Splitting forward-backward penalty scheme for constrained variational problems

TL;DR

This paper introduces a splitting algorithm for solving constrained variational inequalities involving maximal monotone operators, smooth convex functions, and complex constraints represented via penalization functions, with proven convergence properties.

Contribution

It proposes a unified forward-backward splitting scheme for complex variational inequalities with convergence guarantees under various conditions.

Findings

Weak ergodic convergence of the sequence to a solution.

Strong convergence when the operator is strongly monotone or the function is strongly convex.

Weak convergence of the entire sequence when the operator is a subdifferential.

Abstract

We study a forward backward splitting algorithm that solves the variational inequality \begin{equation*} A x +\nabla \Phi(x)+ N_C (x) \ni 0 \end{equation*} where $H$ is a real Hilbert space, $A: H\rightrightarrows H$ is a maximal monotone operator, $\Phi: H\to\mathbb{R}$ is a smooth convex function, and $N_C$ is the outward normal cone to a closed convex set $C\subset H$. The constraint set $C$ is represented as the intersection of the sets of minima of two convex penalization function $\Psi_1:H\to\mathbb{R}$ and $\Psi_2: H\to\mathbb{R}\cup \{+\infty\}$. The function $\Psi_1$ is smooth, the function $\Psi_2$ is proper and lower semicontinuous. Given a sequence $(\beta_n)$ of penalization parameters which tends to infinity, and a sequence of positive time steps $(\lambda_n)$, the algorithm $$ \left\{\begin{array}{rcl} x_1 & \in & H,\\ x_{n+1} & = & (I+\lambda_n A+\lambda_n\beta_n\partial\Psi_2)^{-1}(x_n-\lambda_n\nabla\Phi(x_n)-\lambda_n\beta_n\nabla\Psi_1(x_n)),\ n\geq 1. \end{array}\right. $$ performs forward steps on the smooth parts and backward steps on the other parts. Under suitable assumptions, we obtain weak ergodic convergence of the sequence $(x_n)$ to a solution of the variational inequality. Convergence is strong when either $A$ is strongly monotone or $\Phi$ is strongly convex. We also obtain weak convergence of the whole sequence $(x_n)$ when $A$ is the subdifferential of a proper lower-semicontinuous convex function. This provides a unified setting for several classical and more recent results, in the line of historical research on continuous and discrete gradient-like systems.