Bp Neural Network-based Network Fault Diagnosis System

With the rapid development of computer network technology, the size of the network is becoming larger and larger, and the function of the network is becoming more and more magic, and the sum of people who use the network is growing dramatically. All these cause that it is hard to manage the network, especially to diagnose the network fault. The causes of the network faults are varied, and they always contain a lot of information. So it is very hard for the network manager to solve the entire network faults by the tools which are provided now. Using intelligent methods to diagnose the network faults is the development trend in this field.In this paper, the Artificial Neural Network technology is applied to the network fault diagnosis field, and according to the disadvantages of BP ANN, such as the learning speed is very slow, easy to drop in the local minimal points, and the peculiarity of the network fault diagnosis, a new arithmetic based on Rough Set theory and Momentum Node is brought forward. It is proved that this new arithmetic has improved a lot in the disadvantages of BP ANN. In this paper, a software system base on the improved arithmetic is designed, and it is developed by Microsoft Visual C++ 6.0 and Matlab 6.5 in the environment of Microsoft Windows 2000.The software is used and tested in the actual network environment, the experimental results are very satisfied.