Reinforcement learning approach to implementation of individual controllers in data centre control system

Yulia Berezovskaya; Chen Wei Yang; Valeriy Vyatkin

doi:10.1109/INDIN51773.2022.9976179

Reinforcement learning approach to implementation of individual controllers in data centre control system

Yulia Berezovskaya, Chen Wei Yang, Valeriy Vyatkin

Luleå University of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

Abstract

Contemporary data centres consume electricity on an industrial scale and require control to improve energy efficiency and maintain high availability. The article proposes an idea and structure of the framework supporting development and validation of the multi-agent control for the energy-efficient data centre. The framework comprises two subsystems: the modelling toolbox and the controlling toolbox. This work focuses on such essential components of the controlling toolbox, as an individual controller. The reinforcement learning approach is applied to the controllers' implementation. The server fan controller, named SF agent, is implemented based on the framework infrastructure and reinforcement learning approach. The agent's capability of energy-saving is demonstrated.

Original language	English
Title of host publication	2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022
Publisher	IEEE
Pages	41-46
Number of pages	6
ISBN (Electronic)	978-1-7281-7568-3
DOIs	https://doi.org/10.1109/INDIN51773.2022.9976179
Publication status	Published - 2022
MoE publication type	A4 Conference publication
Event	IEEE International Conference on Industrial Informatics - Perth, Australia Duration: 25 Jul 2022 → 28 Jul 2022

Conference

Conference	IEEE International Conference on Industrial Informatics
Abbreviated title	INDIN
Country/Territory	Australia
City	Perth
Period	25/07/2022 → 28/07/2022

Keywords

data centre
energy-efficient control
modelling
multi-agent control
reinforcement learning

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/INDIN51773.2022.9976179

Cite this

@inproceedings{a61046343bc046d0b228e1952a35429d,

title = "Reinforcement learning approach to implementation of individual controllers in data centre control system",

abstract = "Contemporary data centres consume electricity on an industrial scale and require control to improve energy efficiency and maintain high availability. The article proposes an idea and structure of the framework supporting development and validation of the multi-agent control for the energy-efficient data centre. The framework comprises two subsystems: the modelling toolbox and the controlling toolbox. This work focuses on such essential components of the controlling toolbox, as an individual controller. The reinforcement learning approach is applied to the controllers' implementation. The server fan controller, named SF agent, is implemented based on the framework infrastructure and reinforcement learning approach. The agent's capability of energy-saving is demonstrated.",

keywords = "data centre, energy-efficient control, modelling, multi-agent control, reinforcement learning",

author = "Yulia Berezovskaya and Yang, {Chen Wei} and Valeriy Vyatkin",

note = "Funding Information: ACKNOWLEDGMENT This study was supported by the Swedish Energy Agency (Energimyndigheten) through the Cloudberry Datacentres project under the grant 43090-2 Publisher Copyright: {\textcopyright} 2022 IEEE.; IEEE International Conference on Industrial Informatics, INDIN ; Conference date: 25-07-2022 Through 28-07-2022",

year = "2022",

doi = "10.1109/INDIN51773.2022.9976179",

language = "English",

pages = "41--46",

booktitle = "2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022",

publisher = "IEEE",

address = "United States",

}

Berezovskaya, Y, Yang, CW & Vyatkin, V 2022, Reinforcement learning approach to implementation of individual controllers in data centre control system. in 2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022. IEEE, pp. 41-46, IEEE International Conference on Industrial Informatics, Perth, Australia, 25/07/2022. https://doi.org/10.1109/INDIN51773.2022.9976179

Reinforcement learning approach to implementation of individual controllers in data centre control system. / Berezovskaya, Yulia; Yang, Chen Wei; Vyatkin, Valeriy.
2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022. IEEE, 2022. p. 41-46.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Reinforcement learning approach to implementation of individual controllers in data centre control system

AU - Berezovskaya, Yulia

AU - Yang, Chen Wei

AU - Vyatkin, Valeriy

N1 - Funding Information: ACKNOWLEDGMENT This study was supported by the Swedish Energy Agency (Energimyndigheten) through the Cloudberry Datacentres project under the grant 43090-2 Publisher Copyright: © 2022 IEEE.

PY - 2022

Y1 - 2022

N2 - Contemporary data centres consume electricity on an industrial scale and require control to improve energy efficiency and maintain high availability. The article proposes an idea and structure of the framework supporting development and validation of the multi-agent control for the energy-efficient data centre. The framework comprises two subsystems: the modelling toolbox and the controlling toolbox. This work focuses on such essential components of the controlling toolbox, as an individual controller. The reinforcement learning approach is applied to the controllers' implementation. The server fan controller, named SF agent, is implemented based on the framework infrastructure and reinforcement learning approach. The agent's capability of energy-saving is demonstrated.

AB - Contemporary data centres consume electricity on an industrial scale and require control to improve energy efficiency and maintain high availability. The article proposes an idea and structure of the framework supporting development and validation of the multi-agent control for the energy-efficient data centre. The framework comprises two subsystems: the modelling toolbox and the controlling toolbox. This work focuses on such essential components of the controlling toolbox, as an individual controller. The reinforcement learning approach is applied to the controllers' implementation. The server fan controller, named SF agent, is implemented based on the framework infrastructure and reinforcement learning approach. The agent's capability of energy-saving is demonstrated.

KW - data centre

KW - energy-efficient control

KW - modelling

KW - multi-agent control

KW - reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85145768731&partnerID=8YFLogxK

U2 - 10.1109/INDIN51773.2022.9976179

DO - 10.1109/INDIN51773.2022.9976179

M3 - Conference article in proceedings

AN - SCOPUS:85145768731

SP - 41

EP - 46

BT - 2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022

PB - IEEE

T2 - IEEE International Conference on Industrial Informatics

Y2 - 25 July 2022 through 28 July 2022

ER -

Reinforcement learning approach to implementation of individual controllers in data centre control system

Abstract

Conference

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this