Further properties of the forward-backward envelope with applications to difference-of-convex programming

TL;DR

This paper explores the properties of the forward-backward envelope in nonconvex optimization, establishing conditions for its level-boundedness and Kurdyka-{ extL}ojasiewicz property, and demonstrates its effectiveness in difference-of-convex regularized least squares problems.

Contribution

It provides new theoretical insights into the forward-backward envelope's properties and introduces a practical approach for difference-of-convex problems using BFGS, outperforming standard methods.

Findings

Forward-backward envelope is level-bounded and Kurdyka-{ extL}ojasiewicz with exponent 1/2 under certain conditions.

The proposed method outperforms nonmonotone proximal gradient in large-scale -2 regularized least squares.

Numerical results show efficiency of the BFGS-based approach in large-scale difference-of-convex problems.

Abstract

In this paper, we further study the forward-backward envelope first introduced in [28] and [30] for problems whose objective is the sum of a proper closed convex function and a twice continuously differentiable possibly nonconvex function with Lipschitz continuous gradient. We derive sufficient conditions on the original problem for the corresponding forward-backward envelope to be a level-bounded and Kurdyka-{\L}ojasiewicz function with an exponent of $\frac12$; these results are important for the efficient minimization of the forward-backward envelope by classical optimization algorithms. In addition, we demonstrate how to minimize some difference-of-convex regularized least squares problems by minimizing a suitably constructed forward-backward envelope. Our preliminary numerical results on randomly generated instances of large-scale $\ell_{1-2}$ regularized least squares problems [37] illustrate that an implementation of this approach with a limited-memory BFGS scheme usually outperforms standard first-order methods such as the nonmonotone proximal gradient method in [35].