Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning
The problem of high power consumption in data centers is becoming more and more prominent. In order to improve the energy efficiency of data centers, cooperatively optimizing the energy of IT systems and cooling systems has become an effective way. In this paper, a model-free deep reinforcement lear...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-04-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/14/8/2071 |
| _version_ | 1797538392562466816 |
|---|---|
| author | Ce Chi Kaixuan Ji Penglei Song Avinab Marahatta Shikui Zhang Fa Zhang Dehui Qiu Zhiyong Liu |
| author_facet | Ce Chi Kaixuan Ji Penglei Song Avinab Marahatta Shikui Zhang Fa Zhang Dehui Qiu Zhiyong Liu |
| author_sort | Ce Chi |
| collection | DOAJ |
| description | The problem of high power consumption in data centers is becoming more and more prominent. In order to improve the energy efficiency of data centers, cooperatively optimizing the energy of IT systems and cooling systems has become an effective way. In this paper, a model-free deep reinforcement learning (DRL)-based joint optimization method MAD3C is developed to overcome the high-dimensional state and action space problems of the data center energy optimization. A hybrid AC-DDPG cooperative multi-agent framework is devised for the improvement of the cooperation between the IT and cooling systems for further energy efficiency improvement. In the framework, a scheduling baseline comparison method is presented to enhance the stability of the framework. Meanwhile, an adaptive score is designed for the architecture in consideration of multi-dimensional resources and resource utilization improvement. Experiments show that our proposed approach can effectively reduce energy for data centers through the cooperative optimization while guaranteeing training stability and improving resource utilization. |
| first_indexed | 2024-03-10T12:29:54Z |
| format | Article |
| id | doaj.art-6519e27ec2e8472bb251276a615c33aa |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-10T12:29:54Z |
| publishDate | 2021-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-6519e27ec2e8472bb251276a615c33aa2023-11-21T14:43:43ZengMDPI AGEnergies1996-10732021-04-01148207110.3390/en14082071Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement LearningCe Chi0Kaixuan Ji1Penglei Song2Avinab Marahatta3Shikui Zhang4Fa Zhang5Dehui Qiu6Zhiyong Liu7High Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100095, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100095, ChinaInformation Engineering College, Capital Normal University, Beijing 100048, ChinaInstitute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, ChinaInformation Engineering College, Capital Normal University, Beijing 100048, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100095, ChinaInformation Engineering College, Capital Normal University, Beijing 100048, ChinaHigh Performance Computer Research Center, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100095, ChinaThe problem of high power consumption in data centers is becoming more and more prominent. In order to improve the energy efficiency of data centers, cooperatively optimizing the energy of IT systems and cooling systems has become an effective way. In this paper, a model-free deep reinforcement learning (DRL)-based joint optimization method MAD3C is developed to overcome the high-dimensional state and action space problems of the data center energy optimization. A hybrid AC-DDPG cooperative multi-agent framework is devised for the improvement of the cooperation between the IT and cooling systems for further energy efficiency improvement. In the framework, a scheduling baseline comparison method is presented to enhance the stability of the framework. Meanwhile, an adaptive score is designed for the architecture in consideration of multi-dimensional resources and resource utilization improvement. Experiments show that our proposed approach can effectively reduce energy for data centers through the cooperative optimization while guaranteeing training stability and improving resource utilization.https://www.mdpi.com/1996-1073/14/8/2071data centerenergy efficiencydeep reinforcement learningmulti-agentscheduling algorithmcooling system |
| spellingShingle | Ce Chi Kaixuan Ji Penglei Song Avinab Marahatta Shikui Zhang Fa Zhang Dehui Qiu Zhiyong Liu Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning Energies data center energy efficiency deep reinforcement learning multi-agent scheduling algorithm cooling system |
| title | Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning |
| title_full | Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning |
| title_fullStr | Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning |
| title_full_unstemmed | Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning |
| title_short | Cooperatively Improving Data Center Energy Efficiency Based on Multi-Agent Deep Reinforcement Learning |
| title_sort | cooperatively improving data center energy efficiency based on multi agent deep reinforcement learning |
| topic | data center energy efficiency deep reinforcement learning multi-agent scheduling algorithm cooling system |
| url | https://www.mdpi.com/1996-1073/14/8/2071 |
| work_keys_str_mv | AT cechi cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT kaixuanji cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT pengleisong cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT avinabmarahatta cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT shikuizhang cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT fazhang cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT dehuiqiu cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning AT zhiyongliu cooperativelyimprovingdatacenterenergyefficiencybasedonmultiagentdeepreinforcementlearning |