An Asynchronous Distributed ADMM Algorithm and Efficient Communication Model

In this paper, we present a hierarchical butterfly communication model, which is applied to an asynchronous distributed ADMM algorithm. The goal is to minimize the communication overhead of the distributed ADMM algorithm in the fully connected network. We give a theoretical analysis of the convergen...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2016 IEEE 14th Intl Conf on Dependable, Autonomic and Secure Computing, 14th Intl Conf on Pervasive Intelligence and Computing, 2nd Intl Conf on Big Data Intelligence and Computing and Cyber Science and Technology Congress(DASC/PiCom/DataCom/CyberSciTech) S. 136 - 140
Hauptverfasser: Fang, Ling, Lei, Yongmei
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.08.2016
Schlagworte:
ADMM
Algorithm design and analysis
Analytical models
asynchronous
butterfly
Computational modeling
Computers
Convergence
Data models
hierarchical communication model
Machine learning algorithms
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper, we present a hierarchical butterfly communication model, which is applied to an asynchronous distributed ADMM algorithm. The goal is to minimize the communication overhead of the distributed ADMM algorithm in the fully connected network. We give a theoretical analysis of the convergence of the algorithm with hierarchical butterfly communication model. Experiments show that hierarchical butterfly communication model does not have a great impact on the convergence of the algorithm with the increase of computing nodes and thus effectively improve the performance and scalability of the algorithm.
DOI:10.1109/DASC-PICom-DataCom-CyberSciTec.2016.41