Fast Gradient Methods for Uniformly Convex and Weakly Smooth Problems
arXiv:2103.12349 · doi:10.1007/s10444-022-09943-5
Abstract
In this paper, acceleration of gradient methods for convex optimization problems with weak levels of convexity and smoothness is considered. Starting from the universal fast gradient method which was designed to be an optimal method for weakly smooth problems whose gradients are Hölder continuous, its momentum is modified appropriately so that it can also accommodate uniformly convex and weakly smooth problems. Different from the existing works, fast gradient methods proposed in this paper do not use the restarting technique but use momentums that are suitably designed to reflect both the uniform convexity and the weak smoothness information of the target energy function. Both theoretical and numerical results that support the superiority of proposed methods are presented.
arXiv comment32 pages, 7 figures
References in corpus (7)
- Optimized first-order methods for smooth convex minimization
- Preconditioned Steepest Descent Methods for some Nonlinear Elliptic Equations Involving p-Laplacian Terms
- Primal-dual subgradient methods for minimizing uniformly convex functions
- Generalizing the optimized gradient method for smooth convex minimization
- Adaptive restart of accelerated gradient methods under local quadratic growth condition
- Additive Schwarz Methods for Convex Optimization as Gradient Methods
- Pseudo-linear Convergence of an Additive Schwarz Method for Dual Total Variation Minimization