多种改进模糊概率模型对比研究

Comparative Study of Multi-Improved Probabilistic Fuzzy Model

本文以超标法、层次分析法、信息熵法三种方法分别计算权重的模糊概率模型,依据2003-2013共10年的黄河水系6个国家监测点的水质监测数据,对黄河华北西部段水质进行评价。与传统方法相比,模糊概率评价法用概率处理污染物浓度,用隶属度描述水质等级界限,使评价结果与实际情况更接近;空间评价分析结果显示除宁夏吴忠新墩断面有上游水质优势外,其他断面水质情况只与附近环境相关而与其所处流域位置无关。三种模型结合单因子评价模型在结果和原理上进行比较分析发现,强调了不同类指标因子权重值差异及客观性的层次分析法,是最适合模糊概率模型的权重计算方法,其评价结果更为客观、可靠,可广泛用于河流水体的综合水质评价。

The multi-improved probabilistic fuzzy models calculate the weights by overweight method,information entropy and analytic hierarchy process. The water quality of the Yellow River in the western region of Northern China was assessed by the models based on the monitoring data of six national monitoring sections during 2003 to 2013. Probabilistic fuzzy model uses the probability to calculate pollutant concentrations and the membership degrees to describe the boundaries of water quality level. The model has a better evaluation results which are closer to the actual situation than the traditional models. The paper makes comparative analysis to the multi-improved probabilistic fuzzy model and single factor evaluation model in consequent and principle. The analytic hierarchy process method,which prevents the uncertainties in determining the index weights,is proved to be the most suitable weight calculation method of the probabilistic fuzzy model. The results present reasonable and reliable; the model can be widely used in all kinds of comprehensive evaluation problems.