File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: SI-ADMM: A Stochastic Inexact ADMM Framework for Stochastic Convex Programs

TitleSI-ADMM: A Stochastic Inexact ADMM Framework for Stochastic Convex Programs
Authors
KeywordsAlternating direction method of multiplier (ADMM)
Convex optimization
Stochastic approximation
Stochastic optimization
Issue Date2020
Citation
IEEE Transactions on Automatic Control, 2020, v. 65, n. 6, p. 2355-2370 How to Cite?
AbstractWe consider the structured stochastic convex program requiring the minimization of \mathbb {E}_\xi [\tilde{f}(x,\xi)]+\mathbb {E}_\xi [\tilde{g}(y,\xi)] subject to the constraint Ax + By = b. Motivated by the need for decentralized schemes, we propose a stochastic inexact alternating direction method of multiplier (SI-ADMM) framework where subproblems are solved inexactly via stochastic approximation schemes. we propose a stochastic inexact alternating direction method of multiplier (SI-ADMM) framework where subproblems are solved inexactly via stochastic approximation schemes. Based on this framework, we prove the following: 1) under suitable assumptions on the associated batch-size of samples utilized at each iteration, the SI-ADMM scheme produces a sequence that converges to the unique solution almost surely; 2) if the number of gradient steps (or equivalently, the number of sampled gradients) utilized for solving the subproblems in each iteration increases at a geometric rate, the mean-squared error diminishes to zero at a prescribed geometric rate; and 3) the overall iteration complexity in terms of gradient steps (or equivalently samples) is found to be consistent with the canonical level of \mathcal {O}(1/\epsilon). Preliminary applications on LASSO and distributed regression suggest that the scheme performs well compared to its competitors.
Persistent Identifierhttp://hdl.handle.net/10722/309261
ISSN
2023 Impact Factor: 6.2
2023 SCImago Journal Rankings: 4.501
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorXie, Yue-
dc.contributor.authorShanbhag, Uday V.-
dc.date.accessioned2021-12-15T03:59:51Z-
dc.date.available2021-12-15T03:59:51Z-
dc.date.issued2020-
dc.identifier.citationIEEE Transactions on Automatic Control, 2020, v. 65, n. 6, p. 2355-2370-
dc.identifier.issn0018-9286-
dc.identifier.urihttp://hdl.handle.net/10722/309261-
dc.description.abstractWe consider the structured stochastic convex program requiring the minimization of \mathbb {E}_\xi [\tilde{f}(x,\xi)]+\mathbb {E}_\xi [\tilde{g}(y,\xi)] subject to the constraint Ax + By = b. Motivated by the need for decentralized schemes, we propose a stochastic inexact alternating direction method of multiplier (SI-ADMM) framework where subproblems are solved inexactly via stochastic approximation schemes. we propose a stochastic inexact alternating direction method of multiplier (SI-ADMM) framework where subproblems are solved inexactly via stochastic approximation schemes. Based on this framework, we prove the following: 1) under suitable assumptions on the associated batch-size of samples utilized at each iteration, the SI-ADMM scheme produces a sequence that converges to the unique solution almost surely; 2) if the number of gradient steps (or equivalently, the number of sampled gradients) utilized for solving the subproblems in each iteration increases at a geometric rate, the mean-squared error diminishes to zero at a prescribed geometric rate; and 3) the overall iteration complexity in terms of gradient steps (or equivalently samples) is found to be consistent with the canonical level of \mathcal {O}(1/\epsilon). Preliminary applications on LASSO and distributed regression suggest that the scheme performs well compared to its competitors.-
dc.languageeng-
dc.relation.ispartofIEEE Transactions on Automatic Control-
dc.subjectAlternating direction method of multiplier (ADMM)-
dc.subjectConvex optimization-
dc.subjectStochastic approximation-
dc.subjectStochastic optimization-
dc.titleSI-ADMM: A Stochastic Inexact ADMM Framework for Stochastic Convex Programs-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/TAC.2019.2953209-
dc.identifier.scopuseid_2-s2.0-85086049259-
dc.identifier.volume65-
dc.identifier.issue6-
dc.identifier.spage2355-
dc.identifier.epage2370-
dc.identifier.eissn1558-2523-
dc.identifier.isiWOS:000542947800004-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats