Scholarlink

<p>Multi-access edge computing (MEC) addresses the constraints of limited computing resources and battery-limited devices by deploying computing resources at the network edge. However, most existing works on task offloading focus on single optimization objectives or simple tasks, neglecting the complexity of task types and diverse requirements in real-world applications. To solve this problem, this paper investigates the multi-objective offloading problem for multi-workflow tasks in heterogeneous Internet of Things (IoT) scenarios. The aforementioned problem is formulated as a multi-objective optimization problem with two objectives: average task completion latency and device energy consumption. We propose an improved decomposition-based multi-objective evolutionary algorithm (MOEA/D) incorporating two novel strategies: (1) population initialization based on prior knowledge, and (2) a population distribution-based weight adjustment scheme. Simulation results demonstrate that our algorithm optimally balances delay and energy consumption requirements compared to existing methods, while enhancing the diversity and convergence of non-dominated solutions.</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

Multi-objective multi-workflow task offloading based on evolutionary optimization

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年8月21日

查看全文
  • Teliekebieke Misha,
  • Lisheng Sun,
  • Zheng-yi Chai

摘要

Multi-access edge computing (MEC) addresses the constraints of limited computing resources and battery-limited devices by deploying computing resources at the network edge. However, most existing works on task offloading focus on single optimization objectives or simple tasks, neglecting the complexity of task types and diverse requirements in real-world applications. To solve this problem, this paper investigates the multi-objective offloading problem for multi-workflow tasks in heterogeneous Internet of Things (IoT) scenarios. The aforementioned problem is formulated as a multi-objective optimization problem with two objectives: average task completion latency and device energy consumption. We propose an improved decomposition-based multi-objective evolutionary algorithm (MOEA/D) incorporating two novel strategies: (1) population initialization based on prior knowledge, and (2) a population distribution-based weight adjustment scheme. Simulation results demonstrate that our algorithm optimally balances delay and energy consumption requirements compared to existing methods, while enhancing the diversity and convergence of non-dominated solutions.