Scholarlink

<p>To address the challenges of local congestion and imbalanced resource allocation in UAV-assisted air-ground integrated networks, this study constructs a three-layer network architecture comprising ground edge servers, UAVs, and users. A two-stage optimization framework is proposed, incorporating the node selection and task offloading strategies. First, to mitigate the computational resource limitations arising from the dynamic distribution of multiple users, an evolutionary game-based node selection algorithm is developed to achieve effective dynamic load balancing across offloading nodes. Subsequently, the joint task offloading and power allocation problem is formulated as a Markov decision process, and a reinforcement learning algorithm based on MATD3 is designed to attain the joint optimal control of user offloading ratios and transmission power levels. Simulation results demonstrate that the proposed framework reduces total delay and energy consumption by approximately 33.4% and 29.4%, respectively, outperforming the existing strategies. The framework demonstrates superior scalability and energy efficiency in task-intensive scenarios.</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

Joint node selection and task offloading via evolutionary game and MATD3 in UAV-assisted MEC networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月24日

查看全文
  • Zheng Yao,
  • Puqing Chang,
  • Fahad Khan Khalil,
  • Changhao Duan

摘要

To address the challenges of local congestion and imbalanced resource allocation in UAV-assisted air-ground integrated networks, this study constructs a three-layer network architecture comprising ground edge servers, UAVs, and users. A two-stage optimization framework is proposed, incorporating the node selection and task offloading strategies. First, to mitigate the computational resource limitations arising from the dynamic distribution of multiple users, an evolutionary game-based node selection algorithm is developed to achieve effective dynamic load balancing across offloading nodes. Subsequently, the joint task offloading and power allocation problem is formulated as a Markov decision process, and a reinforcement learning algorithm based on MATD3 is designed to attain the joint optimal control of user offloading ratios and transmission power levels. Simulation results demonstrate that the proposed framework reduces total delay and energy consumption by approximately 33.4% and 29.4%, respectively, outperforming the existing strategies. The framework demonstrates superior scalability and energy efficiency in task-intensive scenarios.