Application of Genetic Neural Network for Predicting the Evolution of Shoal in River

Studying on river shoal evolution is a fundamental work in the science of water conservancy, water conservancy projects and waterways planning, designing, engineering feasiblility. First of all, the neural network model for predicting the evolution of shoal in a river is established, through training the neural network to determine the number of hidden layer´s neural, thus, a more ration neural network structure is detemaned; In the second step, by using genetic algorithm a fittest initial weight value is selectled from the solution group of initial weight values to avoid the blindness in the selection of initial weight value; if the traditional BP algorithm is used in the training, there are still some hidden dangers. Finally, conjugate gradient algorithm is used to improve performance of network. Simulation results show that the method is feasible and effective.