基于熵权的区间型多属性决策方法在湖泊水质评价中的应用

针对现有多属性决策方法在水质评价过程中忽视指标在某一等级内效用相同,导致指标间过度补偿,综合评价效果失真的问题,本文考虑将指标时序或空间的多个监测值转化成区间数以降低数据维度,保留时序信息,同时通过定义指标区间到约定等级区间的距离,体现评价强调达标的特点,进而利用熵权法从指标平均水平达标的一致性和变化幅度的趋同性两方面筛选评价作用显著的指标.该方法在指标测度阶段和指标生成阶段压缩决策矩阵维度,简化了评价过程的计算.最后通过对太湖水域的水质数据的评价案例说明了方法的可行性和实效性.

Lake water quality assessment using entropy-based interval number multi-attribute decision making method

The multi-attribute decision making method ignores the same utility for water quality indicator values in the same grade, leading to overcompensation on water quality indicators and distortions in comprehensive assessment of water quality. In this study, we transform a group of sequential or spatial indicator values into interval numbers in order to reduce data dimensions and retain sequential information. Afterwards, we define the distance between the indicator and acceptable grade intervals to reflect the degree of meeting standard requirements, and then employ the entropy method to assess significant indicators by examining consistency of average values in meeting the standards and convergence of variation. The proposed method reduces dimensions of decision matrices in the phases of indicator screening and measurement and simplifies calculation of the assessment process. The validity and feasibility of the proposed method are examined by assessment of Lake Taihu water quality data.