Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (36): 244-248.DOI: 10.3778/j.issn.10w02-8331.2009.36.070

• 工程与应用 • Previous Articles    

Prediction and adaptive query scheduling of non-dedicated resources on massive open online courses

HOU Yong1,Wushou Sliamu1,Turgun1,ZHOU Yan-hui2,ZHANG Kai2   

  1. 1.School of Information Science and Engineering,Xinjiang University,Urumqi 830046,China
    2.The State Administration of Radio Film and Television Authorities,Service Center,Beijing 100866,China
  • Received:2009-08-21 Revised:2009-10-26 Online:2009-12-21 Published:2009-12-21
  • Contact: HOU Yong

大规模在线课程非专用资源预测与查询调度

侯 勇1,吾守尔·斯拉木1,吐尔根·伊布拉音1,周艳慧2,张 凯2   

  1. 1.新疆大学 信息科学与工程学院,乌鲁木齐 830046
    2.国家广播电影电视总局 机关服务中心,北京 100866
  • 通讯作者: 侯 勇

Abstract: Massive Open Online Course system,MOOC,requires a large number of resources for supporting users to get good services.Resources of MOOC used dynamically and fixed time for rental waste a lot of resources.The method of non-dedicated resources is applied to MOOC,to reduce the overall amount of resources to hire.First of all,using Markov birth-death process to predicate resources for getting non-dedicated resources;secondly,setting up query expected cost matrix,QEC;thirdly,using Adaptive Min-Min and Max-Min,A-MM,algorithm for adaptive query scheduling;finally,experiments are done and show that the prediction error rate is low and the performance of prediction is stable;and A-MM has a good efficiency and balance capacity.

Key words: non-dedicated resource, adaptive query scheduling, prediction of resource

摘要: 大规模在线课程系统需要大量资源作支撑。资源使用的动态性及租用的固定性导致资源被浪费。将“非专用”思想应用于系统,可整体降低资源租用量。首先通过马尔可夫“生灭”过程预测资源状态,筛选出“非专用”资源;接着建立查询期望代价矩阵,使用A-MM(Adaptive Min-Min and Max-Min)算法进行自适应查询处理;最后实验表明,预测错误率低,性能稳定;A-MM有较好执行效率和平衡负载能力。

关键词: 非专用资源, 自适应查询调度, 资源状态预测

CLC Number: