1: \begin{abstract}In this paper we provide a generalization of the

2: Douglas-Rachford splitting (DRS) \lb{and the primal-dual algorithm 

3: \cite{condat,vu}} for solving monotone inclusions

4: in a real Hilbert space involving a general linear operator. The 

5: proposed method \lb{allows for primal and dual non-standard metrics 

6: and }activates the linear operator separately from 

7: the monotone operators appearing in the inclusion\lb{. I}n the 

8: simplest case when the linear operator \lb{has full range}, it reduces 

9: to classical DRS. Moreover, the weak convergence of primal-dual 

10: sequences to \lb{a Kuhn-Tucker point} 

11: is guaranteed, generalizing the main result in \cite{svaiter}. 

12: Inspired \lb{by} \cite{gabay83}, we \lb{also}

13: derive a new Split-ADMM (SADMM) by applying our method to the 

14: dual of a  convex optimization problem involving a linear operator  

15: which can be expressed as the composition of two linear

16:  operators. The proposed SADMM activates one linear operator 

17:  implicitly and the other \lb{one} explicitly\lb{,} and we recover ADMM 

18:  when the 

19:  latter is set as the identity. Connections and comparisons of our 

20:  theoretical results with respect to the literature are provided for 

21:  the main algorithm and SADMM\lb{. T}he flexibility and efficiency 

22:  of \lb{both methods} is illustrated via a 

23: numerical simulations \lb{in total variation image restoration and} a 

24: sparse 

25: minimization problem.

26: \par

27: \bigskip

28: 

29: \noindent \textbf{Keywords.} {\it ADMM, convex optimization, Douglas--Rachford splitting, fixed 

30: point iterations, 

31: monotone operator  theory, quasinonexpansive operators, 

32: splitting algorithms.}

33: \end{abstract}

34: