How different are ranking methods for fuzzy numbers? A numerical study

被引：87

作者：

Brunelli, Matteo ^{[1
]}

Mezei, Jozsef ^{[2
]}

机构：

[1] Aalto Univ, Syst Anal Lab, Helsinki, Finland

[2] Abo Akad Univ, Inst Adv Management Syst Res, Turku, Finland

来源：

INTERNATIONAL JOURNAL OF APPROXIMATE REASONING | 2013年 / 54卷 / 05期

关键词：

Fuzzy numbers; Ranking methods; Defuzzification; Possibility theory; INTERVAL; SETS; APPROXIMATION; VARIANCE; CRITERIA; AREA;

D O I：

10.1016/j.ijar.2013.01.009

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Ranking fuzzy numbers is often a necessary step in many mathematical models, and a large number of ranking methods have been proposed to perform this task. However, few comparative studies exist and nowadays it is still unknown how similar ranking methods are in practice, i.e., how likely they are to induce the same ranking. In this study, by means of numerical simulations, we try to answer this question. We shall discover that there are some very similar methods as well as some outliers. We end the paper interpreting the results and giving some recommendations on the use of ranking methods. (C) 2013 Elsevier Inc. All rights reserved.

引用

页码：627 / 639

页数：13

共 55 条

[51] Ranking fuzzy numbers based on decomposition principle and signed distance [J].