RL-MEOPS: Reinforcement Learning-Driven Memory Energy Optimisation via Hybrid Slack-Aware Control

Key Points

• The central aim is to optimize memory energy consumption using reinforcement learning techniques.
• Implemented a reinforcement learning algorithm for memory management.
• Developed a hybrid slack-aware control system to balance energy and performance.
• Evaluated the system in various computational scenarios.
• Achieved significant energy savings compared to traditional memory management techniques.
• Improved overall system efficiency and performance metrics.
• Demonstrated adaptability of the approach across different computing environments.