Lore: A Learning-based Approach forWorkflow Scheduling in Clouds

Haosong Peng; Chuge Wu; Yufeng Zhan; Yuanqing Xia

doi:10.1145/3538641.3561487

Lore: A Learning-based Approach forWorkflow Scheduling in Clouds

Haosong Peng, Chuge Wu, Yufeng Zhan, Yuanqing Xia

School of Automation

Beijing Institute of Technology

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

5 Citations (Scopus)

Abstract

The workflow scheduling problem is a critical challenge in clouds. Meticulously designed heuristics have attempted to address the intricate decision problem at a high cost. A more general approach is expected to handle different types of workflows and resource configurations. In this paper, a deep reinforcement Learning based apprOach for woRkflow schEduling (Lore) in clouds has been proposed to minimize the completion time of workflows. Moreover, Monte Carlo Tree Search and graph convolutional network are applied to improve performance further. Experimental results show that Lore outperforms the baselines, reducing average makespan by 2 - 10%, and enabling resource utilization increase by up to 20%.

Original language	English
Title of host publication	Proceedings of the 2022 Research in Adaptive and Convergent Systems, RACS 2022
Publisher	Association for Computing Machinery
Pages	47-52
Number of pages	6
ISBN (Electronic)	9781450393980
DOIs	https://doi.org/10.1145/3538641.3561487
Publication status	Published - 3 Oct 2022
Event	2022 Conference on Research in Adaptive and Convergent Systems, RACS 2022 - Virtual, Online, Japan Duration: 3 Oct 2022 → 6 Oct 2022

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	2022 Conference on Research in Adaptive and Convergent Systems, RACS 2022
Country/Territory	Japan
City	Virtual, Online
Period	3/10/22 → 6/10/22

Keywords

deep reinforcement learning
graph convolutional network
monte carlo tree search
workflow scheduling

Access to Document

10.1145/3538641.3561487

Cite this

@inproceedings{709f459da11543aab6effe2148350111,

title = "Lore: A Learning-based Approach forWorkflow Scheduling in Clouds",

abstract = "The workflow scheduling problem is a critical challenge in clouds. Meticulously designed heuristics have attempted to address the intricate decision problem at a high cost. A more general approach is expected to handle different types of workflows and resource configurations. In this paper, a deep reinforcement Learning based apprOach for woRkflow schEduling (Lore) in clouds has been proposed to minimize the completion time of workflows. Moreover, Monte Carlo Tree Search and graph convolutional network are applied to improve performance further. Experimental results show that Lore outperforms the baselines, reducing average makespan by 2 - 10%, and enabling resource utilization increase by up to 20%.",

keywords = "deep reinforcement learning, graph convolutional network, monte carlo tree search, workflow scheduling",

author = "Haosong Peng and Chuge Wu and Yufeng Zhan and Yuanqing Xia",

note = "Publisher Copyright: {\textcopyright} 2022 ACM.; 2022 Conference on Research in Adaptive and Convergent Systems, RACS 2022 ; Conference date: 03-10-2022 Through 06-10-2022",

year = "2022",

month = oct,

day = "3",

doi = "10.1145/3538641.3561487",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

pages = "47--52",

booktitle = "Proceedings of the 2022 Research in Adaptive and Convergent Systems, RACS 2022",

}

Peng, H, Wu, C , Zhan, Y & Xia, Y 2022, Lore: A Learning-based Approach forWorkflow Scheduling in Clouds. in Proceedings of the 2022 Research in Adaptive and Convergent Systems, RACS 2022. ACM International Conference Proceeding Series, Association for Computing Machinery, pp. 47-52, 2022 Conference on Research in Adaptive and Convergent Systems, RACS 2022, Virtual, Online, Japan, 3/10/22. https://doi.org/10.1145/3538641.3561487

Lore: A Learning-based Approach forWorkflow Scheduling in Clouds. / Peng, Haosong; Wu, Chuge ; Zhan, Yufeng et al.
Proceedings of the 2022 Research in Adaptive and Convergent Systems, RACS 2022. Association for Computing Machinery, 2022. p. 47-52 (ACM International Conference Proceeding Series).

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

TY - GEN

T1 - Lore

T2 - 2022 Conference on Research in Adaptive and Convergent Systems, RACS 2022

AU - Peng, Haosong

AU - Wu, Chuge

AU - Zhan, Yufeng

AU - Xia, Yuanqing

PY - 2022/10/3

Y1 - 2022/10/3

N2 - The workflow scheduling problem is a critical challenge in clouds. Meticulously designed heuristics have attempted to address the intricate decision problem at a high cost. A more general approach is expected to handle different types of workflows and resource configurations. In this paper, a deep reinforcement Learning based apprOach for woRkflow schEduling (Lore) in clouds has been proposed to minimize the completion time of workflows. Moreover, Monte Carlo Tree Search and graph convolutional network are applied to improve performance further. Experimental results show that Lore outperforms the baselines, reducing average makespan by 2 - 10%, and enabling resource utilization increase by up to 20%.

AB - The workflow scheduling problem is a critical challenge in clouds. Meticulously designed heuristics have attempted to address the intricate decision problem at a high cost. A more general approach is expected to handle different types of workflows and resource configurations. In this paper, a deep reinforcement Learning based apprOach for woRkflow schEduling (Lore) in clouds has been proposed to minimize the completion time of workflows. Moreover, Monte Carlo Tree Search and graph convolutional network are applied to improve performance further. Experimental results show that Lore outperforms the baselines, reducing average makespan by 2 - 10%, and enabling resource utilization increase by up to 20%.

KW - deep reinforcement learning

KW - graph convolutional network

KW - monte carlo tree search

KW - workflow scheduling

UR - http://www.scopus.com/inward/record.url?scp=85141103309&partnerID=8YFLogxK

U2 - 10.1145/3538641.3561487

DO - 10.1145/3538641.3561487

M3 - Conference contribution

AN - SCOPUS:85141103309

T3 - ACM International Conference Proceeding Series

SP - 47

EP - 52

BT - Proceedings of the 2022 Research in Adaptive and Convergent Systems, RACS 2022

PB - Association for Computing Machinery

Y2 - 3 October 2022 through 6 October 2022

ER -

Lore: A Learning-based Approach forWorkflow Scheduling in Clouds

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this