摘要
用遗传算法和BP算法相结合的混合算法来训练日径流神经网络预测模型的权值,即先通过遗传学习算法进行全局训练,再用权重调整BP算法进行精确训练,这一算法克服了BP算法收敛速度慢、易陷入局部极小等缺陷,实例证明提高了预测精度。
A new method for training Artificial Neural Network(ANN) based daily runoff prediction model is presented. In this method, the genetic algorithm(GA), a general-purpose global search algorithm is used to train the neural network prediction model by updating the weights to minimize the error between the network output and the desired output. It overcomes the limitations of the back-propagation algorithm in slow convergent rate and getting into local optima. The example demonstrates that this method improves the prediction precision.
出处
《西北水电》
2003年第4期1-3,21,共4页
Northwest Hydropower
基金
武汉市晨光计划项目(20005004028)资助.
关键词
遗传算法
日径流预测
神经网络
BP算法
Genetic Algorithm
Artificial Neural Network
runoff prediction
