大数据分析下网络应急资源实时调度仿真研究  

Real-Time Scheduling for Network Emergency Resources under Big Data Analysis

作　　者：张光勇[1] 侯剑[2] ZHANG Guang- yong;HOU Jian(Shandong University of Technology, Zibo Shandong 255000, China;Shandong Normal University, Jinan Shandong 250014, China)

机构地区：[1]山东理工大学,山东淄博255000 [2]山东师范大学,山东济南250014

出　　处：《计算机仿真》2018年第4期445-448,452,共5页Computer Simulation

摘　　要：对网络应急资源的实时调度,能够有效提高大数据下网络资源处理效率。对网络应急资源的调度,首先需将信誉因子引入到资源调度中,并对调度决策进行后期修正,完成已应急资源的高效调度。传统方法组建应急资源实时调度模型,得到调度目标函数,但忽略了对应急资源使用情况做评估,导致资源调度效果不理想。提出基于改进贝叶斯的网络应急资源实时调度方法。采用有向无环图表述每个网络应用,将信誉因子引入到资源实时调度中,对网络资源使用情况进行实时评估,刺激节点在任务结束后释放占用资源,给出资源调度策略,并对调度决策进行后期修正,利用修正的结果进行网络应急资源实时调度。实验结果证明,所提方法在大数据分析的环境下,保障了资源实时调度的高效性,并有效地提升了网络资源利用率。The traditional method ignores the evaluation of service condition of emergency resources, so that the effect of resource scheduling is not satisfactory. This paper presents a real - time scheduling method for network e- mergency resources based on improved Bayesian network. The directed acycline graph is used to describe each net- work application. The credibility factor is introduced to the real - time scheduling of resources, thus working condi- tion of emergency resource is estimated in real time. Then, nodes are activated to release resource at the end of task. The resource scheduling strategy is given and the scheduling decision is corrected. Finally, the corrected result is used to dispatch network emergency resources in real time. Simulation results show that the proposed method ensures the high efficiency of real - time resource scheduling and improves the rate of utilization of network resources in the environment of big data analysis.

关 键 词：大数据 应急资源 实时调度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]