1: \begin{abstract}

2: We investigate the question of tightness of linear programming (LP)

3: relaxation for finding a maximum weight independent set (MWIS) in 

4: sparse random weighted graphs. We show that an edge-based LP relaxation is 

5: asymptotically tight for Erdos-Renyi graph $G(n,c/n)$ for $c \leq 2e$ 

6: and random regular graph $G(n,r)$ for $r\leq 4$ when node weights

7: are i.i.d. with exponential distribution of mean $1$. We establish 

8: these results, through a precise relation between the tightness of 

9: LP relaxation and convergence of the max-product belief propagation 

10: algorithm. We believe that this novel method of understanding 

11: structural properties of combinatorial problems through properties

12: of iterative procedure such as the max-product should be of interest 

13: in its own right. 

14: \end{abstract}

15: