Network Topology and Communication-Computation Tradeoffs in Decentralized Optimization

Angelia Nedich, Alex Olshevsky, Michael G. Rabbat

Research output: Contribution to journalReview article

18 Citations (Scopus)

Abstract

In decentralized optimization, nodes cooperate to minimize an overall objective function that is the sum (or average) of per-node private objective functions. Algorithms interleave local computations with communication among all or a subset of the nodes. Motivated by a variety of applications..decentralized estimation in sensor networks, fitting models to massive data sets, and decentralized control of multirobot systems, to name a few..significant advances have been made toward the development of robust, practical algorithms with theoretical performance guarantees. This paper presents an overview of recent work in this area. In general, rates of convergence depend not only on the number of nodes involved and the desired level of accuracy, but also on the structure and nature of the network over which nodes communicate (e.g., whether links are directed or undirected, static or time varying). We survey the state-of-theart algorithms and their analyses tailored to these different scenarios, highlighting the role of the network topology.

Original languageEnglish (US)
Pages (from-to)953-976
Number of pages24
JournalProceedings of the IEEE
Volume106
Issue number5
DOIs
StatePublished - May 1 2018

Fingerprint

Topology
Communication
Decentralized control
Set theory
Sensor networks

Keywords

  • Consensus algorithms
  • distributed averaging
  • distributed optimization
  • gossip algorithms
  • multiagent systems

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Network Topology and Communication-Computation Tradeoffs in Decentralized Optimization. / Nedich, Angelia; Olshevsky, Alex; Rabbat, Michael G.

In: Proceedings of the IEEE, Vol. 106, No. 5, 01.05.2018, p. 953-976.

Research output: Contribution to journalReview article

Nedich, Angelia ; Olshevsky, Alex ; Rabbat, Michael G. / Network Topology and Communication-Computation Tradeoffs in Decentralized Optimization. In: Proceedings of the IEEE. 2018 ; Vol. 106, No. 5. pp. 953-976.
@article{0eb01ce7bb7d4a4e8de2dd4dff89b652,
title = "Network Topology and Communication-Computation Tradeoffs in Decentralized Optimization",
abstract = "In decentralized optimization, nodes cooperate to minimize an overall objective function that is the sum (or average) of per-node private objective functions. Algorithms interleave local computations with communication among all or a subset of the nodes. Motivated by a variety of applications..decentralized estimation in sensor networks, fitting models to massive data sets, and decentralized control of multirobot systems, to name a few..significant advances have been made toward the development of robust, practical algorithms with theoretical performance guarantees. This paper presents an overview of recent work in this area. In general, rates of convergence depend not only on the number of nodes involved and the desired level of accuracy, but also on the structure and nature of the network over which nodes communicate (e.g., whether links are directed or undirected, static or time varying). We survey the state-of-theart algorithms and their analyses tailored to these different scenarios, highlighting the role of the network topology.",
keywords = "Consensus algorithms, distributed averaging, distributed optimization, gossip algorithms, multiagent systems",
author = "Angelia Nedich and Alex Olshevsky and Rabbat, {Michael G.}",
year = "2018",
month = "5",
day = "1",
doi = "10.1109/JPROC.2018.2817461",
language = "English (US)",
volume = "106",
pages = "953--976",
journal = "Proceedings of the IEEE",
issn = "0018-9219",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Network Topology and Communication-Computation Tradeoffs in Decentralized Optimization

AU - Nedich, Angelia

AU - Olshevsky, Alex

AU - Rabbat, Michael G.

PY - 2018/5/1

Y1 - 2018/5/1

N2 - In decentralized optimization, nodes cooperate to minimize an overall objective function that is the sum (or average) of per-node private objective functions. Algorithms interleave local computations with communication among all or a subset of the nodes. Motivated by a variety of applications..decentralized estimation in sensor networks, fitting models to massive data sets, and decentralized control of multirobot systems, to name a few..significant advances have been made toward the development of robust, practical algorithms with theoretical performance guarantees. This paper presents an overview of recent work in this area. In general, rates of convergence depend not only on the number of nodes involved and the desired level of accuracy, but also on the structure and nature of the network over which nodes communicate (e.g., whether links are directed or undirected, static or time varying). We survey the state-of-theart algorithms and their analyses tailored to these different scenarios, highlighting the role of the network topology.

AB - In decentralized optimization, nodes cooperate to minimize an overall objective function that is the sum (or average) of per-node private objective functions. Algorithms interleave local computations with communication among all or a subset of the nodes. Motivated by a variety of applications..decentralized estimation in sensor networks, fitting models to massive data sets, and decentralized control of multirobot systems, to name a few..significant advances have been made toward the development of robust, practical algorithms with theoretical performance guarantees. This paper presents an overview of recent work in this area. In general, rates of convergence depend not only on the number of nodes involved and the desired level of accuracy, but also on the structure and nature of the network over which nodes communicate (e.g., whether links are directed or undirected, static or time varying). We survey the state-of-theart algorithms and their analyses tailored to these different scenarios, highlighting the role of the network topology.

KW - Consensus algorithms

KW - distributed averaging

KW - distributed optimization

KW - gossip algorithms

KW - multiagent systems

UR - http://www.scopus.com/inward/record.url?scp=85045736476&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045736476&partnerID=8YFLogxK

U2 - 10.1109/JPROC.2018.2817461

DO - 10.1109/JPROC.2018.2817461

M3 - Review article

VL - 106

SP - 953

EP - 976

JO - Proceedings of the IEEE

JF - Proceedings of the IEEE

SN - 0018-9219

IS - 5

ER -