1: \begin{abstract}

2: Consider a set of networked agents endowed with private cost functions         

3: and seeking to find a consensus on the minimizer of the aggregate cost.        

4: A new class of random asynchronous distributed optimization methods is         

5: introduced. The methods generalize the standard Alternating Direction Method

6: of Multipliers (ADMM) to an asynchronous setting where isolated components 

7: of the network are activated in an uncoordinated fashion. The algorithms       

8: rely on the introduction of \emph{randomized} Gauss-Seidel iterations of       

9: a Douglas-Rachford operator for finding zeros of a sum of two monotone         

10: operators. Convergence to the sought minimizers is provided under mild         

11: connectivity conditions. Numerical results sustain our claims.                 

12: \end{abstract}