作　　者： ; ;

机构地区： 广东省水利水电科学研究院

出　　处： 《资源科学》 2007年第1期99-105,共7页

摘　　要： 人工神经网络是复杂非线性科学和人工智能科学的前沿理论,目前在水资源承载力领域的应用研究还很少见。结合神经网络基本原理、方法与水资源承载力基本理论,建立水资源承载力与其影响因子间的定量关系模型———基于神经网络的水资源承载力耦合模型。以惠州市为研究区域,预测了未来水平年不同供水保证率下水资源承载力方案,反映了水资源对社会、经济和生态环境协调发展瓶颈作用;综合评估了研究区域现状和未来的水资源承载力状况,表明了区域水资源利用与社会经济发展匹配性程度。结果表明,神经网络理论能较好解决水资源承载力研究建模问题,计算结果较合理。 ANN is the foreland of nonlinearity and artificial intelligence, and is seldom applied in the field of water resources carrying capacity （WRCC） research. This paper introduced the method of ANN into the WRCCR research firstly. And through establishing the model between WRCC and its influencing factors the coupling WRCC model based on ANN and taking Huizhou city as study region, the quantificational computation of WRCC is discussed. Recently, most of the WRCC models were constructed by coupled system of economy and water resources, and because of the complexity between the two systems, few good results were obtained. This paper explored the complex relationship between these two systems using the ANN model and succeeded in avoiding the difficulties which had happened in other methods. Some method was advanced to compute the ultimate WRCC. Based on the exploitation of water resources and the economic development model, the carrying states of WRCC were forecasted by ANN, and the ANN model was used to classify the WRCC with the index that had been forecasted. Then, this computing process was used to obtain the ultimate WRCC. This paper adopted the method of curve matching error to avoid variable＇ s non-smooth process in the WRCC model. When the ANN forecasting model of WRCC is established, it uses the training of matching error, exports the forecasting error and compensates the forecasting of the dynamic WRCC. In this way, the extension issue of BP network for forecasting is well solved and it improves the forecasting precision of the nonlinear factors in the system of WRCC. The WRCC model based on ANN can simulate the diversified dynamic changing trends of WRCC according to the different developing projects and estimate the region＇s WRCC state. The paper took the Huizhou city as the study the same time the results＇ rationality was analyzed reflect the influence from human being. It can also region and forecasted its WRCC in 2010 and 2020. At contrasting the results computed by the methods of multi-objec

关 键 词： 水资源承载力 神经网络 预测 等级评价 惠州市

领　　域： [自动化与计算机技术] [自动化与计算机技术]