Author :
Tu, Ronghui ; Mao, Yongyi ; Zhao, Jiying
Author_Institution :
Sch. of Inf. Technol. & Eng., Univ. of Ottawa, Ottawa, ON, Canada
fDate :
6/1/2010 12:00:00 AM
Abstract :
The survey propagation (SP) algorithm for solving k-SAT problems has been shown recently as an instance of the belief propagation (BP) algorithm. In this paper, we show that for general constraint-satisfaction problems, SP may not be reducible from BP. We also establish the conditions under which such a reduction is possible. Along our development, we present a unification of the existing SP algorithms in terms of a probabilistically interpretable iterative procedure - weighted probabilistic token passing.
Keywords :
Markov processes; belief networks; constraint theory; message passing; protocols; belief propagation algorithm; general constraint-satisfaction problems; k-SAT problems; probabilistically interpretable iterative procedure; survey propagation algorithm; weighted probabilistic token passing; Algorithm design and analysis; Belief propagation; Heuristic algorithms; Inference algorithms; Iterative algorithms; Markov random fields; NP-complete problem; Physics; Prototypes; Viterbi algorithm; $k$-SAT; $q$-COL; Markov random field; belief propagation; constraint satisfaction; factor graph; message-passing algorithm; survey propagation (SP);
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.2010.2046236
