Studies On Application Of Combination Of Genetic Algorithm And Artificial Neural Network In City Water Environment Ultimate Carrying Capacity

Artificial neural network and Genetic algorithm, the main means applied in this thesis, takes part in the innovative study of the model building and forecasting of city water environment ultimate carrying capacity, aiming at the characters of city water environment ultimate carrying capacity, the artificial neural network model of city water environment ultimate carrying capacity has been established, genetic algorithm were combined for the purpose of weight training in an ANN. Ten important factors have been chosen tentatively as the evaluation index system of water environment carrying capacity. The model has been applied in Yinchuan city, Ningxia Hui Autonomous Region, which lies in arid area of northwest China. The results showed that genetic algorithm combined for the purpose of weight in an ANN was more effective than only ANN.The main fruits are shown as follows:1. This dissertation suggested the definition and evaluation index system of city water environment ultimate carrying capacity through consulting a lot of domestic and foreign references which studied the city water environment carrying capacity.2. To calculate and analyze tentatively on the river water environment capacity in Yinchuan city applying method for calculating water environment capacity in the case of one dimensional, establishing the foundation of industry and agriculture sustainable development about water environment of Yinchuan city.3. Artificial neural network and genetic algorithm were combined, which was applied that the establishment of model of city water environment ultiamte carrying capacity , based on analysis of respective character of artificial neural network and genetic algorithm.Regarding Yinchuan city as case, the thesis has forecasted the city water environment carrying capacity. The model shows three characteristics .The first is high resolution. Back-Propagation Network simulates the thought mode of human being, and needn't establish some fixed modes before judging the thing. It judges the directly and rationally according to essential characterisic, and its information is abundant. The second is wide application. The model may be used other cities in arid area of northwest China. The third is powrful practicability. The trained network only gives a input and gets the result of assessment immediately. It is easy to use.4. Adopting the MATLAB language to program and solve the model of city water environment ultimate carrying capacity.