Research on fault diagnosis model of computer network based on GA-BP neural network

The popularity of the computer has prompted obvious changes in people"s work and life style at this stage, providing people with quality services, but affected by a variety of factors, computer network failures occur more frequently, and then reduce its functionality, in order to ensure the healthy and stable operation of the local area network, the back propagation (BP) neural network model was optimized by genetic algorithm (GA), and the GA-BP neural network model was constructed. This paper introduces the basic model construction process and principle, and simulate computer network failures, validate models, and perform data collection and validation. The results show that the optimized model (GABP neural network) is superior to BP neural network in stability, prediction accuracy and convergence speed, and can be used in computer fault diagnosis, the nonlinear approximation and local optimization ability of BP neural network are combined with the global search characteristics of genetic algorithm. Therefore, GA-BP can predict computer network fault more scientifically and reasonably, and provide an effective implementation path and development strategy for improving the efficiency of computer fault diagnosis.