云计算环境下基于时间和可靠性的调度策略  

Task scheduling strategy based on service time and reliability in cloud computing

作　　者：余盛季[1] 魏恺明 李强[1] 郭迅[1] Yu Shengji;Wei Kaiming;Li Qiang;Guo Xun(School of Computer Science & Engineering, University of Electronic Science & Technology of China, Chengdu 610000 , China)

机构地区：[1]电子科技大学计算机科学与工程学院,成都610000

出　　处：《计算机应用研究》2016年第9期2673-2678,共6页Application Research of Computers

摘　　要：云计算中,调度策略对于用户任务的运行和服务提供方均有很大的影响,直接影响到云平台的运行效果。针对传统任务调度算法难以适用于云计算环境这一问题进行了研究,采用由星型结构演化而成的多代理模型,给出了多代理模型的体系结构,然后针对多代理系统,提出了基于服务时间和可靠性为优化目标的任务调度策略,并采用模拟退火算法进行求解。最后进行了仿真实验,实验结果表明提出的调度策略是高效的。In cloud computing, task scheduling strategy has a great impact on the operating results of the user5 s task and the service provider’ s profit. Therefore, it has a direct impact on the cloud computing platform. To solve the problem that the traditional task scheduling algorithms could not suitable for cloud computing, this paper proposed a task scheduling strategy.Firstly, based on multi-agent system (M A S) extended from star structure, it described the multi-agent system structure.Then, it put forword a scheduling strategy considering the service time and reliability as optimization targets, and employed a simulated annealing algorithm to solve the optimization problem. Finally, simulation experiments show the efficiency of the proposed scheduling strategy.

关 键 词：多代理系统 调度策略 服务时间 服务可靠性 多目标优化 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]