A Summary Evaluating And Analysis Of The Method Of Water Demand Prediction Based On The Artificial Neural Network

With the increase of population and fast development of economy, the shortage of water resources has become a restricted factor to the sustainable development of national economy and society. It is a very important task to research thoroughly water demand and predict water demand of our trades.In this paper, we mainly introduce the method of water demand prediction based on the artificial neural network, and further we evaluate and analyse it. There are five chapters in this paper. In the first chapter, methods of water demand prediction and questions and causations of water demand prediction are introduced. Important factors related to water demand are found by analyzing our country's water consumption datas. In the second chapter, learning process of BP network, error revised method and BP algorithm are introduced. According to BP algorithm, a predictive program is programmed with C language and validated by XOR question. Aiming at defects of BP network, some better measures to improve performance of BP network are also introduced. In the third chapter, we train our country's water consumption using synthesis method and subentry method based on prediction system of BP network designed, and predict water demand in the recent fifty years. In the fourth chapter, we analyze and compare predictive achievements of some domestic departments and experts in different predictable ways. Furthermore, our results in this paper are compared to the previous works too. The fifth chapter summarizes how to devise BP network.The experimental result shows that datas predicted by using BP network in this paper are very approximate to those of the expert by the water requirement per capita.