基于正态分布区间数的多属性决策方法

Approach of Multi-attribute Decision-making Based on Normal Distribution Interval Number

针对属性权重完全未知且属性值为区间数的多属性决策问题,提出了一种基于正态分布区间数的决策方法。首先基于正态分布的基本性质,针对两区间数存在相互交叉部分的情形,提出了正态分布下区间数之间相互比较的可能度概念,给出了区间数比较的一种新方法。然后根据离差最大化原理给出了确定属性权重的一种新方法,并给出了一种区间数的多属性决策方法。最后通过一个算例说明了此方法的实用性和简易性。

For multiple decision making problems, in which the attribute weight values are completely unknown and the attribute values are intervals,a decision making approach is given based on normal distribution intervals.First based on the basic properties of normal distribution,for two intervals between which there are crossed parts,the concept on the possibility for comparing two intervals is proposed and a method is given for comparing intervals.Then the weights of attributes are obtained based on the maximizing deviation,and a method of interval multi-attribute decision-making is obtained.Finally an example shows the effectiveness and facility of this method.