A new class of distributed optimization algorithms: Application to regression of distributed data

S. Sundhar Ram, Angelia Nedich, V. V. Veeravalli

Research output: Contribution to journalArticle

60 Citations (Scopus)

Abstract

In a distributed optimization problem, the complete problem information is not available at a single location but is rather distributed among different agents in a multi-agent system. In the problems studied in the literature, each agent has an objective function and the network goal is to minimize the sum of the agents objective functions over a constraint set that is globally known. In this paper, we study a generalization of the above distributed optimization problem. In particular, the network objective is to minimize a function of the sum of the individual objective functions over the constraint set. The outer function and the constraint set are known to all the agents. We discuss an algorithm and prove its convergence, and then discuss extensions to more general and complex distributed optimization problems. We provide a motivation for our algorithms through the example of distributed regression of distributed data.

Original languageEnglish (US)
Pages (from-to)71-88
Number of pages18
JournalOptimization Methods and Software
Volume27
Issue number1
DOIs
StatePublished - Feb 1 2012
Externally publishedYes

Fingerprint

Distributed Optimization
Distributed Algorithms
Optimization Algorithm
Regression
Objective function
Optimization Problem
Minimise
Multi-agent Systems
Multi agent systems
Class

Keywords

  • convex optimization
  • distributed optimization
  • distributed regression

ASJC Scopus subject areas

  • Control and Optimization
  • Software
  • Applied Mathematics

Cite this

A new class of distributed optimization algorithms : Application to regression of distributed data. / Sundhar Ram, S.; Nedich, Angelia; Veeravalli, V. V.

In: Optimization Methods and Software, Vol. 27, No. 1, 01.02.2012, p. 71-88.

Research output: Contribution to journalArticle

@article{4355cc55c44648389ee78445f7130c81,
title = "A new class of distributed optimization algorithms: Application to regression of distributed data",
abstract = "In a distributed optimization problem, the complete problem information is not available at a single location but is rather distributed among different agents in a multi-agent system. In the problems studied in the literature, each agent has an objective function and the network goal is to minimize the sum of the agents objective functions over a constraint set that is globally known. In this paper, we study a generalization of the above distributed optimization problem. In particular, the network objective is to minimize a function of the sum of the individual objective functions over the constraint set. The outer function and the constraint set are known to all the agents. We discuss an algorithm and prove its convergence, and then discuss extensions to more general and complex distributed optimization problems. We provide a motivation for our algorithms through the example of distributed regression of distributed data.",
keywords = "convex optimization, distributed optimization, distributed regression",
author = "{Sundhar Ram}, S. and Angelia Nedich and Veeravalli, {V. V.}",
year = "2012",
month = "2",
day = "1",
doi = "10.1080/10556788.2010.511669",
language = "English (US)",
volume = "27",
pages = "71--88",
journal = "Optimization Methods and Software",
issn = "1055-6788",
publisher = "Taylor and Francis Ltd.",
number = "1",

}

TY - JOUR

T1 - A new class of distributed optimization algorithms

T2 - Application to regression of distributed data

AU - Sundhar Ram, S.

AU - Nedich, Angelia

AU - Veeravalli, V. V.

PY - 2012/2/1

Y1 - 2012/2/1

N2 - In a distributed optimization problem, the complete problem information is not available at a single location but is rather distributed among different agents in a multi-agent system. In the problems studied in the literature, each agent has an objective function and the network goal is to minimize the sum of the agents objective functions over a constraint set that is globally known. In this paper, we study a generalization of the above distributed optimization problem. In particular, the network objective is to minimize a function of the sum of the individual objective functions over the constraint set. The outer function and the constraint set are known to all the agents. We discuss an algorithm and prove its convergence, and then discuss extensions to more general and complex distributed optimization problems. We provide a motivation for our algorithms through the example of distributed regression of distributed data.

AB - In a distributed optimization problem, the complete problem information is not available at a single location but is rather distributed among different agents in a multi-agent system. In the problems studied in the literature, each agent has an objective function and the network goal is to minimize the sum of the agents objective functions over a constraint set that is globally known. In this paper, we study a generalization of the above distributed optimization problem. In particular, the network objective is to minimize a function of the sum of the individual objective functions over the constraint set. The outer function and the constraint set are known to all the agents. We discuss an algorithm and prove its convergence, and then discuss extensions to more general and complex distributed optimization problems. We provide a motivation for our algorithms through the example of distributed regression of distributed data.

KW - convex optimization

KW - distributed optimization

KW - distributed regression

UR - http://www.scopus.com/inward/record.url?scp=84855928432&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84855928432&partnerID=8YFLogxK

U2 - 10.1080/10556788.2010.511669

DO - 10.1080/10556788.2010.511669

M3 - Article

AN - SCOPUS:84855928432

VL - 27

SP - 71

EP - 88

JO - Optimization Methods and Software

JF - Optimization Methods and Software

SN - 1055-6788

IS - 1

ER -