基于理想关联度的不确定多属性决策方法

针对只有部分属性权重信息且属性值以区间数形式给出的不确定多属性决策问题,提出了一种逼近理想关联度的决策分析方法。首先改进了文4]给出的区间数决策矩阵的规范化方法;然后提出了利用期望-方差区间数排序方法求解理想最优方案;最后依据关联系数矩阵给出了属性权重信息不完全的区间数多属性决策问题的求解方法,其核心是求解线型规划得到属性权重,进而根据各个方案与理想最优方案的综合关联度大小进行排序。特别地,给出了属性权重完全未知的简洁方法,文后的实例验证了方法的可行性和有效性。

A Method for Uncertainty Multiple Attribute Decision Making Based on Ideal Incidence Degree

The uncertain multiple attribute decision-making problem is investigated,in which the information about attribute weights is partly known.An analysis method for similarity to ideal incidence degree is proposed.In the method,firstly the ideal optimization scheme is derived by the expectation-variance ranking method;secondly incidence coefficient matrix is obtained based on the grey incidence degree theories;thirdly a linear programming model is developed to achieve attribute value as a key step;finally the priority of schemes is gotten by the size of synthesis incidence degree.Especially,the fact that the information about attribute weights is completely unknown is studied.The example shows the feasibility and effectiveness of the proposed method.