CARTAD: Compiler-Assisted Reinforcement Learning for Thermal-Aware Task Scheduling and DVFS on Multicores.
Di LiuShi-Gui YangZhenli HeMingxiong ZhaoWeichen LiuPublished in: IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. (2022)
Keyphrases
- reinforcement learning
- function approximation
- infrared
- optimal policy
- reinforcement learning algorithms
- programming language
- power plant
- resource management
- temporal difference
- general purpose
- model free
- learning process
- robotic control
- finite element analysis
- learning algorithm
- parallel algorithm
- markov decision processes
- multi agent
- real robot
- energy aware
- scheduling algorithm
- state space
- supervised learning
- software systems
- computing systems
- dynamic programming
- transfer learning
- thermal images
- highly optimized
- transition model
- power reduction
- grid computing
- distributed memory machines