Scholarlink

<p>Cloud computing is a rapidly growing technology that offers on-demand services via the Internet. Workflow scheduling, an NP-hard problem in high-performance computing (HPC), presents significant challenges in finding polynomial-time solutions, as most current mechanisms are designed for conventional computing platforms. Specific meta-heuristic algorithms proposed for workflow scheduling often fail to provide a global optimal solution, becoming trapped in local optima. In this article, we propose a multi-objective model-based hybrid algorithm to address this issue. First, the Grey Wolf Optimization (GWO) algorithm is modified to enhance its exploration capability, resulting in the Enhanced GWO (EGWO). Subsequently, a new meta-heuristic algorithm, PSOEGWO, is developed by hybridizing Particle Swarm Optimization (PSO) with EGWO to solve the workflow scheduling problem. The proposed algorithm is evaluated using complex scientific workflow datasets, including Montage, CyberShake, SIPHT, and Inspiral. Simulation results demonstrate that PSOEGWO outperforms state-of-the-art algorithms such as GWO, PSO-GWO, and PSO. The proposed algorithm (PSOEGWO) reduces total execution cost by approximate 29.83%, 3.58%, and 3.63% compared to GWO, PSO, and PSOGWO, respectively. Similarly, the proposed algorithm reduces the total execution time by 20.06%, 6.10%, and 10.67% compared to GWO, PSO, and PSOGWO, respectively.</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

PSOEGWO: An Efficient Workflow Scheduling Algorithm for Clouds

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • Chotu Lal,
  • Harish Sharma,
  • Neeraj Arora

摘要

Cloud computing is a rapidly growing technology that offers on-demand services via the Internet. Workflow scheduling, an NP-hard problem in high-performance computing (HPC), presents significant challenges in finding polynomial-time solutions, as most current mechanisms are designed for conventional computing platforms. Specific meta-heuristic algorithms proposed for workflow scheduling often fail to provide a global optimal solution, becoming trapped in local optima. In this article, we propose a multi-objective model-based hybrid algorithm to address this issue. First, the Grey Wolf Optimization (GWO) algorithm is modified to enhance its exploration capability, resulting in the Enhanced GWO (EGWO). Subsequently, a new meta-heuristic algorithm, PSOEGWO, is developed by hybridizing Particle Swarm Optimization (PSO) with EGWO to solve the workflow scheduling problem. The proposed algorithm is evaluated using complex scientific workflow datasets, including Montage, CyberShake, SIPHT, and Inspiral. Simulation results demonstrate that PSOEGWO outperforms state-of-the-art algorithms such as GWO, PSO-GWO, and PSO. The proposed algorithm (PSOEGWO) reduces total execution cost by approximate 29.83%, 3.58%, and 3.63% compared to GWO, PSO, and PSOGWO, respectively. Similarly, the proposed algorithm reduces the total execution time by 20.06%, 6.10%, and 10.67% compared to GWO, PSO, and PSOGWO, respectively.