.png){kind=icon}
.png){kind=icon}
Return
Achieving Linear Convergence in Distributed Asynchronous Multiagent Optimization
.png){kind=icon}
OA
AI
DOI：10.1109/TAC.2020.2977940.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This article studies multiagent (convex and nonconvex) optimization over static digraphs. We propose a general distributed asynchronous algorithmic framework whereby 1) agents can update their local variables as well as communicate with their neighbors at any time, without any form of coordination; and 2) they can perform their local computations using (possibly) delayed, out-of-sync information from the other agents. Delays need not be known to the agent or obey any specific profile, and can also be time-varying (but bounded). The algorithm builds on a tracking mechanism that is robust against asynchrony (in the above sense), whose goal is to estimate locally the average of agents' gradients. When applied to strongly convex functions, we prove that it converges at an R-linear (geometric) rate as long as the step-size is sufficiently small. A sublinear convergence rate is proved, when nonconvex problems and/or diminishing, uncoordinated step-sizes are considered. To the best of our knowledge, this is the first distributed algorithm with provable geometric convergence rate in such a general asynchronous setting. Preliminary numerical results demonstrate the efficacy of the proposed algorithm and validate our theoretical findings.
Keywords:
Delays
Convergence
Optimization
Convex functions
Distributed algorithms
Robustness
Indexes
Asynchrony
delay
directed graphs
distributed optimization
linear convergence
nonconvex optimization
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
7
Papers:
1.3W
Citations:
6.7W
