基于遗传与BP混合算法神经网络预测模型及应用

提出用遗传学习算法和权重调整BP算法相结合的混合算法来训练模糊模式识别神经网络预测模型；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，使网络收敛速度加快和避免局部极小。作为实例，以新疆雅马渡站的实测径流资料和相应的前期4个预报因子实测数据作为样本进行训练并用以预测雅马渡站的年径流量。结果表明，该方法具有收敛速度快和预测精度高的特点。

Neural network prediction model and its application based on GA and BP

A new method for training the neural network prediction model with fuzzy pattern recognition is presented. In this method, the genetic algorithm(GA), a general purpose global search algorithm is used to train the neural network prediction model with updating the weights to minimize the error between the network output and the desired output. Then the back propagation (BP) algorithm is used to further train the neural network prediction model with fuzzy pattern recognition. This method is used to speed up the convergence and improve the performance. To demonstrate the procedures and performance of this neural network-training algorithm, the case of Yamadu, Xinjiang, is analyzed and discussed. The recorded data of runoff and its four affecting factors are adopted as training pattern in order to predict annual runoff for Yamadu.