Abstract:Abstract. The augmented Lagrangian method (ALM) provides a benchmark for solving the 5 canonical convex programming problem with linear equality constraints; it also inspires a number of 6 variants for various applications in different areas. Among them, the proximal version of ALM, by 7 adding an additional positive-definite quadratic proximal term to the estimation of primal variable, 8 is often implemented to tackle many problems and its convergence has been well studied in the liter9 ature. The most recent study reveals an interesting fact that the positive definite regularization term 10 is not necessarily required to ensure the convergence of the proximal ALM, which hence results in 11 an indefinite linearized ALM with a larger step-size for speeding up convergence. The literature on 12 the inequality-constrained ALM, however, is relatively few. In this paper, exploiting the variational 13 inequality structure of the indefinite linearized ALM, we present a novel indefinite linearized ALM 14 framework to study the linear inequality constrained convex optimization problem, which gives rise 15 to a new ALM-based method. The proposed method enjoys great advantages in two folds: first, it 16 reduces the computation difficulty of the core subproblem in the original inequality-constrained ALM 17 to the one in the linearized equality-constrained ALM; second, the relaxation of the proximal regu18 larization term allows a bigger step size to potentially reduce required convergence iterations. The 19 extensive numerical experiments on some convex-relaxed image segmentation models demonstrate 20 the efficiency of the novel method. 21

An Augmented Lagrangian Method for Non-Lipschitz Nonconvex Programming

Augmented Lagrangian Methods for Convex Matrix Optimization Problems

A Smoothing Augmented Lagrangian Method for Nonconvex, Nonsmooth Constrained Programs and Its Applications to Bilevel Problems

Nonconvex and Nonsmooth Optimization with Generalized Orthogonality Constraints: an Approximate Augmented Lagrangian Method

The Augmented Lagrangian Method for Mathematical Programs with Vertical Complementarity Constraints Based on Inexact Scholtes Regularization

Proximal augmented lagrangian method for convex

Augmented Lagrangian optimization under fixed-point arithmetic

Smoothing Augmented Lagrangian Method for Nonsmooth Constrained Optimization Problems.

An Augmented Lagrangian Method for State Constrained Linear Parabolic Optimal Control Problems

An Augmented Lagrangian Method for a Class of Inverse Quadratic Programming Problems

Inexact Augmented Lagrangian Methods for Conic Programs: Quadratic Growth and Linear Convergence

An Inexact Augmented Lagrangian Framework for Nonconvex Optimization with Nonlinear Constraints

Indefinite Linearized Augmented Lagrangian Method for Convex Programming with Linear Inequality Constraints

A New Insight on Augmented Lagrangian Method with Applications in Machine Learning

The inexact power augmented Lagrangian method for constrained nonconvex optimization

A New Augmented Lagrangian Method for MPCCs - Theoretical and Numerical Comparison with Existing Augmented Lagrangian Methods

On the Convergence Properties of a Second-Order Augmented Lagrangian Method for Nonlinear Programming Problems with Inequality Constraints

Global Convergence Analysis of the Power Proximal Point and Augmented Lagrangian Method

A first-order augmented Lagrangian method for constrained minimax optimization

An Efficient Augmented Lagrangian Method with Applications to Total Variation Minimization.

A First-Order Primal-Dual Method for Nonconvex Constrained Optimization Based on the Augmented Lagrangian