A support vector machine neural network is proposed for performing VBR video traffic prediction.

A support vector machine model is proposed to predict the VBR video traffic.

A novel dynamic bandwidth allocation algorithm named the hysteresis algorithm is presented based on VBR video traffic prediction in this paper, The buffer size requirement and cell loss rate are both improved significantly compared with threshold algorithm presented by Girish and the allocation scheme based on GOP scene detection presented by Youssef.

This paper studies the placement method of VBR video objects on the multi-zone disks, the organizing methods of the video files, and the disk I/O smooth scheduling and admission controlling policies.

A new type of VBR video multiplex strategy is proposed based on the identification of different frame types.

This paper summarizes all kinds of the bandwidth smoothing algorithms for prerecorded VBR video,compares the performances of these algorithms ,induces and formalizes these smoothing models,and gives the generalization description of them.

MPEG VBR video stream cannot be well modeled because of its complexity.

When usage of network resources and dynamic bandwidth allocation are considered, one important part is VBR video modeling and its traffic prediction.

It is capable of describing the video service in the case of scenery change, and its specific feature of Gamma distribution follows fairly well the bit rate of a VBR video service.