Distributed Optimization with Gradient Tracking Over Heterogeneous Delay-Prone Directed Networks

Evagoras Makridis, Gabriele Oliva, Kasagatta Ramesh Narahari, Mohammadreza Doostmohammadian, Usman A. Khan, Themistoklis Charalambous

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

12 Downloads (Pure)

Abstract

In this paper, we address the distributed optimization problem over unidirectional networks with possibly time-invariant heterogeneous bounded transmission delays. In particular, we propose a modified version of the Accelerated Distributed Directed OPTimization (ADD-OPT) algorithm, herein called Robustified ADD-OPT (R-ADD-OPT), which is able to solve the distributed optimization problem, even when the communication links suffer from heterogeneous but bounded transmission delays. We show that if the gradient step-size of the R-ADD-OPT algorithm is within a certain range, which also depends on the maximum time delay in the network, then the nodes are guaranteed to converge to the optimal solution of the distributed optimization problem. The range of the gradient step-size that guarantees convergence can be computed a priori based on the maximum time delay in the network.

Original languageEnglish
Title of host publication2024 European Control Conference, ECC 2024
PublisherIEEE
Pages2312-2319
Number of pages8
ISBN (Electronic)978-3-9071-4410-7
DOIs
Publication statusPublished - 2024
MoE publication typeA4 Conference publication
EventEuropean Control Conference - Stockholm, Sweden
Duration: 25 Jun 202428 Jun 2024

Publication series

Name2024 European Control Conference, ECC 2024

Conference

ConferenceEuropean Control Conference
Abbreviated titleECC
Country/TerritorySweden
CityStockholm
Period25/06/202428/06/2024

Keywords

  • bounded delays
  • directed graphs
  • distributed optimization
  • gradient tracking
  • unidirectional networks

Fingerprint

Dive into the research topics of 'Distributed Optimization with Gradient Tracking Over Heterogeneous Delay-Prone Directed Networks'. Together they form a unique fingerprint.

Cite this