Scholarlink

With the assistance of edge-cloud computing, computation tasks can be migrated among edge-cloud servers to improve the efficiency of task processing. Since traditional heuristic optimization methods usually take multiple iterations, they can not adapt to the dynamic environment. Deep reinforcement learning (DRL) has recently gained significant attention in solving task decision-making problems. The neural network frameworks can extract potential features from input task data in edge-cloud computing. However, the cloud-edge network’s structure information features are ignored. To tackle this issue, we propose an efficient task migration and resource allocation (ETMRA) algorithm based on DRL, and the model takes advantage of graph-based relational inference capability from graph convolutional networks (GCN), which also incorporates the robust exploration and self-evolutionary capabilities of Soft Actor-Critic (SAC). ETMRA could adaptively choose the edge/cloud servers to migrate the task by interacting with the edge-cloud environment. Simulation results demonstrate that compared with three DRL-based methods, our ETMRA can effectively and efficiently explore and discover near-optimal migration decisions with lower total energy consumption of the system, task response time, and task Service Level Agreements violations rate.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A GCN-Based DRL Approach for Task Migration and Resource Allocation in Heterogeneous Edge-Cloud Environments

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yang Wang,
  • Juan Chen,
  • Zongling Wu,
  • Peng Chen,
  • Fengchun Zhang,
  • Junfeng Hao

摘要

With the assistance of edge-cloud computing, computation tasks can be migrated among edge-cloud servers to improve the efficiency of task processing. Since traditional heuristic optimization methods usually take multiple iterations, they can not adapt to the dynamic environment. Deep reinforcement learning (DRL) has recently gained significant attention in solving task decision-making problems. The neural network frameworks can extract potential features from input task data in edge-cloud computing. However, the cloud-edge network’s structure information features are ignored. To tackle this issue, we propose an efficient task migration and resource allocation (ETMRA) algorithm based on DRL, and the model takes advantage of graph-based relational inference capability from graph convolutional networks (GCN), which also incorporates the robust exploration and self-evolutionary capabilities of Soft Actor-Critic (SAC). ETMRA could adaptively choose the edge/cloud servers to migrate the task by interacting with the edge-cloud environment. Simulation results demonstrate that compared with three DRL-based methods, our ETMRA can effectively and efficiently explore and discover near-optimal migration decisions with lower total energy consumption of the system, task response time, and task Service Level Agreements violations rate.