Applied Research of Pantograph Carbon Slide Formulation Optimization Based on Genetic Neural Network

摘要

Applying BP neural network improved by GA to build a nonlinear multi-objective model between formulation ingredient and mechanical properties of electric locomotive pantograph carbon slide. And on the basis of test data improved neural network is trained; finally we can get a genetic neural network model. Therefore, when given specific formulation of carbon slide plate, it is able to predict the corresponding mechanical properties. The research shows that the trained genetic neural network can accurately predict relevant properties of carbon slide plate and be confirmed the practical significance of the algorithm.