Application And Research Of The BP Neural Network Algorithm In The Transmission Network System On QOS

As mobile transmission business diversification, the development of heavy and complicated, the mobile transmission network requirement also increasingly excellence. In the traditional network transmission, processing most belongs to no time limit on the application of the system. For example at present we are still using web applications, E-mail Settings, and so on. But the data in business is to make up most of the time delay, jam, throughput and lost has great requirements such as network applications. So reliable transmission network become numerous business application data one of the important factors.Neural network is simulation of artificial intelligence had built a mathematical model. It is a parallel processing, distribution of type structure; Each processing unit can have any branch; Processing unit of output signal of also can be arbitrary mathematical model; To each part processing unit can be local operations. These are neural network can be used the advantage of network communication problems. In the neural network in use in the bp neural network, and the BP neural network has good fault tolerance and excellent associative memory function, and it itself is has strong ability to learn and adapt to the ability, the excellent training method can make each processing units to achieve the desired output.This paper using bp neural network learning ability and adaptive ability, make whole transmission network in the routing, continuously adjust its weight value spread back until the choice to optimal path. And this paper in order to adapt to the actual transmission network routing condition, use waxman-salama model of network topology generation method, simulate the effective network structure, carries on the simulation.The simulation results show that, the paper designed based on the bp neural network, and to realize the QOS performance of the optimal solution of the method is feasible, and considering two optimizing targets to meet when only consider a significantly better than optimization purposes.