neural Networks 2015-05-13 v1 Statistical Mechanics Discrete Mathematics

Abstract

We introduce a version of the cavity method for diluted mean-field spin models that allows the computation of thermodynamic quantities similar to the Franz-Parisi quenched potential in sparse random graph models. This method is developed in the particular case of partially decimated random constraint satisfaction problems. This allows to develop a theoretical understanding of a class of algorithms for solving constraint satisfaction problems, in which elementary degrees of freedom are sequentially assigned according to the results of a message passing procedure (belief-propagation). We confront this theoretical analysis to the results of extensive numerical simulations.

Keywords

Cite

@article{arxiv.0904.3395,
  title  = {On the cavity method for decimated random constraint satisfaction problems and the analysis of belief propagation guided decimation algorithms},
  author = {Federico Ricci-Tersenghi and Guilhem Semerjian},
  journal= {arXiv preprint arXiv:0904.3395},
  year   = {2015}
}

Comments

32 pages, 24 figures