Scholarlink

<p>The growing demand for efficient cloud computing solutions has highlighted the need for advanced task-scheduling techniques that minimize execution time and energy consumption. An appropriate fitness function for nature-inspired optimization may lead to improvement in performance. Our research aims to bridge this gap by proposing novel optimization-based task-scheduling techniques using nature-inspired meta-heuristic algorithms. An enhanced initialization approach for particle swarm optimization (PSO) is introduced, and an ellipsoid function as the fitness function to optimize key performance metrics, including makespan, execution time, energy consumption, and degree of imbalance, is applied. Additionally, we explore a genetic approach for further optimization. The proposed algorithms, longest job-based optimization (LJPBO), efficient task scheduling-based optimization (ETSBO), dynamic completion time-based optimization (DCTBO), and dynamic completion time-based genetic algorithm (DCTBGA), are evaluated against existing algorithms using performance measures. The results demonstrate substantial improvements of proposed algorithms LJPBO, ETSBO, DCTBO, and DCTBGA, with respect to MCTPSO with reductions in energy consumption of 50.6901%, 8.9008%, 20.2066%, and 78.0909%, respectively, for 100 VMs and 400 tasks. The experiments were conducted on a Windows machine with an Intel i7-6700 processor running at 3.4&#xa0;GHz and 8&#xa0;GB of memory. Further, MinMin PSO task scheduling algorithm is proposed, which achieved the lowest makespan.</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 novel hybrid meta heuristic approach for optimization of makespan and handling imbalance in task scheduling in cloud environments

  • 期刊论文

  • 发表日期：

    2025年7月9日

查看全文
  • Akanksha Tandon,
  • Anand Nayyar,
  • Sanjeev Patel

摘要

The growing demand for efficient cloud computing solutions has highlighted the need for advanced task-scheduling techniques that minimize execution time and energy consumption. An appropriate fitness function for nature-inspired optimization may lead to improvement in performance. Our research aims to bridge this gap by proposing novel optimization-based task-scheduling techniques using nature-inspired meta-heuristic algorithms. An enhanced initialization approach for particle swarm optimization (PSO) is introduced, and an ellipsoid function as the fitness function to optimize key performance metrics, including makespan, execution time, energy consumption, and degree of imbalance, is applied. Additionally, we explore a genetic approach for further optimization. The proposed algorithms, longest job-based optimization (LJPBO), efficient task scheduling-based optimization (ETSBO), dynamic completion time-based optimization (DCTBO), and dynamic completion time-based genetic algorithm (DCTBGA), are evaluated against existing algorithms using performance measures. The results demonstrate substantial improvements of proposed algorithms LJPBO, ETSBO, DCTBO, and DCTBGA, with respect to MCTPSO with reductions in energy consumption of 50.6901%, 8.9008%, 20.2066%, and 78.0909%, respectively, for 100 VMs and 400 tasks. The experiments were conducted on a Windows machine with an Intel i7-6700 processor running at 3.4 GHz and 8 GB of memory. Further, MinMin PSO task scheduling algorithm is proposed, which achieved the lowest makespan.