Scholarlink

<p>Optimal resource utilization stands as the primary challenge for cloud workflow scheduling conducted by service providers. Achieving conflicting objectives becomes extremely difficult when the scheduler must address requirements related to execution time, costs, and energy consumption parameters. We model the workflow problem as a constrained multi-objective optimization problem. In addition to dealing with the conflicting objectives above, the formulated problem also contains task execution dependencies, resource capacity limits, deadline limits, and energy consumption thresholds. This paper introduces a novel enhanced hybrid algorithm, which is a strategic and implemented combination of the Artificial Bee Colony (ABC) and Genetic Algorithm (GA) using a special staged architecture to solve the formulated constrained multi-objective optimization problem. The essential innovations consist of problem specific, feasibility maintaining genetic operators and dynamic multi-constraint handling mechanism, which all together guarantee robust and efficient scheduling. ABC procedures enable exploration in the hybrid approach while GA provides local exploitation. The proposed hybrid model incorporates the features of ABC-GA integration with adaptive penalty functions and implementation-based genetic operators that preserve feasibility constraints. Standard workflow benchmarks in experimental tests show that the hybrid ABC-GA solution produces superior results compared to ABC, GA, and other existing advanced methods. Experimental results reveal that the proposed ABC-GA hybrid algorithm can execute a 100-task workflow in 412.8&#xa0;s, at a total cost of 18.2 $, consuming 4.05 kWh of energy. Using ratios, the proposed ABC-GA hybrid system achieves a 14% faster makespan duration, an 11% reduction in operating costs, and 9% lower energy consumption than its closest competitor. Using different workflow sizes, the proposed ABC-GA hybrid system consistently outperforms modern algorithms such as GA, ABC, and other hybrid algorithms.</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

An enhanced hybrid artificial bee colony and genetic algorithm for multi-objective workflow scheduling in the cloud

  • 期刊论文

  • 发表日期：

    2026年2月2日

查看全文
  • Medhat A. Tawfeek,
  • Ibrahim Alrashdi,
  • Madallah Alruwaili,
  • Bader Aldughayfiq,
  • Hisham Allahem

摘要

Optimal resource utilization stands as the primary challenge for cloud workflow scheduling conducted by service providers. Achieving conflicting objectives becomes extremely difficult when the scheduler must address requirements related to execution time, costs, and energy consumption parameters. We model the workflow problem as a constrained multi-objective optimization problem. In addition to dealing with the conflicting objectives above, the formulated problem also contains task execution dependencies, resource capacity limits, deadline limits, and energy consumption thresholds. This paper introduces a novel enhanced hybrid algorithm, which is a strategic and implemented combination of the Artificial Bee Colony (ABC) and Genetic Algorithm (GA) using a special staged architecture to solve the formulated constrained multi-objective optimization problem. The essential innovations consist of problem specific, feasibility maintaining genetic operators and dynamic multi-constraint handling mechanism, which all together guarantee robust and efficient scheduling. ABC procedures enable exploration in the hybrid approach while GA provides local exploitation. The proposed hybrid model incorporates the features of ABC-GA integration with adaptive penalty functions and implementation-based genetic operators that preserve feasibility constraints. Standard workflow benchmarks in experimental tests show that the hybrid ABC-GA solution produces superior results compared to ABC, GA, and other existing advanced methods. Experimental results reveal that the proposed ABC-GA hybrid algorithm can execute a 100-task workflow in 412.8 s, at a total cost of 18.2 $, consuming 4.05 kWh of energy. Using ratios, the proposed ABC-GA hybrid system achieves a 14% faster makespan duration, an 11% reduction in operating costs, and 9% lower energy consumption than its closest competitor. Using different workflow sizes, the proposed ABC-GA hybrid system consistently outperforms modern algorithms such as GA, ABC, and other hybrid algorithms.