Score: 0

Smoothness of the Augmented Lagrangian Dual in Convex Optimization

Published: May 3, 2025 | arXiv ID: 2505.01824v1

By: Jingwang Li, Vincent Lau

Potential Business Impact:

Makes math problems with limits easier to solve.

Business Areas:
A/B Testing Data and Analytics

This paper investigates the general linearly constrained optimization problem: $\min_{x \in \R^d} f(x) \ \st \ A x = b$, where $f: \R^n \rightarrow \exs$ is a closed proper convex function, $A \in \R^{p \times d}$, and $b \in \R^p$. We establish the following results without requiring additional regularity conditions: (1) the augmented Lagrangian dual function $\phi_{\rho}(\lambda) = \inf_x \cL_{\rho}(x, \lambda)$ is $\frac{1}{\rho}$-smooth everywhere; and (2) the solution to $\min_{x \in \R^d} \cL_{\rho}(x, \lambda)$ exists for any dual variable $\lambda \in \R^p$, where $\rho > 0$ is the augmented parameter and $\cL_{\rho}(x, \lambda) = f(x) + \dotprod{\lambda, A x - b} + \frac{\rho}{2}\norm{A x - b}^2$ is the augmented Lagrangian. These findings significantly relax the strong assumptions commonly imposed in existing literature to guarantee similar properties.

Country of Origin
🇭🇰 Hong Kong

Page Count
6 pages

Category
Mathematics:
Optimization and Control