基于自适应分段广延指数模型的IPTV用户点播行为

IPTV user's on-demand behavior based on adaptive segmental stretched exponential model

为优化IPTV CDN系统配置,并改善视频调度与存储策略,使得用户获得更优质的视频访问体验,提出了一种基于自适应分段广延指数(adaptive segmental stretched exponential,ASSE)模型的IPTV用户点播行为建模方法。同时,采用中国电信股份有限公司(以下简称中国电信)IPTV实际用户点播行为数据对提出的ASSE模型进行视频访问概率建模仿真。实验表明,与传统建模方法相比,提出的ASSE模型不仅可以达到简化数学模型和减少计算的目的,而且可以更好地提高曲线拟合的精度,与实际数据具有良好的一致性。

The IPTV user＇s on-demand behavior model based on adaptive segmental stretched exponential（ASSE） was proposed to optimize the configuration of IPTV CDN system, to improve the video scheduling and storage strategy, and to provide users with better experience for video access. At the same time, the probability of the video access by using the actual IPTV user＇s on-demand data of China Telecom was simulated. Experiments show that compared with the traditional modeling methods, the ASSE model proposed not only achieves the purpose of simplifying the mathematical model and reducing the computation, but also improves the accuracy of the curve fitting. It has good consistency with the actual data.