1: \begin{abstract} 

2: We describe how the powerful ``Divide and Concur'' algorithm for constraint

3: satisfaction can be derived as a special case of a message-passing version of 

4: the Alternating Direction Method of Multipliers (ADMM) algorithm for convex

5: optimization, and introduce an improved message-passing algorithm based on ADMM/DC

6: by introducing

7: three distinct weights for messages, 

8: with ``certain'' and ``no opinion'' weights, as well as the standard weight 

9: used in ADMM/DC. The ``certain'' messages allow our 

10: improved algorithm to implement constraint propagation as a special case, while

11: the ``no opinion'' messages speed convergence for some problems by making the algorithm focus

12: only on active constraints. We describe how our three-weight version of ADMM/DC

13: can give greatly improved performance for non-convex problems such as circle packing and 

14: solving large Sudoku puzzles, while

15: retaining the exact performance of ADMM for convex problems. We also describe the

16: advantages of our algorithm compared to other message-passing algorithms based

17: upon belief propagation.

18: 

19: \end{abstract}