Environmental Science

CLC number: X82

Using TOPSIS Model with Modified Entropy Weight for Water Quality Assessment in Guizhou Province, China

基于改进熵权的TOPSIS模型在中国贵州省水质评价中的应用

¹ IRI Consulting Group, Beijing 100037, China
² School of Political Science International Relations, Tongji University, Shanghai 200092, China

Received: 3 July 2024

Abstract

Although the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) model has been widely applied in water quality assessment by numerous studies, several common limitations remain unresolved. Specifically: 1) Subjective elements in methods such as fuzzy theory and the analytic hierarchy process (AHP) may distort evaluation outcomes; 2) The utilization of raw sample data is insufficient when constructing evaluation matrices; 3) The traditional entropy weight method in TOPSIS merely reflects statistical characteristics of the final matrix while neglecting richer information embedded in raw datasets. To address these issues, we proximate probability distribution function of various indicators by using cubic spline interpolation and fully exploit information in the existing massive sample data. In this paper, the entropy weight method is enhanced based on the concept mentioned above and integrated with TOPSIS model to construct a novel evaluation model. Furthermore, the experimental analysis using wastewater monitoring data from Guizhou Province, China, verifies its practicality, and its results provide valuable references for local water environmental management.

摘要

诸多研究已将TOPSIS模型应用于水质评价，但仍有共性缺陷亟待解决。具体表现为：带有主观性要素方法，如模糊理论、层次分析法等可能会扭曲评估结果；构建评价矩阵时原始样本数据利用不足；传统TOPSIS中的熵权法仅能反映最终矩阵的统计特性，却忽视了原始数据蕴含的更丰富信息。为解决上述问题，本文利用三次样条插值法近似拟合各指标的概率分布函数，深度挖掘既有海量样本数据的信息价值，并在此基础上对熵权法进行改进，与TOPSIS模型相结合，构建了一种新型评价模型。最后，以中国贵州省废水监测数据为例验证了该模型的可行性，其结果为当地区域水环境治理提供了一定的参考价值。