.png){kind=icon}
返回
A Two-Stage Multi-Objective Task Scheduling Framework Based on Invasive Tumor Growth Optimization Algorithm for Cloud Computing
.png){kind=icon}
PRE
AI
DOI：10.1007/s10723-023-09665-y.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Task scheduling in cloud computing is usually required to achieve multiple goals from the perspective of cloud service providers, users, environmental benefits, and so on. However, there are often conflictions among these goals, and the constraints might be diverse and strict. Since scheduling strategies need to be made efficiently and effectively, multi-objective task scheduling optimization becomes a huge challenge. Aiming at collaboratively optimizing three conflicting goals, including batch task completion time, energy consumption and idle resource costs, this paper proposes a multi-objective scheduling framework MSITGO based on Invasive Tumor Growth Optimization (ITGO). MSITGO utilizes the characteristics of tumor cell growth model and adopts the Pareto optimal model and packing problem model to perform a fine-grained and efficient search in solution space, which effectively enhances the diversity of solutions and increases the speed of convergence. In addition, considering an entire task processing procedure, MSITGO assembles the task scheduling process into two stages as machine assignment and timeslot allocation, to further improve the task scheduling performance and reduce unreasonable allocations. Experimental results on real-world cluster data from Alibaba show that MSITGO can provide a better solution to the proposed multi-objective task scheduling problem compared with other state-of-the-art algorithms.
Keyword:
Multi-objective optimization
Task scheduling
Cloud computing
MSITGO
期刊
IF:
2.9
论文数:
761
被引数:
1.2K
机构
引用论文
.png){kind=icon}
分享收藏Improved many-objective particle swarm optimization algorithm for scientific workflow scheduling in cloud computing面向云计算科学工作流调度的改进多目标粒子群算法
PREAI
分享收藏
查看更多内容