视频点播系统用户行为仿真平台

大型VOD（Video On Demand)系统需要设计用户许可控制、流调度等一系列算法，然而为研究此类算法而建立大型系统并不现实。本文提出VOD系统用户行为仿真平台，模拟大型VOD系统的用户行为并为各类算法提供仿真运行的平台。该平台描述了VOD系统用户行为的一般数学模型，刻划了典型用户进入系统行为分布、节目选择分布和VCR操作分布，设计了用户行为仿真平台结构，并完成了该平台，此平台提出的变强度用户流的概念，使对实际用户行为的仿真更加真实。实验表明，VOD系统用户行为仿真平台客观的反映了VOD系统的用户行为，同时还为其他算法研究和算法验证提供方便易用的基础。

Simulation Platform for User Actions in the Video-on-Demand System

User admission schemes and streaming scheduling algorithms are important for huge-scale VOD (Video-On-Demand) system, but it is not practical to build real system to validate these algorithms. In the paper, a simulation platform for user actions in VOD system is proposed to simulate huge-scale system and to provide a testing platform for designing the various algorithms. A general mathematics model for user actions in VOD system is built to present the distribution of user visiting action, distribution of program selection and distribution of VCR functions. The platform is structural designed and implemented. A definition of variational frequency for user visiting is also proposed in the platform to make the simulation more objective. The simulation indicates that the platform objectively reflects the user actions in VOD system, and it is also provide an easy-used basic for designing and validating the algorithms.