Water Supply Warning System Based On Optimized Combination With Neural Network

With the rapid development of social economy, people's demand and dependence on water resources is increasing. As the rapid development of human society in the course of the predatory exploitation of water resources and the irrational use, in addition to its limited water resources and uneven, making the growing contradiction between supply and demand of water resources, aquatic ecosystems have been seriously damaged, water resources Security issues become more important.Water security as an important component of urban water supply security is to maintain the normal operation of the most basic protection. Changes in water quantity and water quality changes are early-warning security system in urban water supply the most important factor to consider in this paper draw on the comprehensive analysis of relevant laws and regulations and a large number of literature based on the proposed establishment of a moving average forecast, multi-line Regression prediction, based on an exponential smoothing forecast combination based on neural network optimization of water supply an early warning mechanism.The paper takes the Hechuan area of Chongqing for example, and the main contents of the thesis are as follows:First, in the variety of existing analysis and prediction model of BP neural network on the basis of a study, presented optimization based on neural network model of forecasting method of water supply.Secondly, in the variety of existing analysis and prediction model of BP neural network on the basis of a study, presented optimization based on neural network model of forecasting method of water supply.And then on the basis of the analysis and comparison of the various common prediction models, with hechuan district water information available, choose a moving average forecasting, prediction of multivariate linear regression, exponential smoothing optimal combination forecasting model based on neural network prediction model. Neural network model for optimized combination prediction model can be effective in preventing a single prediction of deviation, forecasting accuracy is high, forecast effect is better, especially in the case of more accurate historical data, use of combination forecasting model of water supply has more akin to actual consumption values.Lastly Combined with existing year in Chongqing Hechuan district water supply information, the establishment of a water supply prediction model based on BP neural network for years, and initially built years of water supply in the area of early warning mechanisms.