1: \begin{abstract}

2:     The paper proposes a linesearch for a primal-dual method. Each

3: iteration of the linesearch requires to update only the dual (or primal)

4: variable. For many problems, in particular for regularized least

5: squares, the linesearch does not require any additional matrix-vector

6: multiplications. We prove convergence of the proposed method under

7: standard assumptions. We also show an ergodic $O(1/N)$ rate of

8: convergence for our method.  In case one or both of the prox-functions

9: are strongly convex, we modify our basic method to get a  better convergence rate.

10: Finally, we propose a linesearch for a saddle point

11: problem with an additional smooth term. Several numerical experiments confirm the

12: efficiency of our proposed methods.

13: \end{abstract}

14: