Scholarlink

<p>This paper addresses the multi-objective task scheduling problem with task deadlines and scheduling budget constraints, and proposes an improved particle swarm optimization algorithm (Quality of Service Guided Multi-objective Discrete Particle Swarm Optimization, QoS-MDPSO). The algorithm aims to simultaneously minimize the maximum completion time and total execution cost of the task set, while maximizing the overall credibility of the task set. The overall credibility is defined as the sum of the product of the task criticality and the credibility of the assigned virtual machine. The QoS-MDPSO algorithm introduces an adaptive learning strategy that dynamically adjusts the learning parameters of particles, enhancing their exploration ability in the search space and effectively avoiding local optima. Additionally, the algorithm incorporates a differentiated update mechanism, optimizing particles based on their fitness levels. High-fitness particles are further refined, while low-fitness particles are encouraged to explore the solution space more comprehensively, thus improving particle quality and enhancing the algorithm’s global optimization capability. Experimental results show that compared to other methods, QoS-MDPSO demonstrates significant performance improvements: the maximum completion time is reduced by 19.69%, the total execution cost is reduced by 4.39%, and the overall credibility is increased by 7.04%.</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

Multi-objective cloud scheduling based on particle swarm optimization algorithm with deadline and budget constraints

  • 期刊论文

  • 发表日期：

    2025年9月11日

查看全文
  • Chen Ouyang,
  • Yi-Wen Zhang

摘要

This paper addresses the multi-objective task scheduling problem with task deadlines and scheduling budget constraints, and proposes an improved particle swarm optimization algorithm (Quality of Service Guided Multi-objective Discrete Particle Swarm Optimization, QoS-MDPSO). The algorithm aims to simultaneously minimize the maximum completion time and total execution cost of the task set, while maximizing the overall credibility of the task set. The overall credibility is defined as the sum of the product of the task criticality and the credibility of the assigned virtual machine. The QoS-MDPSO algorithm introduces an adaptive learning strategy that dynamically adjusts the learning parameters of particles, enhancing their exploration ability in the search space and effectively avoiding local optima. Additionally, the algorithm incorporates a differentiated update mechanism, optimizing particles based on their fitness levels. High-fitness particles are further refined, while low-fitness particles are encouraged to explore the solution space more comprehensively, thus improving particle quality and enhancing the algorithm’s global optimization capability. Experimental results show that compared to other methods, QoS-MDPSO demonstrates significant performance improvements: the maximum completion time is reduced by 19.69%, the total execution cost is reduced by 4.39%, and the overall credibility is increased by 7.04%.