Scholarlink

<p>Cloud computing has rapidly evolved to become an indispensable technology across various domains, offering scalable and on-demand services tailored to user requirements. Despite its benefits, this paradigm introduces numerous challenges, particularly in task scheduling a critical factor influencing system performance in Infrastructure-as-a-Service (IaaS) environments. Task scheduling in IaaS clouds remains a challenging multi-objective optimization problem, especially under dynamic workloads and resource heterogeneity. However, many existing metaheuristic-based methods suffer from issues such as premature convergence and stagnation in local optima, leading to suboptimal task-to-resource mapping. To address these limitations, this paper presents a novel hybrid optimization algorithm which integrates the exploration strength of the Dwarf Mongoose Optimization (DMO) algorithm with the diversity-enhancing Chaotic Local Search (CLS) strategy. The algorithm is implemented in the CloudSim simulation environment and evaluated using real-world workload traces GoCJ and HPC2N under both fixed and variable virtual machines settings. The performance is assessed based on metric including makespan, execution cost, and resource utilization. Results reveal that CLSDMO consistently achieves lower makespan values, recording a makespan of 314.37 compared to 321.99, 326.86, and 336.94 for the other algorithms respectively at 1000 tasks. Statistical t-tests were performed to assess the significance of observed improvements. CLSDMO showed statistically significant improvement (<i>p</i> &lt; 0.05) demonstrating reductions in makespan with high confidence levels, indicating robust scalability and scheduling quality.</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

A chaotic local search-based dwarf mongoose optimization algorithm for multi-objective task scheduling in cloud computing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月30日

查看全文
  • Olanrewaju Lawrence Abraham,
  • Md Asri Ngadi,
  • Johan Bin Mohamad Sharif,
  • Mohd Kufaisal Mohd Sidik

摘要

Cloud computing has rapidly evolved to become an indispensable technology across various domains, offering scalable and on-demand services tailored to user requirements. Despite its benefits, this paradigm introduces numerous challenges, particularly in task scheduling a critical factor influencing system performance in Infrastructure-as-a-Service (IaaS) environments. Task scheduling in IaaS clouds remains a challenging multi-objective optimization problem, especially under dynamic workloads and resource heterogeneity. However, many existing metaheuristic-based methods suffer from issues such as premature convergence and stagnation in local optima, leading to suboptimal task-to-resource mapping. To address these limitations, this paper presents a novel hybrid optimization algorithm which integrates the exploration strength of the Dwarf Mongoose Optimization (DMO) algorithm with the diversity-enhancing Chaotic Local Search (CLS) strategy. The algorithm is implemented in the CloudSim simulation environment and evaluated using real-world workload traces GoCJ and HPC2N under both fixed and variable virtual machines settings. The performance is assessed based on metric including makespan, execution cost, and resource utilization. Results reveal that CLSDMO consistently achieves lower makespan values, recording a makespan of 314.37 compared to 321.99, 326.86, and 336.94 for the other algorithms respectively at 1000 tasks. Statistical t-tests were performed to assess the significance of observed improvements. CLSDMO showed statistically significant improvement (p < 0.05) demonstrating reductions in makespan with high confidence levels, indicating robust scalability and scheduling quality.