作者: Y. Liang
DOI: 10.1109/TSMCC.2003.818492
关键词:
摘要: In this paper, we systematically investigate the long-term, online, real-time variable-bit-rate (VBR) video traffic prediction, which is key and complicated component for advanced predictive dynamic bandwidth control allocation framework future networks Internet multimedia services. We focus on neural network-based approach prediction demonstrate that performance robustness of network predictors can be significantly improved through multiresolution learning. show predictor trained learning (called NN predictor) successfully predict various real-world VBR up to hundreds frames in advance, then lays a solid foundation mechanism. Also, control/allocation based long-term discussed detail.