Scholarlink

The swift advancement of Mobile Edge Computing (MEC) and cloud terminals has amplified the need for effective strategies in task offloading and resource allocation. This research tackles the challenges of task offloading within MEC environments, emphasizing the potential of cloud terminals as a pivotal application scenario. By enhancing task execution performance and efficiency within MEC systems, this study introduces an algorithm leveraging Deep Q Network (DQN) to optimize both task offloading and resource distribution. Utilizing neural network architecture, the DQN algorithm stabilizes the learning process and boosts decision efficiency. The method is designed to dynamically allocate computing resources, facilitating the transfer of tasks from cloud terminals to edge servers and cloud endpoints. Simulation outcomes reveal that the DQN-based approach substantially reduces task completion time and energy consumption, outperforming traditional methods. The integration of cloud terminals in MEC frameworks provides a robust solution for managing the rising computational demands of contemporary applications. The results underscore the potential of advanced learning techniques in enhancing MEC capabilities, ultimately fostering the development of more responsive and energy-efficient mobile networks.

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

Deep Reinforcement Learning-Based Task Offloading Strategy for Cloud Terminal and Edge Computing

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Dong Tian,
  • Yuanshuo Wang,
  • Yiming Dang,
  • Ligang Ren

摘要

The swift advancement of Mobile Edge Computing (MEC) and cloud terminals has amplified the need for effective strategies in task offloading and resource allocation. This research tackles the challenges of task offloading within MEC environments, emphasizing the potential of cloud terminals as a pivotal application scenario. By enhancing task execution performance and efficiency within MEC systems, this study introduces an algorithm leveraging Deep Q Network (DQN) to optimize both task offloading and resource distribution. Utilizing neural network architecture, the DQN algorithm stabilizes the learning process and boosts decision efficiency. The method is designed to dynamically allocate computing resources, facilitating the transfer of tasks from cloud terminals to edge servers and cloud endpoints. Simulation outcomes reveal that the DQN-based approach substantially reduces task completion time and energy consumption, outperforming traditional methods. The integration of cloud terminals in MEC frameworks provides a robust solution for managing the rising computational demands of contemporary applications. The results underscore the potential of advanced learning techniques in enhancing MEC capabilities, ultimately fostering the development of more responsive and energy-efficient mobile networks.