1: \begin{abstract}

2: In this paper we propose a distributed dual gradient  algorithm for

3: minimizing  linearly constrained separable  convex problems and

4: analyze its rate of convergence. In particular, we prove that under

5: the assumption of  strong convexity and Lipshitz continuity of the

6: gradient of the primal objective function we have a global error

7: bound type property for the dual problem. Using this error bound

8: property we devise a  fully distributed dual gradient scheme, i.e. a

9: gradient  scheme based on a weighted step size,  for which we derive

10: global linear rate of convergence for both dual and primal

11: suboptimality and for primal feasibility violation.  Many real

12: applications, e.g. distributed model predictive control, network

13: utility maximization or optimal power flow, can be posed as linearly

14: constrained separable  convex problems for which dual gradient type

15: methods from literature  have sublinear convergence rate. In the

16: present paper we prove for the first time that in fact we can

17: achieve linear convergence rate for such algorithms when they are

18: used for solving these applications.  Numerical simulations are also

19: provided to confirm our theory.

20: \end{abstract}