1: \begin{abstract}

2: We develop two new variants of alternating direction methods of multipliers (ADMM) and two parallel primal-dual decomposition algorithms to solve a wide range class of constrained convex optimization problems.

3: Our approach relies on a novel combination of the augmented Lagrangian framework, partial alternating/linearization scheme, Nesterov's acceleration technique, and adaptive strategy.

4: The proposed algorithms have the following new features compared to existing ADMM variants.

5: First, they have a Nesterov's acceleration step on the primal variables instead of the dual ones as in several  ADMM variants.

6: Second, they possess an optimal $\BigO{\frac{1}{k}}$-convergence rate guarantee in a non-ergodic sense without any smoothness or strong convexity-type assumption, where $k$ is the iteration counter. 

7: When one objective term is strongly convex, our algorithm achieves an optimal $\BigO{\frac{1}{k^2}}$-non-ergodic rate.

8: Third, our methods have better per-iteration complexity than standard ADMM due to the linearization step in the second subproblem.

9: Fourth, we provide a set of conditions to derive update rules for algorithmic parameters, and give a concrete update for these parameters as an example.

10: Finally, when the objective function is separable, our methods can naturally be implemented in a parallel fashion.

11: We also study two extensions of our methods and a connection to existing  primal-dual methods.

12: We verify our theoretical development via different numerical examples and compare our methods with some existing state-of-the-art algorithms.

13: \end{abstract}

14: