人工神经网络在工业污染源综合评价中的应用

现有污染源评价方法很多，但均为线性方程，由于环境、经济、社会指标间相互关系非常复杂，既存在多重共线性，又存在非线性关系，这使得应用现有方法对污染源进行环境经济属性综合评价时难以做到周全。以深圳市为例，首次引入人工神经网络（ANN）对工业行业污染源进行综合评价，并采用等标污染负荷和综合判别指数两种方法进行对比。结果表明，ANN法的评价结果最优。由于ANN具有模拟任意非线性连续函数的能力及自学习的特点，因此适用于污染源的综合评价。

Applying Artificial Neural Network in Integrated Assessment of Industry Pollution Source

The environmental and economic attribution of integrated pollution source assessment is the main criterion in total waste load for controlling and allocating.The pollution source management is putting greater concern in the research on integrated pollution source assessment.Several linear pollution source assessment methods have been developed.However due to the multicolinearity and non-linearity relationship among the environment,economy and society,the integrated environmental-economic attribute of pollution source is not fully considered in all these linear methods.Artificial Neural Network(ANN) was applied to assess the integrated environmental-economic attributes of industry pollution source in Shenzhen.Equal-standard pollution load and integrated discriminator index were used for comparison.The results of comparison indicate that ANN method is the best successful one.The effectiveness of ANN method for integrated pollution source assessment is due to its ability of simulating arbitrary nonlinear continuous function and self-adapting.