• Home
  • Search
  • Research on Thermal Management Algorithm for Chips with Few-Shot Deep Reinforcement Learning Integrating Graph Neural Networks
  • https://doi.org/10.1109/iicspi66775.2025.11438001Copy DOI Icon

Research on Thermal Management Algorithm for Chips with Few-Shot Deep Reinforcement Learning Integrating Graph Neural Networks

  • Dec 26, 2025
  • Jing Liu +1 more
Show More
  • Abstract
  • Literature Map
  • References
  • Similar Papers
Abstract

To address the shortcomings of traditional thermal management algorithms, such as insufficient accuracy under dynamic loads, poor generalization of DRL algorithms with few shots, and superficiality in the fusion of GNN and DRL, this paper proposes a graph attention-enhanced meta-learning deep reinforcement learning algorithm (GAM-SSDRL). The algorithm constructs a three-level closed-loop architecture of “thermal field graph modeling - few-shot decision-making dynamic collaborative optimization.” It accurately captures spatial correlations of the thermal field through a graph attention network, achieves efficient few-shot training based on an improved MAML framework, and adaptively matches feature extraction and decision execution through a dynamic collaborative strategy. Experiments based on a 16 -core CPU chip simulation model show that GAM-SSDRL performs optimally under high, medium, and low load scenarios, with an average temperature deviation as low as 1.6° C, representing reductions of 50.0 %, 23.8 %, and 15. 8 % compared to traditional PID, conventional DRL, and GNN+DRL, respectively. The thermal runaway response time is 44.3 ms, a reduction of over 50.6 %; the average chip power consumption is 78.9 W, a reduction of 18.3 %; and only 100 training samples are needed to achieve the target accuracy, improving sample efficiency by over 54.5 %. This algorithm fills the technical gap in precise control of dynamic thermal fields with small sample sizes, providing a new solution for thermal management of high-power-density chips, and has significant theoretical and engineering value.

Similar Papers
  • Research Article
  • Citations57

A new ensemble deep graph reinforcement learning network for spatio-temporal traffic volume forecasting in a freeway network

  • Jan 29, 2022
  • Digital Signal Processing
  • Pan Shang +5
  • PDF
  • Research Article
  • Citations5

Deep imitation reinforcement learning with expert demonstration data

  • Oct 31, 2018
  • The Journal of Engineering
  • Menglong Yi +3
  • Research Article
  • Citations22

Physical-assisted multi-agent graph reinforcement learning enabled fast voltage regulation for PV-rich active distribution network

  • Aug 31, 2023
  • Applied Energy
  • Yongdong Chen +5
  • PDF
  • Research Article
  • Citations22

A Review of Mobile Robot Path Planning Based on Deep Reinforcement Learning Algorithm

  • Dec 01, 2021
  • Journal of Physics: Conference Series
  • Yanwei Zhao +2
  • Research Article

Research on Dependency-Aware Service Migration Strategy in the Internet of Vehicles Integrating a Graph Attention Network and Deep Reinforcement Learning

  • Jan 16, 2026
  • Applied Sciences
  • Ying Liu +2
  • Research Article

Adaptive Routing for Meshed QKD Networks of Flexible Size Using Deep Reinforcement Learning

  • Feb 16, 2026
  • Photonics
  • Tim Johann +2
  • Research Article
  • Citations5

Distributed Routing and Data Scheduling in IPNs With GNN-Based Multiagent DRL

  • Jun 15, 2025
  • IEEE Internet of Things Journal
  • Xixuan Zhou +6
  • Research Article
  • Citations1

DDPG Agent to Swing Up and Balance Cart- Pole System

  • Apr 09, 2021
  • International Journal of Advanced Research in Science, Communication and Technology
  • Buvanesh Pandian V
  • Conference Article

A Deep Reinforcement Learning Framework for Optimizing Data Center Cooling Systems

  • Sep 19, 2025
  • Yang Wei +1
  • Dissertation
  • Citations1

Deep reinforcement learning-based dynamic scheduling

  • Jan 01, 2022
  • Renke Liu
  • Conference Article
  • Citations1

Research on AI-driven personalized learning path planning and effectiveness under dual-system teaching mode

  • Apr 18, 2025
  • Ling Chen
  • Research Article
  • Citations3

Dynamic Reconfiguration Method of Active Distribution Networks Based on Graph Attention Network Reinforcement Learning

  • Apr 17, 2025
  • Energies
  • Chen Guo +2
  • Research Article

A general algorithm for eliminating critical conditions for solving the problem of controlling a real walking robot based on deep reinforcement learning methods

  • Mar 01, 2025
  • Программные системы и вычислительные методы
  • Vasily Vasil'Evich Kashko +1
  • Research Article
  • Citations17

A New Fingerprint and Graph Hybrid Neural Network for Predicting Molecular Properties.

  • Jul 25, 2024
  • Journal of chemical information and modeling
  • Qingtian Zhang +3
  • Research Article

Dynamic aircraft maintenance stand scheduling via graph attention network and deep reinforcement learning

  • Mar 19, 2026
  • Annals of Operations Research
  • Runxia Guo +2
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.