1: \begin{abstract}

2: We introduce a novel approach for analyzing the performance of first-order black-box

3: optimization methods. We focus on smooth unconstrained convex minimization over the Euclidean

4: space $\real^d$. Our approach relies on the observation that by definition, the worst case

5: behavior of a black-box optimization method is by itself an optimization problem, which we

6: call the Performance Estimation Problem (PEP). We formulate and analyze the PEP for two

7: classes of first-order algorithms. We first apply this approach on the classical gradient

8: method and derive a new and tight analytical bound on its performance. We then consider a

9: broader class of first-order black-box methods, which among others, include the so-called

10: heavy-ball method and the fast gradient schemes. We show that for this broader class, it is

11: possible to derive new numerical bounds on the performance of these methods by solving an

12: adequately relaxed convex semidefinite PEP. Finally, we show an efficient procedure for

13: finding optimal step sizes  which results in a first-order black-box method that achieves best

14: performance.

15: % with

16: %respect to a slightly relaxed version of the (PEP) defining the new bounds. %Our approach

17: %gives rise to some intriguing, often nontrivial analysis developments leading us to suggest a

18: %number of conjectures and raise open questions for future research.

19: \medskip

20: 

21: \noindent{\bf Keywords:} Performance of First-Order Algorithms, Rate of Convergence,

22: Complexity, Smooth Convex Minimization, Duality, Semidefinite Relaxations, Fast Gradient

23: Schemes, Heavy Ball method.

24: 

25: \end{abstract}