BP神经网络在水资源承载能力预测中的应用

采用三层的BP神经网络结构，选取与水资源承载力密切相关的6个社会经济指标，根据1995～2004年10年时间序列指标数据，运用MATLAB中改进的BP神经网络算法建立了长春市水资源需求量预测模型，通过预测值和检验值的误差比较，表明预测模型的精度较高。参考长春市“十一五”规划期间的社会和经济发展目标，预测得到“十一五”期间长春市水资源需求量，对比现有供水能力，“十一五”期间水资源承载能力无法满足社会和经济的发展要求，为实现资源、社会和经济的协调发展，从开源和节流两个方面提出了水资源的可持续利用对策。

Application of BP neural network to prediction of carrying capacity of water resources

With application of three-layers BP neural network, six social and economical indexes related to the demands of water resources are selected, and then the Predication Model of Water Resources Demands for Changchun is established with the improved BP neural network algorithms in the MATLAB in accordance with the indexes data. Through comparison, the error between the prediction value and the test value shows that the precision of the prediction model is higher. Based on the target of the social and economic development of the period of the 11 th Five Year Plan for Changchun, the demands of water resources from the city during the period of the 11 th Five Year Plan are predicted. Compared with the current water supply capacity, the requirements of the social and economic development can not be satisfied by the carrying capacity of water resources during the period concerned. Therefore, a countermeasure for the sustainable utilization of water resources is put forward from the aspects of saving and development of water resources, so as to realize the harmonious resources, social and economic developments therein.