Accelerated Decentralized Constraint-Coupled Optimization: A Dual$^2$ Approach

TL;DR

This paper introduces two accelerated algorithms for decentralized constraint-coupled optimization, achieving faster convergence and lower complexity by leveraging a novel dual$^2$ approach.

Contribution

The paper proposes a new dual$^2$ method and two accelerated algorithms with improved convergence rates and complexity bounds for decentralized optimization.

Findings

Both algorithms guarantee asymptotic convergence under milder conditions.

Under additional assumptions, they achieve linear convergence rates.

Numerical experiments confirm theoretical advantages and practical efficiency.

Abstract

In this paper, we focus on a class of decentralized constraint-coupled optimization problem: $\min_{x_i \in \mathbb{R}^{d_i}, i \in \mathcal{I}; y \in \mathbb{R}^p}$ $\sum_{i=1}^n\left(f_i(x_i) + g_i(x_i)\right) + h(y) \ \text{s.t.} \ \sum_{i=1}^{n}A_ix_i = y$, over an undirected and connected network of $n$ agents. Here, $f_i$, $g_i$, and $A_i$ represent private information of agent $i \in \mathcal{I} = \{1, \cdots, n\}$, while $h$ is public for all agents. Building on a novel dual$^2$ approach, we develop two accelerated algorithms to solve this problem: the inexact Dual$^2$ Accelerated (iD2A) gradient method and the Multi-consensus inexact Dual$^2$ Accelerated (MiD2A) gradient method. We demonstrate that both iD2A and MiD2A can guarantee asymptotic convergence under a milder condition on $h$ compared to existing algorithms. Furthermore, under additional assumptions, we establish linear convergence rates and derive significantly lower communication and computational complexity bounds than those of existing algorithms. Several numerical experiments validate our theoretical analysis and demonstrate the practical superiority of the proposed algorithms.