Time rescaling of a primal-dual dynamical system with asymptotically vanishing damping

TL;DR

This paper introduces a time-rescaled primal-dual dynamical system with vanishing damping for convex optimization with linear constraints, demonstrating improved convergence rates and asymptotic optimality under certain conditions.

Contribution

It presents a novel time rescaling of a primal-dual dynamical system with asymptotically vanishing damping, achieving faster convergence and new trajectory convergence results.

Findings

Fast convergence of primal-dual gap, feasibility, and objective value.

Improved convergence rates depend on the rescaling parameter.

Weak convergence of trajectories to primal-dual optimal solutions.

Abstract

In this work, we approach the minimization of a continuously differentiable convex function under linear equality constraints by a second-order dynamical system with an asymptotically vanishing damping term. The system under consideration is a time rescaled version of another system previously found in the literature. We show fast convergence of the primal-dual gap, the feasibility measure, and the objective function value along the generated trajectories. These convergence rates now depend on the rescaling parameter, and thus can be improved by choosing said parameter appropriately. When the objective function has a Lipschitz continuous gradient, we show that the primal-dual trajectory asymptotically converges weakly to a primal-dual optimal solution to the underlying minimization problem. We also exhibit improved rates of convergence of the gradient along the primal trajectories and of the adjoint of the corresponding linear operator along the dual trajectories. Even in the unconstrained case, some trajectory convergence result seems to be new. We illustrate the theoretical outcomes through numerical experiments.