BP神经网络法预测唐山市需水量

需水量预测研究已成为当前水资源规划与管理研究中的重要课题之一.本文设定不同的神经网络运行次数,根据预测结果进行误差分析,BP神经网络在运行5 000次时,具有高度的可信度和可行性.应用5 000次运行次数的BP神经网络模型对唐山市规划水平年的需水量进行预测.最后引入人均综合用水量概念,结果表明,预测结果在理论上和实际上都具有可行性.

Prediction of urban water demand in Tangshan City with BP neutral network method

The present paper aims to introduce the authors' research on the methods to predict the water demand of Tangshan area and the actual water resources planning and management for the time being.Once different neutral network operation times are chosen and the errors between the prediction data and the original data are analyzed,the paper has made a conclusion that BP neutral network is of high reliability and good feasibility under the 5 000 operation times.Then BP neutral network with 5 000 operation times can be applied to predict Tangshan urban water demand in the different future years.And,finally,the comprehensive water demand per capita is put forward,thus,the prediction results prove to be feasible both in theory and in practice.It should be pointed that the BP nentral network has its own advantages,such as higher reliability,better feasibility and veracity,and it can avoid successfully fussy calculation.But applying BP neutral network to predict urban water demand belongs to across subject,the function mechanism of connotative nerve cell and the choice of node number are very difficult,therefore,it will be further studied and discussed,especially how to ascertain connotative node number.