Scholarlink

<p>Industrial Internet of Things (IIoT) is a novel industrial production paradigm that integrates cloud computing, edge computing to meet industrial application requirements and improve manufacturing efficiency. In IIoT, cloud-edge resource scheduling optimizes resource allocation dynamically based on task demands, ensuring efficient processing, low-latency responses, and enhanced system performance while avoiding resource idleness. Existing research on resource scheduling in dynamic edge cloud environments faces difficulties in adapting to workload variations and neglecting the heterogeneity of computing resources, leading to imbalanced resource allocation and limited system performance. We adopt the A3CIW (Asynchronous Advantage Actor-Critic with Importance Weights) algorithm based on reinforcement learning to address the resource scheduling problem of computational tasks based on cloud-edge collaborative in IIoT, to reduce the response time and energy consumption of task processing. Compared to some commonly used scheduling algorithms, the experiment shows that the proposed method improves energy consumption, response time by approximately 6.92%, 13.47%.</p>

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 cloud edge resources scheduling method based on reinforcement learning in industrial internet of things

  • 期刊论文

  • 发表日期：

    2025年10月9日

查看全文
  • Yuzhen Zhang,
  • Xuhui Zhao,
  • Ming Liu,
  • Jianghui Liu,
  • Ruijuan Zheng

摘要

Industrial Internet of Things (IIoT) is a novel industrial production paradigm that integrates cloud computing, edge computing to meet industrial application requirements and improve manufacturing efficiency. In IIoT, cloud-edge resource scheduling optimizes resource allocation dynamically based on task demands, ensuring efficient processing, low-latency responses, and enhanced system performance while avoiding resource idleness. Existing research on resource scheduling in dynamic edge cloud environments faces difficulties in adapting to workload variations and neglecting the heterogeneity of computing resources, leading to imbalanced resource allocation and limited system performance. We adopt the A3CIW (Asynchronous Advantage Actor-Critic with Importance Weights) algorithm based on reinforcement learning to address the resource scheduling problem of computational tasks based on cloud-edge collaborative in IIoT, to reduce the response time and energy consumption of task processing. Compared to some commonly used scheduling algorithms, the experiment shows that the proposed method improves energy consumption, response time by approximately 6.92%, 13.47%.