1: \begin{abstract}

2: In this note we study an iterative belief propagation (IBP) algorithm and demonstrate it's ability to solve sparse combinatorial optimization problems. Similar to simulated annealing (SA) \cite{Kirkpatrick1983_SA}, our IBP algorithm attempts to sample from the Boltzmann distribution of the objective function but also uses belief propagation (BP) \cite{Braunstein2005_SP, Pearl1982_BP} to improve convergence.

3: \end{abstract}

4: