Scalable network adaptation for cloud-rans

An imitation learning approach

Yifei Shen, Yuanming Shi, Jun Zhang, Khaled Letaief

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Network adaptation is essential for the efficient operation of Cloud-RANs. Unfortunately, it leads to highly intractable mixed-integer nonlinear programming problems. Existing solutions typically rely on convex relaxation, which yield performance gaps that are difficult to quantify. Meanwhile, global optimization algorithms such as branch-and-bound can find optimal solutions but with prohibitive computational complexity. In this paper, to obtain near-optimal solutions at affordable complexity, we propose to approximate the branch-and-bound algorithm via machine learning. Specifically, the pruning procedure in branch-and-bound is formulated as a sequential decision problem, followed by learning the oracle's action via imitation learning. A unique advantage of this framework is that the training process only requires a small dataset, and it is scalable to problem instances with larger dimensions than the training setting. This is achieved by identifying and leveraging the problem-size independent features. Numerical simulations demonstrate that the learning based framework significantly outperforms competing methods, with computational complexity much lower than the traditional branch-and-bound algorithm.

Original languageEnglish
Title of host publication2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages614-618
Number of pages5
ISBN (Electronic)9781728112954
DOIs
Publication statusPublished - 20 Feb 2019
Externally publishedYes
Event2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Anaheim, United States
Duration: 26 Nov 201829 Nov 2018

Publication series

Name2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

Conference

Conference2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018
CountryUnited States
CityAnaheim
Period26/11/1829/11/18

Fingerprint

Computational complexity
Nonlinear programming
Global optimization
Learning systems
Computer simulation

Keywords

  • Branch-and-bound
  • Cloud-RAN
  • Green communications
  • Imitation learning
  • Pruning

ASJC Scopus subject areas

  • Information Systems
  • Signal Processing

Cite this

Shen, Y., Shi, Y., Zhang, J., & Letaief, K. (2019). Scalable network adaptation for cloud-rans: An imitation learning approach. In 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings (pp. 614-618). [8646503] (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/GlobalSIP.2018.8646503

Scalable network adaptation for cloud-rans : An imitation learning approach. / Shen, Yifei; Shi, Yuanming; Zhang, Jun; Letaief, Khaled.

2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 614-618 8646503 (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Shen, Y, Shi, Y, Zhang, J & Letaief, K 2019, Scalable network adaptation for cloud-rans: An imitation learning approach. in 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings., 8646503, 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 614-618, 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018, Anaheim, United States, 26/11/18. https://doi.org/10.1109/GlobalSIP.2018.8646503
Shen Y, Shi Y, Zhang J, Letaief K. Scalable network adaptation for cloud-rans: An imitation learning approach. In 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 614-618. 8646503. (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings). https://doi.org/10.1109/GlobalSIP.2018.8646503
Shen, Yifei ; Shi, Yuanming ; Zhang, Jun ; Letaief, Khaled. / Scalable network adaptation for cloud-rans : An imitation learning approach. 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 614-618 (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings).
@inproceedings{51ae33f88c164f20910b448fd491da1a,
title = "Scalable network adaptation for cloud-rans: An imitation learning approach",
abstract = "Network adaptation is essential for the efficient operation of Cloud-RANs. Unfortunately, it leads to highly intractable mixed-integer nonlinear programming problems. Existing solutions typically rely on convex relaxation, which yield performance gaps that are difficult to quantify. Meanwhile, global optimization algorithms such as branch-and-bound can find optimal solutions but with prohibitive computational complexity. In this paper, to obtain near-optimal solutions at affordable complexity, we propose to approximate the branch-and-bound algorithm via machine learning. Specifically, the pruning procedure in branch-and-bound is formulated as a sequential decision problem, followed by learning the oracle's action via imitation learning. A unique advantage of this framework is that the training process only requires a small dataset, and it is scalable to problem instances with larger dimensions than the training setting. This is achieved by identifying and leveraging the problem-size independent features. Numerical simulations demonstrate that the learning based framework significantly outperforms competing methods, with computational complexity much lower than the traditional branch-and-bound algorithm.",
keywords = "Branch-and-bound, Cloud-RAN, Green communications, Imitation learning, Pruning",
author = "Yifei Shen and Yuanming Shi and Jun Zhang and Khaled Letaief",
year = "2019",
month = "2",
day = "20",
doi = "10.1109/GlobalSIP.2018.8646503",
language = "English",
series = "2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "614--618",
booktitle = "2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings",

}

TY - GEN

T1 - Scalable network adaptation for cloud-rans

T2 - An imitation learning approach

AU - Shen, Yifei

AU - Shi, Yuanming

AU - Zhang, Jun

AU - Letaief, Khaled

PY - 2019/2/20

Y1 - 2019/2/20

N2 - Network adaptation is essential for the efficient operation of Cloud-RANs. Unfortunately, it leads to highly intractable mixed-integer nonlinear programming problems. Existing solutions typically rely on convex relaxation, which yield performance gaps that are difficult to quantify. Meanwhile, global optimization algorithms such as branch-and-bound can find optimal solutions but with prohibitive computational complexity. In this paper, to obtain near-optimal solutions at affordable complexity, we propose to approximate the branch-and-bound algorithm via machine learning. Specifically, the pruning procedure in branch-and-bound is formulated as a sequential decision problem, followed by learning the oracle's action via imitation learning. A unique advantage of this framework is that the training process only requires a small dataset, and it is scalable to problem instances with larger dimensions than the training setting. This is achieved by identifying and leveraging the problem-size independent features. Numerical simulations demonstrate that the learning based framework significantly outperforms competing methods, with computational complexity much lower than the traditional branch-and-bound algorithm.

AB - Network adaptation is essential for the efficient operation of Cloud-RANs. Unfortunately, it leads to highly intractable mixed-integer nonlinear programming problems. Existing solutions typically rely on convex relaxation, which yield performance gaps that are difficult to quantify. Meanwhile, global optimization algorithms such as branch-and-bound can find optimal solutions but with prohibitive computational complexity. In this paper, to obtain near-optimal solutions at affordable complexity, we propose to approximate the branch-and-bound algorithm via machine learning. Specifically, the pruning procedure in branch-and-bound is formulated as a sequential decision problem, followed by learning the oracle's action via imitation learning. A unique advantage of this framework is that the training process only requires a small dataset, and it is scalable to problem instances with larger dimensions than the training setting. This is achieved by identifying and leveraging the problem-size independent features. Numerical simulations demonstrate that the learning based framework significantly outperforms competing methods, with computational complexity much lower than the traditional branch-and-bound algorithm.

KW - Branch-and-bound

KW - Cloud-RAN

KW - Green communications

KW - Imitation learning

KW - Pruning

UR - http://www.scopus.com/inward/record.url?scp=85063101511&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063101511&partnerID=8YFLogxK

U2 - 10.1109/GlobalSIP.2018.8646503

DO - 10.1109/GlobalSIP.2018.8646503

M3 - Conference contribution

T3 - 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

SP - 614

EP - 618

BT - 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -