1: \begin{abstract}

2: Applying the max-product (and belief-propagation) algorithms to loopy

3: graphs is now quite popular for best assignment problems. This is

4: largely due to their low computational complexity and impressive performance

5: in practice. Still, there is no general understanding of the conditions

6: required for convergence and/or the optimality of converged solutions.

7: This paper presents an analysis of both attenuated max-product (AMP)

8: decoding and weighted min-sum (WMS) decoding for LDPC codes which

9: guarantees convergence to a fixed point when a weight parameter, $\beta$,

10: is sufficiently small. It also shows that, if the fixed point satisfies

11: some consistency conditions, then it must be both the linear-programming

12: (LP) and maximum-likelihood (ML) solution. 

13: 

14: For $(d_{v},d_{c})$-regular LDPC codes, the weight must satisfy $\beta(d_{v}-1)\leq1$

15: whereas the results proposed by Koetter and Frey require instead that

16: $\beta(d_{v}-1)(d_{c}-1)<1$. A counterexample which shows a fixed

17: point might not be the ML solution if $\beta(d_{v}-1)>1$ is also

18: given. Finally, connections are explored with recent work by Arora

19: et al. on the threshold of LP decoding.\end{abstract}

20: