基于支持向量机的有效带宽估计算法的研究

高性能网络设计的关键是对性能参数进行建模和估计的能力，通信业务在网络上传输时，其带宽要求与通信质量密切相关，合理分配带宽能很好提高网络性能，合理分配网络带宽的首要任务是能够根据网络信息的变化，预测带宽使用情况，为此，本文采用支持向量机（SVM）对网络业务流数据进行预测估计，通过训练样本，从而获得样本以外数据的分布规律，在此基础上，设计了一种网络传输有效带宽使用的估计算法，实验表明，该估计算法具有较高的训练效率和很高的估计精确度。

Research on Available Bandwidth Estimation Algorithm Based on Support Vector Machine

Modeling and estimation of performance parameters play an important role in the design of high performance network. The available bandwidth is related closely to QoS when the traffic transfers in the network. The rational allocation of the bandwidth can improve network performance. Based on the theory of Support Vector Machine, we obtain distribution law of other data by analyzing and training the sample data. An available bandwidth estimation algorithm is implemented in this paper. Simulation results show that this estimation algorithm has higher training efficiency and estimation precision.