Energy-Efficient Scheduling for Real-Time Systems Based on Deep Q-Learning Mode

被引:112
作者
Zhang, Qingchen [1 ,2 ]
Lin, Man [2 ]
Yang, Laurence T. [1 ,2 ]
Chen, Zhikui [3 ]
Li, Peng [3 ]
机构
[1] Univ Elect Sci & Technol China, Sch Elect Engn, Chengdu 611731, Peoples R China
[2] St Francis Xavier Univ, Dept Comp Sci, Antigonish, NS B2G 2W5, Canada
[3] Dalian Univ, Sch Software Technol, Dalian 116620, Peoples R China
来源
IEEE TRANSACTIONS ON SUSTAINABLE COMPUTING | 2019年 / 4卷 / 01期
关键词
Energy consumption; stacked auto-encoder; dynamic voltage and frequency scaling; Q-learning; POWER MANAGEMENT; DESIGN; ALGORITHM;
D O I
10.1109/TSUSC.2017.2743704
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Energy saving is a critical and challenging issue for real-time systems in embedded devices because of their limited energy supply. To reduce the energy consumption, a hybrid dynamic voltage and frequency scaling (DVFS) scheduling based on Q-learning (QL-HDS) was proposed by combining energy-efficient DVFS techniques. However, QL-HDS discretizes the system state parameters with a certain step size, resulting in a poor distinction of the system states. More importantly, it is difficult for QL-HDS to learn a system for various task sets with a Q-table and limited training sets. In this paper, an energy-efficient scheduling scheme based on deep Q-learning model is proposed for periodic tasks in real-time systems (DQL-EES). Specially, a deep Q-learning model is designed by combining a stacked auto-encoder and a Q-learning model. In the deep Q-learning model, the stacked auto-encoder is used to replace the Q-function for learning the Q-value of each DVFS technology for any system state. Furthermore, a training strategy is devised to learn the parameters of the deep Q-learning model based on the experience replay scheme. Finally, the performance of the proposed scheme is evaluated by comparison with QL-HDS on different simulation task sets. Results demonstrated that the proposed algorithm can save average 4.2% energy than QL-HDS.
引用
收藏
页码:132 / 141
页数:10
相关论文
共 36 条
[1]  
[Anonymous], 2017, CONCRETE
[2]   Power-aware scheduling for periodic real-time tasks [J].
Aydin, H ;
Melhem, R ;
Mossé, D ;
Mejía-Alvarez, P .
IEEE TRANSACTIONS ON COMPUTERS, 2004, 53 (05) :584-600
[3]   A survey of design techniques for system-level dynamic power management [J].
Benini, L ;
Bogliolo, A ;
De Micheli, G .
IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, 2000, 8 (03) :299-316
[4]   Hybrid power management in real time embedded systems: an interplay of DVFS and DPM techniques [J].
Bhatti, Muhammad Khurram ;
Belleudy, Cecile ;
Auguin, Michel .
REAL-TIME SYSTEMS, 2011, 47 (02) :143-162
[5]   Application-Oriented Sensor Network Architecture for Dependable Structural Health Monitoring [J].
Bhuiyan, Md Zakirul Alam ;
Wang, Guojun ;
Wu, Jie ;
Xing Xiaofei ;
Liu, Xiangyong .
2015 IEEE 21ST PACIFIC RIM INTERNATIONAL SYMPOSIUM ON DEPENDABLE COMPUTING (PRDC), 2015, :91-98
[6]   Distributed Control Scheme for Package-Level State-of-Charge Balancing of Grid-Connected Battery Energy Storage System [J].
Cai, He ;
Hu, Guoqiang .
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2016, 12 (05) :1919-1929
[7]  
[初佳兰 Chu Jialan], 2015, [海洋科学, Marine Sciences], V39, P72
[8]   Workload Change Point Detection for Runtime Thermal Management of Embedded Systems [J].
Das, Anup ;
Merrett, Geoff V. ;
Tribastone, Mirco ;
Al-Hashimi, Bashir M. .
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS, 2016, 35 (08) :1358-1371
[9]   System-Level Power Management Using Online Learning [J].
Dhiman, Gaurav ;
Rosing, Tajana Simunic .
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS, 2009, 28 (05) :676-689
[10]   Context-Aware Energy Enhancements for Smart Mobile Devices [J].
Donohoo, Brad K. ;
Ohlsen, Chris ;
Pasricha, Sudeep ;
Xiang, Yi ;
Anderson, Charles .
IEEE TRANSACTIONS ON MOBILE COMPUTING, 2014, 13 (08) :1720-1732