Abstract:In order to solve the problem of BP neural network, like local minimum and convergence rate, the author effectively combined DNA algorithm and neural network. By utilizing global searching ability of DNA algorithm and optimizing the initial weights and threshold value of the optimized network, the two problems of the BP neural network would be solved and the accuracy and rapidity of BP neural network diagnosis would be further improved. Taking the fault diagnosis of switch control circuit as the object of the study, this paper establishes fault diagnosis model of BP neural network optimized by DNA algorithm and conducts simulation analysis for fault diagnosis model with MATLAB software. The result demonstrates that the generalization ability and the accuracy of BP neural network optimized by DNA algorithm are better than BP neural network.