1: \begin{abstract}

2: We show that the exact worst-case performance of fixed-step

3: first-order methods for {unconstrained optimization of} smooth (possibly strongly) convex functions

4: can be obtained by solving convex programs.

5: 

6: Finding the worst-case performance of a black-box first-order method

7: is formulated as an optimization problem over a set of smooth (strongly)

8: convex functions and initial conditions. We develop closed-form necessary and

9: sufficient conditions for smooth (strongly) convex interpolation,

10: which provide a finite representation for those functions. This

11: allows us to reformulate the worst-case performance estimation problem

12: as an equivalent finite dimension-independent semidefinite

13: optimization problem, whose exact solution can be recovered up to numerical

14: precision. Optimal solutions to this performance estimation problem

15: provide both worst-case performance bounds and explicit functions

16: matching them, as our smooth (strongly) convex interpolation procedure is

17: constructive.

18: 

19: Our works build on those of Drori and Teboulle

20: in {[Math.\@ Prog.\@ 145 (1-2), 2014]} who introduced and solved relaxations of the

21: performance estimation problem for smooth convex functions.

22: 

23: We apply our approach to different fixed-step first-order methods with

24: several performance criteria, including objective function accuracy and

25: gradient norm. We conjecture several numerically supported worst-case

26: bounds on the performance of the {fixed-step} gradient, fast gradient and optimized {gradient} methods, both in the smooth convex and the smooth strongly

27: convex cases, and deduce tight estimates of the optimal step size for

28: the gradient method.

29: %\keywords{Smooth convex minimization \and Smooth convex interpolation \and First-order methods \and Worst-case analysis \and Rates of convergence \and Semidefinite programming}

30: %%\PACS{PACS code1 \and PACS code2 \and more}

31: %\subclass{90C25 \and 90C30 \and 90C60 \and 68Q25 \and 90C22 }

32: \end{abstract}

33: