A data-driven approach to stochastic network optimization

Tianyi Chen; Aryan Mokhtari; Xin Wang; Alejandro Ribeiro; Georgios B. Giannakis

doi:10.1109/GlobalSIP.2016.7905894

A data-driven approach to stochastic network optimization

Tianyi Chen, Aryan Mokhtari, Xin Wang, Alejandro Ribeiro, Georgios B. Giannakis

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Scopus citations

Abstract

This paper considers the long-term network resource allocation problem subject to queue stability. The dynamic problem is first reformulated as a static stochastic programming. To tackle the resultant static programming, we study its dual problem which contains finite number of variables in oppose to the primal problem that has infinite dimension. A novel online framework is developed by formulating the dual stochastic optimization as empirical risk minimization. We first propose an offline scheme for batch training which linearly converges to the optimal dual argument in expectation. The offline approach is further extended to the online setting which successfully converges to the statistical accuracy of the adaptive training set with high probability. It is both theoretically and numerically established that the novel approach can significantly improve delay and convergence of existing network optimization schemes.

Original language	English (US)
Title of host publication	2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	510-514
Number of pages	5
ISBN (Electronic)	9781509045457
DOIs	https://doi.org/10.1109/GlobalSIP.2016.7905894
State	Published - Apr 19 2017
Event	2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Washington, United States Duration: Dec 7 2016 → Dec 9 2016

Publication series

Name	2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings

Other

Other	2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016
Country/Territory	United States
City	Washington
Period	12/7/16 → 12/9/16

Bibliographical note

Publisher Copyright:
© 2016 IEEE.

Keywords

Data-driven
Machine learning
Network resource allocation
Statistical learning
Stochastic optimization

Publisher link

10.1109/GlobalSIP.2016.7905894

Find It

Find It at U of M Libraries

Cite this

Chen, T., Mokhtari, A., Wang, X., Ribeiro, A., & Giannakis, G. B. (2017). A data-driven approach to stochastic network optimization. In 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings (pp. 510-514). Article 7905894 (2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/GlobalSIP.2016.7905894

A data-driven approach to stochastic network optimization. / Chen, Tianyi; Mokhtari, Aryan; Wang, Xin et al.
2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. p. 510-514 7905894 (2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Chen, T, Mokhtari, A, Wang, X, Ribeiro, A & Giannakis, GB 2017, A data-driven approach to stochastic network optimization. in 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings., 7905894, 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 510-514, 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016, Washington, United States, 12/7/16. https://doi.org/10.1109/GlobalSIP.2016.7905894

Chen T, Mokhtari A, Wang X, Ribeiro A, Giannakis GB. A data-driven approach to stochastic network optimization. In 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2017. p. 510-514. 7905894. (2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings). doi: 10.1109/GlobalSIP.2016.7905894

Chen, Tianyi ; Mokhtari, Aryan ; Wang, Xin et al. / A data-driven approach to stochastic network optimization. 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 510-514 (2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings).

@inproceedings{6aa5fc40f3b04076a6dfb6aed397ce39,

title = "A data-driven approach to stochastic network optimization",

abstract = "This paper considers the long-term network resource allocation problem subject to queue stability. The dynamic problem is first reformulated as a static stochastic programming. To tackle the resultant static programming, we study its dual problem which contains finite number of variables in oppose to the primal problem that has infinite dimension. A novel online framework is developed by formulating the dual stochastic optimization as empirical risk minimization. We first propose an offline scheme for batch training which linearly converges to the optimal dual argument in expectation. The offline approach is further extended to the online setting which successfully converges to the statistical accuracy of the adaptive training set with high probability. It is both theoretically and numerically established that the novel approach can significantly improve delay and convergence of existing network optimization schemes.",

keywords = "Data-driven, Machine learning, Network resource allocation, Statistical learning, Stochastic optimization",

author = "Tianyi Chen and Aryan Mokhtari and Xin Wang and Alejandro Ribeiro and Giannakis, {Georgios B.}",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 ; Conference date: 07-12-2016 Through 09-12-2016",

year = "2017",

month = apr,

day = "19",

doi = "10.1109/GlobalSIP.2016.7905894",

language = "English (US)",

series = "2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "510--514",

booktitle = "2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings",

}

TY - GEN

T1 - A data-driven approach to stochastic network optimization

AU - Chen, Tianyi

AU - Mokhtari, Aryan

AU - Wang, Xin

AU - Ribeiro, Alejandro

AU - Giannakis, Georgios B.

PY - 2017/4/19

Y1 - 2017/4/19

N2 - This paper considers the long-term network resource allocation problem subject to queue stability. The dynamic problem is first reformulated as a static stochastic programming. To tackle the resultant static programming, we study its dual problem which contains finite number of variables in oppose to the primal problem that has infinite dimension. A novel online framework is developed by formulating the dual stochastic optimization as empirical risk minimization. We first propose an offline scheme for batch training which linearly converges to the optimal dual argument in expectation. The offline approach is further extended to the online setting which successfully converges to the statistical accuracy of the adaptive training set with high probability. It is both theoretically and numerically established that the novel approach can significantly improve delay and convergence of existing network optimization schemes.

AB - This paper considers the long-term network resource allocation problem subject to queue stability. The dynamic problem is first reformulated as a static stochastic programming. To tackle the resultant static programming, we study its dual problem which contains finite number of variables in oppose to the primal problem that has infinite dimension. A novel online framework is developed by formulating the dual stochastic optimization as empirical risk minimization. We first propose an offline scheme for batch training which linearly converges to the optimal dual argument in expectation. The offline approach is further extended to the online setting which successfully converges to the statistical accuracy of the adaptive training set with high probability. It is both theoretically and numerically established that the novel approach can significantly improve delay and convergence of existing network optimization schemes.

KW - Data-driven

KW - Machine learning

KW - Network resource allocation

KW - Statistical learning

KW - Stochastic optimization

UR - http://www.scopus.com/inward/record.url?scp=85019071642&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85019071642&partnerID=8YFLogxK

U2 - 10.1109/GlobalSIP.2016.7905894

DO - 10.1109/GlobalSIP.2016.7905894

M3 - Conference contribution

AN - SCOPUS:85019071642

T3 - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings

SP - 510

EP - 514

BT - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016

Y2 - 7 December 2016 through 9 December 2016

ER -

A data-driven approach to stochastic network optimization

Abstract

Publication series

Other

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this