水资源配置方案综合评价中的信息损失测度研究

水资源配置的多属性特征使得在方案综合评价中依据不同的聚合方法,能够获得不完全一致的方案排序结果。本文从水资源配置方案评价的决策矩阵及方案排序结果中的信息量大小入手,利用信息熵理论和斯皮尔曼等级相关系数提出衡量水资源配置方案综合评价中信息损失的香农斯皮尔曼测度。并构造由2种权重(熵权法权重、证据理论权重)与3种评价模型(简单加权法、模糊优选法、TOPSIS)组合而成的6种综合评价模型。最后,将香农斯皮尔曼测度运用到天津市水资源配置方案综合评价过程中的信息损失的衡量,并与文献中的综合评价模型所得结果进行比较。结果表明,采用证据理论权重-模糊优选评价模型所得评价结果的绝对信息损失和相对信息损失最小。通过水资源配置方案综合评价中的信息损失测度研究,能够使得水资源配置方案综合评价过程更加透明,并为决策部门选取理想的水资源配置方案综合评价方法提供决策支持。

Research on The Measure of Information Loss in Composite Evaluation of Water Resource Allocation Schemes

The characterof multiple attributes in water resource allocation makes the ranking of schemes incompletely consistent by using different aggregating methods. This paper discusses the quantity of information loss from given decision-making metrics to ranking results. Then information entropy theory and Spearman’s rank correlation coefficient are adopted to measure information loss in evaluation of water resource allocation schemes and the Shannon-Spearman measure is proposed. Six composite evaluation models are constructed by combining two weight sets(weight based on entropy and weight based on evidence theory)and three evaluation models (namely sample additive weighting, fuzzy selection method and TOPSIS). Finally, the Shannon-Spearman measure is adopted to calculate information loss in the process of composite evaluation of Tianjin water resource allocation schemes. And composite evaluationresultsare compared with a given composite evaluation model in reference. The results show that the result from combining weight based on entropy and fuzzy selection method obtains the least absolute/relative information loss in seven results. Through the measure of information loss in evaluation of water resource allocation schemes, it could make evaluation process more transparent and provide policy-making departmentwith decision support for choosing ideal composite evaluation method of water resource allocation schemes.