A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs

W. Jiang; T. Charalambous

doi:10.1109/CDC51059.2022.9992389

A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

Abstrakti

In this paper, we study the unconstrained optimization problem in a distributed way over directed strongly connected communication graphs. We propose an algorithm, which combines techniques of both gradient descent (GD) and finite-time exact ratio consensus (FTERC). Different from the techniques of average or dynamic average consensus with asymptotic convergence or techniques of finite-time “approximate” consensus with inexact accuracy in the literature, with the help of FTERC for gradient tracking, our proposed distributed FTERC based GD algorithm has a faster convergence rate related to the optimization iteration number and a larger step-size upper bound compared with other algorithms, as demonstrated in the simulations.

Alkuperäiskieli	Englanti
Otsikko	2022 IEEE 61st Conference on Decision and Control (CDC)
Kustantaja	IEEE
Sivut	6848-6854
Sivumäärä	7
ISBN (elektroninen)	978-1-6654-6761-2
DOI - pysyväislinkit	https://doi.org/10.1109/CDC51059.2022.9992389
Tila	Julkaistu - 2022
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisuussa
Tapahtuma	IEEE Conference on Decision and Control - Cancun, Meksiko Kesto: 6 jouluk. 2022 → 9 jouluk. 2022 Konferenssinumero: 61

Julkaisusarja

Nimi	Proceedings of the IEEE Conference on Decision & Control
ISSN (elektroninen)	2576-2370

Conference

Conference	IEEE Conference on Decision and Control
Lyhennettä	CDC
Maa/Alue	Meksiko
Kaupunki	Cancun
Ajanjakso	06/12/2022 → 09/12/2022

Pääsy asiakirjaan

10.1109/CDC51059.2022.9992389

OpenUrl-saatavuus

Koko teksti

Muut tiedostot ja linkit

Linkki julkaisuun Scopuksessa

Siteeraa tätä

@inproceedings{810a2c75f3a84a3ab7ef658b03173d9e,

title = "A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs",

abstract = "In this paper, we study the unconstrained optimization problem in a distributed way over directed strongly connected communication graphs. We propose an algorithm, which combines techniques of both gradient descent (GD) and finite-time exact ratio consensus (FTERC). Different from the techniques of average or dynamic average consensus with asymptotic convergence or techniques of finite-time “approximate” consensus with inexact accuracy in the literature, with the help of FTERC for gradient tracking, our proposed distributed FTERC based GD algorithm has a faster convergence rate related to the optimization iteration number and a larger step-size upper bound compared with other algorithms, as demonstrated in the simulations.",

keywords = "Gradient methods, Upper bound, Costs, Additives, Heuristic algorithms, Directed graphs, Approximation algorithms, Distributed optimization, gradient tracking, finite-time consensus, directed graphs, gradient descent",

author = "W. Jiang and T. Charalambous",

year = "2022",

doi = "10.1109/CDC51059.2022.9992389",

language = "English",

series = "Proceedings of the IEEE Conference on Decision & Control",

publisher = "IEEE",

pages = "6848--6854",

booktitle = "2022 IEEE 61st Conference on Decision and Control (CDC)",

address = "United States",

note = "IEEE Conference on Decision and Control, CDC ; Conference date: 06-12-2022 Through 09-12-2022",

}

Jiang, W & Charalambous, T 2022, A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs. julkaisussa 2022 IEEE 61st Conference on Decision and Control (CDC). Proceedings of the IEEE Conference on Decision & Control, IEEE, Sivut 6848-6854, IEEE Conference on Decision and Control, Cancun, Meksiko, 06/12/2022. https://doi.org/10.1109/CDC51059.2022.9992389

A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs. / Jiang, W.; Charalambous, T.

2022 IEEE 61st Conference on Decision and Control (CDC). IEEE, 2022. s. 6848-6854 (Proceedings of the IEEE Conference on Decision & Control).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

TY - GEN

T1 - A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs

AU - Jiang, W.

AU - Charalambous, T.

N1 - Conference code: 61

PY - 2022

Y1 - 2022

N2 - In this paper, we study the unconstrained optimization problem in a distributed way over directed strongly connected communication graphs. We propose an algorithm, which combines techniques of both gradient descent (GD) and finite-time exact ratio consensus (FTERC). Different from the techniques of average or dynamic average consensus with asymptotic convergence or techniques of finite-time “approximate” consensus with inexact accuracy in the literature, with the help of FTERC for gradient tracking, our proposed distributed FTERC based GD algorithm has a faster convergence rate related to the optimization iteration number and a larger step-size upper bound compared with other algorithms, as demonstrated in the simulations.

AB - In this paper, we study the unconstrained optimization problem in a distributed way over directed strongly connected communication graphs. We propose an algorithm, which combines techniques of both gradient descent (GD) and finite-time exact ratio consensus (FTERC). Different from the techniques of average or dynamic average consensus with asymptotic convergence or techniques of finite-time “approximate” consensus with inexact accuracy in the literature, with the help of FTERC for gradient tracking, our proposed distributed FTERC based GD algorithm has a faster convergence rate related to the optimization iteration number and a larger step-size upper bound compared with other algorithms, as demonstrated in the simulations.

KW - Gradient methods

KW - Upper bound

KW - Costs

KW - Additives

KW - Heuristic algorithms

KW - Directed graphs

KW - Approximation algorithms

KW - Distributed optimization

KW - gradient tracking

KW - finite-time consensus

KW - directed graphs

KW - gradient descent

UR - http://www.scopus.com/inward/record.url?scp=85143406802&partnerID=8YFLogxK

U2 - 10.1109/CDC51059.2022.9992389

DO - 10.1109/CDC51059.2022.9992389

M3 - Conference contribution

T3 - Proceedings of the IEEE Conference on Decision & Control

SP - 6848

EP - 6854

BT - 2022 IEEE 61st Conference on Decision and Control (CDC)

PB - IEEE

T2 - IEEE Conference on Decision and Control

Y2 - 6 December 2022 through 9 December 2022

ER -

A Fast Finite-Time Consensus based Gradient Method for Distributed Optimization over Digraphs

Abstrakti

Julkaisusarja

Conference

Pääsy asiakirjaan

OpenUrl-saatavuus

Muut tiedostot ja linkit

Sormenjälki

Siteeraa tätä