Scholarlink

<p>This paper proposes DCOA-TS, a discrete adaptation of the Coati Optimization Algorithm for energy-aware task scheduling in heterogeneous cloud environments. DCOA-TS replaces continuous updates with tailored discrete operators swap, combine, and crossover to balance global exploration and local exploitation while preserving the assignment feasibility. It minimizes the weighted objective by integrating the makespan, energy consumption, and deadline-violation penalties. Extensive CloudSim Plus experiments across medium to large workloads show that DCOA-TS consistently outperforms classical heuristics, leading to metaheuristics, and reinforcement learning baselines, achieving up to 87.6% reduction in makespan (DCOA-TS: 9656&#xa0;s vs.&#xa0;FCFS: 77,919&#xa0;s at 100,000 tasks), substantial energy savings (up to 87% at 5000 tasks), and a lower average waiting time. All improvements are statistically significant (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(p &lt; 10^{-9}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>p</mi> <mo>&lt;</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>9</mn> </mrow> </msup> </mrow> </math></EquationSource> </InlineEquation>, Wilcoxon signed-rank test, 30 seeds). The gains stem from preferentially dispatching compute-intensive tasks to high-performance VMs and suppressing low-efficiency host activities, improving both utilization balance and power efficiency. The sensitivity analysis indicates that iterations have a larger impact than population size, underscoring the value of deeper exploitation in discrete scheduling. These results indicate that DCOA-TS as a strong candidate for large-scale, energy-efficient scheduling in heterogeneous clouds, delivering improved throughput, reduced power consumption, and better queuing fairness.</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

Energy-efficient task scheduling in heterogeneous cloud computing using discrete Coati Optimization Algorithm

  • 期刊论文

  • 发表日期：

    2026年5月22日

查看全文
  • Abdeldjalil Ledmi,
  • Makhlouf Ledmi,
  • Mohammed El Habib Souidi,
  • Nabil Azizi,
  • Hichem Haouassi

摘要

This paper proposes DCOA-TS, a discrete adaptation of the Coati Optimization Algorithm for energy-aware task scheduling in heterogeneous cloud environments. DCOA-TS replaces continuous updates with tailored discrete operators swap, combine, and crossover to balance global exploration and local exploitation while preserving the assignment feasibility. It minimizes the weighted objective by integrating the makespan, energy consumption, and deadline-violation penalties. Extensive CloudSim Plus experiments across medium to large workloads show that DCOA-TS consistently outperforms classical heuristics, leading to metaheuristics, and reinforcement learning baselines, achieving up to 87.6% reduction in makespan (DCOA-TS: 9656 s vs. FCFS: 77,919 s at 100,000 tasks), substantial energy savings (up to 87% at 5000 tasks), and a lower average waiting time. All improvements are statistically significant ( \(p < 10^{-9}\) p < 10 - 9 , Wilcoxon signed-rank test, 30 seeds). The gains stem from preferentially dispatching compute-intensive tasks to high-performance VMs and suppressing low-efficiency host activities, improving both utilization balance and power efficiency. The sensitivity analysis indicates that iterations have a larger impact than population size, underscoring the value of deeper exploitation in discrete scheduling. These results indicate that DCOA-TS as a strong candidate for large-scale, energy-efficient scheduling in heterogeneous clouds, delivering improved throughput, reduced power consumption, and better queuing fairness.