Distributed zeroth-order online optimization with communication delays

Inoue, Keito; Hayashi, Naoki; Takai, Shigemasa

IET Control Theory and Applications, 2025, 19(1), e12759

Number of Access:339（2025-08-14 02:03 Counts）

Identifier to cite or link to this item: https://hdl.handle.net/11094/100341

Link to primary information 

File Format Terms of use Size Views Date.Available Description information
IETControlTheoryAppl_19_1_e12759 pdf None 1.24 MB 20 2025.02.18  
Creative Commons : Attribution - NonCommercial

Item Information

Output File Export EndNote Basic Export Mendeley

Title
Distributed zeroth-order online optimization with communication delays
Creator
CreatorName
Inoue, Keito
NRID
1000080637752 1000080637752
researchmap permalink
zu3rv2h zu3rv2h
Scopus Author ID
30967580300 30967580300
CreatorName
Hayashi, Naoki
NRID
1000060243177 1000060243177
researchmap permalink
read0014096 read0014096
Scopus Author ID
7201470914 7201470914
CreatorName
Takai, Shigemasa
Subject
bandit feedback
communication delays
distributed online optimization
multiagent system
Description
Inoue K., Hayashi N., Takai S. Distributed zeroth-order online optimization with communication delays. IET Control Theory and Applications 19, e12759 (2025); https://doi.org/10.1049/cth2.12759.
Abstract
This paper investigates distributed online optimization in a networked multiagent system, where each agent has its own private objective and constraint functions that vary over time. In many real-world scenarios, computing the gradient of the cost function can be challenging, especially when agents have limited computational capabilities. Moreover, communication delays are common in practical networked systems due to various factors. This paper considers a unified framework for distributed online optimization that can handle bandit feedback and communication delays feedback simultaneously. A distributed primal-dual algorithm is proposed that utilizes bandit feedback, in which the agents estimate the gradients of their objective and constraint functions by sampling the function values. An enlarged network model that incorporates the delayed information exchanged among the agents is introduced. Through theoretical analysis, it is shown that the proposed algorithm achieves sublinear upper bounds on both the dynamic regret and the constraint violation despite communication delays.
Publisher
John Wiley and Sons Inc
Source Title
IET Control Theory and Applications
Volume (Issue)
19(1)
Page
e12759
Date of Issued
2025-01-02
Language
English
Handle URL
https://hdl.handle.net/11094/100341
PISSN
17518644
eISSN
17518652
Relation.isIdenticalTo
DOI (Publisher Version)
https://doi.org/10.1049/cth2.12759
Access Rights
open access
CopyRight
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
oaire:version
Version of Record
Category
学術雑誌論文 Journal Article