一种基于最小偏差的联系数多属性决策方法

Minimum—deviation—based Method for Multiple Attributes Decision Making with Connection Number

针对属性权重和决策矩阵均为区间数的多属性决策问题,提出了一种基于最小偏差的联系数排序方法.构建了一个区间数上下界偏差函数,并根据偏差越小越好的原则将区间数权重转化为确定性的权重向量.将区间决策矩阵转化成联系数决策矩阵,求出满意度最高的联系数,再根据各个方案与最高满意度的比较得出差异度,进而排序.并通过例证说明该方法的可行性与合理性.

A minimum--deviation--based method is proposed for multiple attribute decision making problems in which both the attribute weights and the elements of decision matrix are interval numbers. By establishing a deviation function of interval bounds and applying the small deviation principle, the interval evaluations are transformed into connection numbers. The interval decision matrix is then transformed into connection number matrix, and the highest satisfaction connection number is calculated. Programs are aeeordingly sorted by the differenee degrees obtained by comparing the related highest satisfaction connection numbers. The effeetiveness of the proposed method is illustrated by an example.