贝叶斯MLP神经网络在IP承载网性能预测的研究

Research on performance prediction of IP network by Bayesian MLP neural network

通过对贝叶斯MLP神经网络的学习,发现它在处理大量的实时数据时,具有较高的精准度还能够保证IP承载网的强泛化能力,并且预测结果容易达到运营商的预期。为此尝试将贝叶斯MLP神经网络引入目前IP承载网性能预测,并且通过实验验证贝叶斯MLP神经网络在IP承载网中进行性能预测的可行性结论。

The method of Bayesian MLP neural network not only has higher precision in the processing of large amounts of real- time data, but also ensure the generalization ability of IP carrying network, and it is easier to achieve the operator＇s expected from the prediction results. In the paper, Bayesian MLP neural network is introduced into the IP network for performance prediction. The results of simulation show the algorithm of Bayesian MLP neural network can predict the performance of IP network and evaluate its effectiveness.