ChinaCom2008-Multimedia Communications Symposium

Research Article

An Adaptive Model of Task Scheduling Based on QoS Feedback Control

  • @INPROCEEDINGS{10.1109/CHINACOM.2008.4685209,
        author={Qi Qi and Jianxin Liao and Xiaomin Zhu and Chun Wang and Hai Huang},
        title={An Adaptive Model of Task Scheduling Based on QoS Feedback Control},
        proceedings={ChinaCom2008-Multimedia Communications Symposium},
        keywords={Non-real time task; QoS feedback control; task schedulin},
  • Qi Qi
    Jianxin Liao
    Xiaomin Zhu
    Chun Wang
    Hai Huang
    Year: 2008
    An Adaptive Model of Task Scheduling Based on QoS Feedback Control
    DOI: 10.1109/CHINACOM.2008.4685209
Qi Qi1,*, Jianxin Liao1,*, Xiaomin Zhu1,*, Chun Wang1,*, Hai Huang1,*
  • 1: State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing, China
*Contact email:,,,,


The traditional RR (Round Robin) task scheduling algorithms and real-time feedback control models are hard to satisfy the QoS changing demands for non-real-time tasks on data value-added services systems. An adaptive model of task scheduling is proposed in this paper based on the IDP (Integrated Data value-added services Platform) project. The model includes a hierarchical scheduling algorithm, taking feedback QoS information, such as message sending success rate and the user’s attention degree into account while controlling the task execution, so the system can automatically adapt to the shaking of the bandwidth of network links and the changing of user’s demand. Experiments on the performance of the algorithm are taken in IDP in the laboratory. Result of the experiment show that the variance of system operational efficiency is less than 0.0478 and the task unit response time is decreasing function for user’s inquiring times. So the adaptive task scheduling model can effectively solve the scheduling of real-time tasks and non-realtime tasks in data value-added services system, and it is also useful of other system by plugging in some feedback controller module.