English

EXTRA: An Exact First-Order Algorithm for Decentralized Consensus Optimization

Optimization and Control 2014-12-02 v4

Abstract

Recently, there have been growing interests in solving consensus optimization problems in a multi-agent network. In this paper, we develop a decentralized algorithm for the consensus optimization problem minxRp fˉ(x)=1ni=1nfi(x),\min\limits_{x\in\mathbb{R}^p}~\bar{f}(x)=\frac{1}{n}\sum\limits_{i=1}^n f_i(x), which is defined over a connected network of nn agents, where each function fif_i is held privately by agent ii and encodes the agent's data and objective. All the agents shall collaboratively find the minimizer while each agent can only communicate with its neighbors. Such a computation scheme avoids a data fusion center or long-distance communication and offers better load balance to the network. This paper proposes a novel decentralized EXact firsT-ordeR Algorithm (abbreviated as EXTRA) to solve the consensus optimization problem. "exact" means that it can converge to the exact solution. EXTRA can use a fixed large step size, {which is independent of the network size}, and has synchronized iterations. The local variable of every agent ii converges uniformly and consensually to an exact minimizer of fˉ\bar{f}. In contrast, the well-known decentralized gradient descent (DGD) method must use diminishing step sizes in order to converge to an exact minimizer. EXTRA and DGD have the same choice of mixing matrices and similar per-iteration complexity. EXTRA, however, uses the gradients of last two iterates, unlike DGD which uses just that of last iterate. EXTRA has the best known convergence rates among the existing first-order decentralized algorithms. Specifically, if fif_i's are convex and have Lipschitz continuous gradients, EXTRA has an ergodic convergence rate O(1k)O(\frac{1}{k}) in terms of the first-order optimality residual. If fˉ\bar{f} is also restricted strongly convex, EXTRA converges to an optimal solution at a linear rate O(Ck)O(C^{-k}) for some constant C>1C>1.

Keywords

Cite

@article{arxiv.1404.6264,
  title  = {EXTRA: An Exact First-Order Algorithm for Decentralized Consensus Optimization},
  author = {Wei Shi and Qing Ling and Gang Wu and Wotao Yin},
  journal= {arXiv preprint arXiv:1404.6264},
  year   = {2014}
}
R2 v1 2026-06-22T03:58:16.942Z