A single-loop proximal-conditional-gradient penalty method

TL;DR

This paper introduces a novel single-loop proximal-conditional-gradient penalty method for convex optimization with linear constraints, achieving explicit convergence rates and demonstrating effectiveness on matrix completion problems.

Contribution

It proposes a new single-loop algorithm combining proximal-gradient and conditional-gradient steps with explicit convergence analysis for constrained convex problems.

Findings

Objective deviations decay at rate t^{- ext{min}\{ ext{ extmu}, ext{ extnu},1/2 ight"]}

Feasibility violations decay at rate t^{-1/2}

Numerical results on low rank Hankel matrix completion show effectiveness.

Abstract

We consider the problem of minimizing a convex separable objective (as a separable sum of two proper closed convex functions $f$ and $g$) over a linear coupling constraint. We assume that $f$ can be decomposed as the sum of a smooth part having H\"older continuous gradient (with exponent $\mu\in(0,1]$) and a nonsmooth part that admits efficient proximal mapping computations, while $g$ can be decomposed as the sum of a smooth part having H\"older continuous gradient (with exponent $\nu\in(0,1]$) and a nonsmooth part that admits efficient linear oracles. Motivated by the recent works [1,49], we propose a single-loop variant of the standard penalty method, which we call a single-loop proximal-conditional-gradient penalty method (proxCG$^{\rm pen}_{1\ell}$), for this problem. In each iteration of proxCG$^{\rm pen}_{1\ell}$, we successively perform one proximal-gradient step involving $f$ and one conditional-gradient step involving $g$ on the quadratic penalty function, followed by an update of the penalty parameter. We present explicit rules for updating the penalty parameter and the stepsize in the conditional-gradient step in each iteration. Under a standard constraint qualification and domain boundedness assumption, we show that the objective value deviations (from the optimal value) along the sequence generated decay in the order of $t^{-\min\{\mu,\nu,1/2\}}$ with the associated feasibility violations decaying in the order of $t^{-1/2}$. Moreover, if the nonsmooth parts are indicator functions and the extended objective is a KL function with exponent $\alpha\in[0,1)$, then the distances to the optimal solution set along the sequence generated by proxCG$^{\rm pen}_{1\ell}$ decay asymptotically at a rate of $t^{-(1-\alpha)\min\{\mu,\nu,1/2\}}$. Finally, we illustrate numerically the behavior of proxCG$^{\rm pen}_{1\ell}$ on solving low rank Hankel matrix completion problems.