FiDRL: Flexible Invocation-Based Deep Reinforcement Learning for DVFS Scheduling in Embedded Systems

Keyphrases

• Deep Reinforcement Learning (deep RL) 100%
• Embedded Systems 100%
• Dynamic Voltage Frequency Scaling 100%
• Invocation 100%
• Reinforcement Learning Agent 50%
• Energy Consumption 33%
• Resource Consumption 16%
• Energy Reduction 16%
• Platform Design 16%
• Experiment Results 16%
• Miniaturization 16%
• Cost Reduction 16%
• Hybrid Algorithm 16%
• Overall Energy 16%
• Pruning 16%
• Energy Conservation 16%
• Learning-based 16%
• Embedded Computing 16%
• Overall System 16%
• Temporal Perspective 16%
• Action Space 16%
• Scaling Approach 16%
• Energy Overhead 16%
• Multi-task Scheduling 16%
• Deep Reinforcement Learning Model 16%
• Execution Energy 16%
• Agent Deployment 16%
• System Energy Consumption 16%

Computer Science

• Deep Reinforcement Learning 100%
• Embedded Systems 100%
• Dynamic Voltage 100%
• Energy Consumption 42%
• Learning Agent 42%
• Hybrid Algorithm 14%
• task-scheduling 14%
• Embedded Computing 14%
• Resource Consumption 14%