改进的BP算法在黄河下游枯季径流预测中的应用

本在黄河下游地区应用多层前向人工神经网络理论，通过改进BP算法，建立下游枯季径流预测的BP神经网络模型，使用花园口-利津水站26年的完整序列测流资料训练和检验网络并用于预测，结果表明，通过本次研究建立的BP网络模型是合理的，可靠的，它较好地反映了黄河下游的枯季径流规律，可为今后黄河流域水资源的统一调度，管理，尤其是预防黄河下游再次出现断流提供科学依据。

APPLICATION OF IMPROVED BP ALGORITHM TO DRY SEASON RUN-OFF PREDICTION IN THE LOWER YELLOW RIVER

BP algorithm is one of the most important algorithms of the artificial neural network. It has a wide application range. The improved BP algorithm based on the Fletcher Reeves algorithm is described in the present paper. By using the improved BP algorithm, the problems of slow convergence speed and local optimization solution are solved. By analyzing the influencing factors the BP model of the dry season run off prediction in lower Yellow River is set up. Then the training and the test with the recorded data of 26 years of Huayuankou Station and Lijin Station of Yellow River based on the BP model is given. The results of the training and test show that the BP model is reasonable and reliable. It is concluded that the model may be used to illustrate the regulations of dry season run off in lower Yellow River. A run off prediction for the next year of the lower Yellow River is also carried out. The results show that the BP model may provide scientific bases for integrated management and administration of the Yellow River basin water resources.