1: \begin{abstract}

2: The alternating direction method with multipliers (ADMM) has been

3: one of most powerful and successful methods for solving  various

4: composite problems. The convergence of the conventional ADMM (i.e.,

5: 2-block) for convex objective functions has been justified for a

6: long time, and its convergence for nonconvex objective functions

7: has, however, been established very recently. The multi-block ADMM,

8: a natural extension of ADMM, is a widely used scheme and has also

9: been found very useful in solving various nonconvex optimization

10: problems. It is thus expected to establish  convergence theory of

11: the multi-block ADMM under nonconvex frameworks. In this paper we

12: present a Bregman modification of 3-block ADMM and establish its

13: convergence for a large family of nonconvex functions. We further extend

14: the convergence results to the $N$-block case ($N \geq 3$),

15: which underlines the feasibility of multi-block ADMM applications in

16: nonconvex settings. Finally, we present a simulation study and a

17: real-world application to support the correctness of the obtained

18: theoretical assertions.\\

19: 

20: \noindent{\sc Keywords}: nonconvex regularization, alternating

21: direction method, subanalytic function, K-L inequality, Bregman

22: distance.

23: \end{abstract}

24: