Abstract:The generalized Lasso is a remarkably versatile and extensively utilized model across a broad spectrum of domains, including statistics, machine learning, and image science. Among the optimization techniques employed to address the challenges posed by this model, saddle-point methods stand out for their effectiveness. In particular, the primal-dual hybrid gradient (PDHG) algorithm has emerged as a highly popular choice, celebrated for its robustness and efficiency in finding optimal solutions. Recently, the underlying mechanism of the PDHG algorithm has been elucidated through the high-resolution ordinary differential equation (ODE) and the implicit-Euler scheme as detailed in [Li and Shi, 2024a]. This insight has spurred the development of several accelerated variants of the PDHG algorithm, originally proposed by [Chambolle and Pock, 2011]. By employing discrete Lyapunov analysis, we establish that the PDHG algorithm with iteration-varying step sizes, converges at a rate near $O(1/k^2)$. Furthermore, for the specific setting where $\tau_{k+1}\sigma_k = s^2$ and $\theta_k = \tau_{k+1}/\tau_k \in (0, 1)$ as proposed in [Chambolle and Pock, 2011], an even faster convergence rate of $O(1/k^2)$ can be achieved. To substantiate these findings, we design a novel discrete Lyapunov function. This function is distinguished by its succinctness and straightforwardness, providing a clear and elegant proof of the enhanced convergence properties of the PDHG algorithm under the specified conditions. Finally, we utilize the discrete Lyapunov function to establish the optimal linear convergence rate when both the objective functions are strongly convex.

On the Local and Superlinear Convergence of a Parameterized DFP Method

Convergence analysis of the DFP algorithm for unconstrained optimization problems on Riemannian manifolds

A New Filled Function Method with Two Parameters for Global Optimization

An inertial hybrid DFPM-based algorithm for constrained nonlinear equations with applications

A Novel Adaptive Filter Algorithm Based on Dfp Technique

U-D factorization-based nonlinear programming method and its application in neural network training

A NOTE ON THE CONVERGENCE OF THE DFP ALGORITHM ON QUADRATIC UNIFORMLY CONVEX FUNCTIONS

On the Global Convergence of A Parameter-Adjusting Levenberg-Marquardt Method

Superlinear Convergence of an Interior Point Algorithm on Linear Semi-definite Feasibility Problems

Updating to Optimal Parametric Values by Memory-Dependent Methods: Iterative Schemes of Fractional Type for Solving Nonlinear Equations

Convergence and Superconvergence Analysis of a Nonconforming Finite Element Variable-Time-Step BDF2 Implicit Scheme for Linear Reaction-Diffusion Equations

Fitted L1-ADI Scheme for Improving Convergence of Two-Dimensional Delay Fractional Equations

A New Dai-Liao Conjugate Gradient Method with Optimal Parameter Choice

A parameterized Douglas-Rachford Splitting algorithm for nonconvex optimization

A Superlinear Convergence Framework for Kurdyka-{\L}ojasiewicz Optimization

A Lyapunov Analysis of Accelerated PDHG Algorithms

A superlinear convergence iterative framework for Kurdyka-{\L}ojasiewicz optimization and application

Linear Convergence of Randomized Primal-Dual Coordinate Method for Large-scale Linear Constrained Convex Programming

Diagonal and normal with Toeplitz-block splitting iteration method for space fractional coupled nonlinear Schrödinger equations with repulsive nonlinearities

A NOTE ON THE OPTIMAL PARAMETER OF BABAIE-KAFAKI'S THREE-TERM CONJUGATE GRADIENT METHOD