1: \begin{abstract}

2:   We propose a framework to use Nesterov's accelerated method for

3:   constrained convex optimization problems. Our approach consists of

4:   first reformulating the original problem as an unconstrained

5:   optimization problem using a continuously differentiable exact

6:   penalty function. This reformulation is based on replacing the

7:   Lagrange multipliers in the augmented Lagrangian of the original

8:   problem by Lagrange multiplier functions. The expressions of these

9:   Lagrange multiplier functions, which depend upon the gradients of

10:   the objective function and the

11:   constraints, 

12:   can make the unconstrained penalty function non-convex in general

13:   even if the original problem is convex. We establish sufficient

14:   conditions on the objective function and the constraints of the

15:   original problem under which the unconstrained penalty function is

16:   convex. This enables us to use Nesterov's accelerated gradient

17:   method for unconstrained convex optimization and achieve a

18:   guaranteed rate of convergence which is better than the

19:   state-of-the-art first-order algorithms for constrained convex

20:   optimization. Simulations illustrate our results.

21: \end{abstract}

22: