A multi-setpoint cooling control approach for air-cooled data centers using the deep Q-network algorithm

被引：6
作者：
Chen, Yaohua [1 ]
Guo, Weipeng [2 ]
Liu, Jinwen [3 ]
Shen, Songyu [4 ,6 ]
Lin, Jianpeng [5 ]
Cui, Delong [5 ,7 ]
机构：
[1] Xiamen Customs, Technol Ctr, Xiamen, Peoples R China
[2] Guangzhou Digital Cities Inst Co Ltd, Guangzhou, Peoples R China
[3] Comprehens Technol Serv Ctr Quanzhou Customs, Quanzhou, Peoples R China
[4] Third Res Inst Minist Publ Secur, Shanghai, Peoples R China
[5] Guangdong Univ Petrochem Technol, Maoming, Peoples R China
[6] Affiliated Hangtian Finest Informat Technol Co ltd, Guangzhou, Peoples R China
[7] Guangdong Univ Petrochem Technol, Maoming 525000, Peoples R China
来源：
MEASUREMENT & CONTROL | 2024年 / 57卷 / 06期
基金：
中国国家自然科学基金;
关键词：
Data center; deep reinforcement learning; thermal modeling; cooling control; energy saving; TEMPERATURE PREDICTION; MANAGEMENT;
D O I：
10.1177/00202940231216543
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
Cooling systems provide a safe thermal environment for the reliable operation of IT equipment in data centers (DCs) while generating significant energy consumption. Therefore, to achieve energy savings in cooling system control under dynamic thermal distribution in DCs, this paper proposes a multi-setpoint cooling control approach based on deep reinforcement learning (DRL). Firstly, a thermal model based on the XGBoost algorithm is constructed to precisely evaluate the thermal distribution in the rack room to guide real-time cooling control. Secondly, a multi-set point cooling control approach based on the deep Q-network algorithm (DQN-MSP) is designed to finely regulate the supply air temperature of each air conditioner by capturing the thermal fluctuations to ensure the dynamic balance of cooling supply and demand. Finally, we adopt the extended CloudSimPy simulation tool and the real workload trace of the PlanetLab system to evaluate the effectiveness and performance of the proposed approach. The simulation results show that the proposed control solution effectively reduces the cooling energy consumption by over 2.4% by raising the average air supply temperature of the air conditioner while satisfying the thermal constraints.
引用
收藏
页码：782 / 793
页数：12
相关论文
共 37 条
[11]   Advances in data center thermal management [J].
Fulpagare, Yogesh ;
Bhargav, Atul .
RENEWABLE & SUSTAINABLE ENERGY REVIEWS, 2015, 43 :981-996
[12]   Rapid Temperature Predictions in Data Centers using Multi-Parameter Proper Orthogonal Decomposition [J].
Ghosh, Rajat ;
Joshi, Yogendra .
NUMERICAL HEAT TRANSFER PART A-APPLICATIONS, 2014, 66 (01) :41-63
[13]   A Taxonomy and Future Directions for Sustainable Cloud Computing: 360 Degree View [J].
Gill, Sukhpal Singh ;
Buyya, Rajkumar .
ACM COMPUTING SURVEYS, 2019, 51 (05)
[14]   Thermal Prediction for Efficient Energy Management of Clouds Using Machine Learning [J].
Ilager, Shashikant ;
Ramamohanarao, Kotagiri ;
Buyya, Rajkumar .
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, 2021, 32 (05) :1044-1056
[15]   Exploration in deep reinforcement learning: A survey [J].
Ladosz, Pawel ;
Weng, Lilian ;
Kim, Minwoo ;
Oh, Hyondong .
INFORMATION FUSION, 2022, 85 :1-22
[16]  
Lazic N, 2018, ADV NEUR IN, V31
[17]   Transforming Cooling Optimization for Green Data Center via Deep Reinforcement Learning [J].
Li, Yuanlong ;
Wen, Yonggang ;
Tao, Dacheng ;
Guan, Kyle .
IEEE TRANSACTIONS ON CYBERNETICS, 2020, 50 (05) :2002-2013
[18]  
Lillicrap T. P., 2019, Continuous control with deep reinforcement learning
[19]   Thermal prediction for Air-cooled data center using data Driven-based model [J].
Lin, Jianpeng ;
Lin, Weiwei ;
Lin, Wenjun ;
Wang, Jiangtao ;
Jiang, Hongliang .
APPLIED THERMAL ENGINEERING, 2022, 217
[20]   A Two-Stage Framework for the Multi-User Multi-Data Center Job Scheduling and Resource Allocation [J].
Lin, Jianpeng ;
Cui, Delong ;
Peng, Zhiping ;
Li, Qirui ;
He, Jieguang .
IEEE ACCESS, 2020, 8 (08) :197863-197874
1234