年均径流预测的遗传神经网络模型研究

Study and Application on Genetic-Neural Network Prediction Model for the Annual Average Runoff of Reservoir

为了进一步提高BP网络模型对径流预测的精度,采用遗传算法优化了BP网络初始的权值和阈值。实例研究结果表明:该方法克服了传统BP网络极易陷入局部极小值点等缺点,提出的遗传神经网络预测模型能够提高预测精度。

In order to improve the predication accuracy of back propagation network model for annual runoff of reservoir,back propagation network had been used to optimize the weights and threshold by genetic algorithm of prediction model.The study results show that,this method overcomes the shortcomings that traditional back propagation network traps into local minima easily,the genetic-neural network model has a higher accuracy than that of traditional BP network.