Scholarlink

<p>The Cloud-Edge-End (C-E-E) paradigm presents challenges in scheduling dependent tasks due to diverse resource instances, complex network topologies, and task interdependencies. This study proposes a novel multi-level task queue scheduling (MLTQS) strategy to address the understudied issue of scheduling tasks with parallel dependencies for synchronous execution. MLTQS comprises parallel and sequential scheduling stages, optimizing combinations of parallel tasks and resources. We introduce a mutant Grey Wolf Optimizer with Levy flight and Random walk (GWOLR) to solve the two-stage problem, employing fusion encoding and partial mutation to enhance population diversity. A new hunting strategy integrating Levy flight and Random walk balances global exploration and local exploitation. Simulation results demonstrate the efficacy of our approach, with the enhanced algorithm showing optimal performance in C-E-E task scheduling scenarios, contributing to efficient resource allocation in complex distributed computing environments.</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

Scheduling dependent-tasks over cloud-edge-end: a mutant Grey Wolf optimization approach with levy flight and random walk

  • 期刊论文

  • 发表日期：

    2025年9月15日

查看全文
  • Junpeng Cai,
  • Yingbo Wu

摘要

The Cloud-Edge-End (C-E-E) paradigm presents challenges in scheduling dependent tasks due to diverse resource instances, complex network topologies, and task interdependencies. This study proposes a novel multi-level task queue scheduling (MLTQS) strategy to address the understudied issue of scheduling tasks with parallel dependencies for synchronous execution. MLTQS comprises parallel and sequential scheduling stages, optimizing combinations of parallel tasks and resources. We introduce a mutant Grey Wolf Optimizer with Levy flight and Random walk (GWOLR) to solve the two-stage problem, employing fusion encoding and partial mutation to enhance population diversity. A new hunting strategy integrating Levy flight and Random walk balances global exploration and local exploitation. Simulation results demonstrate the efficacy of our approach, with the enhanced algorithm showing optimal performance in C-E-E task scheduling scenarios, contributing to efficient resource allocation in complex distributed computing environments.