1: \begin{abstract}

2: Consider a set of $N$ agents seeking to solve distributively the 

3: minimization problem $\inf_{x} \sum_{n = 1}^N f_n(x)$ where the convex 

4: functions $f_n$ are local to the agents. 

5: The popular Alternating Direction Method of Multipliers has the

6: potential to handle distributed optimization problems of this kind.

7: \modifPB{We provide a general reformulation of the problem

8: and obtain a class of distributed algorithms which encompass

9:  various network architectures.}

10: The rate of convergence of \modifPB{our} method is considered. 

11: It is assumed that the infimum of the problem is reached at a point $x_\star$, 

12: the functions $f_n$ are twice differentiable

13: at this point and $\sum \nabla^2 f_n(x_\star) > 0$

14: in the positive definite ordering of symmetric matrices.

15: With these assumptions, it is shown that the convergence to the consensus 

16: $x_\star$ is linear and the exact rate is provided. 

17: Application examples where this rate can be optimized with respect to the ADMM free parameter $\rho$ are also given.

18: \end{abstract}

19: