1: \begin{abstract}

2: The standard assumption for proving linear convergence of first

3: order methods for smooth convex optimization is the strong convexity

4: of the objective function, an assumption which does not hold for

5: many practical applications. In this paper, we derive linear

6: convergence rates of several first order methods for solving

7: smooth non-strongly convex constrained optimization problems, i.e.

8: involving an objective function with a Lipschitz continuous gradient

9: that satisfies some relaxed strong convexity condition. In

10: particular, in the case of  smooth constrained convex optimization,

11: we provide several relaxations of the strong convexity conditions and

12: prove that they  are sufficient for getting linear convergence for

13: several first order methods such as projected gradient, fast

14: gradient and feasible descent methods. We also provide examples of

15: functional classes  that satisfy our proposed relaxations of strong

16: convexity conditions.  Finally, we show that the

17: proposed relaxed strong convexity conditions cover important

18: applications ranging from solving  linear systems, Linear

19: Programming, and dual formulations of linearly constrained convex

20: problems.

21: \end{abstract}