1: \begin{abstract}

2: In decentralized consensus optimization, a connected network of

3: agents collaboratively minimize the sum of their local objective

4: functions over a common decision variable, where their information

5: exchange is restricted between the neighbors. To this end, one can

6: first obtain a problem reformulation and then apply the

7: alternating direction method of multipliers (ADMM). The method

8: applies iterative computation at the individual agents and

9: information exchange between the neighbors. This approach has been

10: observed to converge quickly and deemed powerful. This paper

11: establishes its linear convergence rate for the decentralized

12: consensus optimization problem with strongly convex local

13: objective functions. The theoretical convergence rate is

14: explicitly given in terms of the network topology, the properties

15: of local objective functions, and the algorithm parameter. This

16: result is not only a performance guarantee but also a guideline

17: toward accelerating the ADMM convergence.

18: \end{abstract}