BMW-TOPSIS: A generalized TOPSIS model based on three-way decision
BMW-TOPSIS: A generalized TOPSIS model based on three-way decision
复制标题
DOI：
10.1016/j.ins.2022.06.018
复制
发表时间：
2022-06
期刊：
影响因子：
--
通讯作者：
Yumei Wang;Peide Liu;Yiyu Yao
中科院分区：
文献类型：
--
作者：
Yumei Wang;Peide Liu;Yiyu Yao
TOPSIS（Technique for Order Preference by Similarity to Ideal Solution）采用一对正理想解和一对负理想解作为两个参考点对一组决策方案进行排序。在某些情况下，到两个极端参考点的距离的折衷可能不一定有意义。受三人思考的三路决策理论的启发（例如，两个相反的极点和一个第三中间点），本文通过增加一个第三中间参考点来推广经典的TOPSIS。我们使用一个共同的设置系统地调查基于参考点的TOPSIS风格的多标准决策方法。特别地，我们检查三类方法：a）基于最佳参考点的模型（即，B-TOPSIS）和基于最差参考点的模型（即，W-TOPSIS），B）经典的基于最佳和最差参考点的模型（即，BW-TOPSIS），以及c）新的基于最佳、平均和最差参考点的模型（即，BMW-TOPSIS法）。这三类分别是单向TOPSIS、双向TOPSIS和三向TOPSIS。在单因素和双因素TOPSIS的基础上，给出了三因素TOPSIS的两种具体方法。实验结果与现有的TOPSIS方法进行比较，表明BMW-TOPSIS模型是可行且有效的。
TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) uses a pair of a positive ideal solution and a negative ideal solution as two reference points to rank a set of decision alternatives. In some situations, a trade-off of the distances to the two extreme reference points may not necessarily be meaningful. Inspired by the theory of three-way decision as thinking in threes (e.g., two opposite poles and a third middle), in this paper we generalize the classical TOPSIS by adding a third middle reference point. We use a common setting for investigating systematically reference-point-based TOPSIS-style multi-criteria decision-making methods. In particular, we examine three classes of approaches: a) a best reference point based model (i.e., B-TOPSIS) and a worst reference point based model (i.e., W-TOPSIS), b) the classical best and worst reference points based model (i.e., BW-TOPSIS), and c) a new best, mean, and worst reference points based model (i.e., BMW-TOPSIS). The three classes are one-way TOPSIS, two-way TOPSIS, and three-way TOPSIS, respectively. Based on one-way and two-way TOPSISs, we give two specific methods of three-way TOPSIS. The experimental results, compared with the existing TOPSIS methods, show that the BMW-TOPSIS model is feasible and effective.