"奖优罚劣"的动态多指标灰色关联度模型研究

针对动态多指标系统决策特点 ,利用灰色关联分析方法提出了一种新的“奖优罚劣”的动态多指标决策模型。该模型不仅对指标进行初始化处理时使用了“奖优罚劣”原则 ,而且通过构造相关矩阵 ,利用特征根、特征向量的性质 ,构造最优指标体系 ,并由此确定最优指标体系中各指标的权重。通过计算决策对象与理想矩阵、负理想矩阵的关联度 ,进而得到决策对象的排序结果。模型中充分考虑了指标的成长特性 ,将其整个系统的成长特性进行比较 ,作为关联度的一部分。在经济决策的实际问题中应用该模型 ,取得较好的效果

Research on Decision Making Model of the Dynamic Multiple-Attribute System for "Rewarding Good and Punishing Bad"

In this paper, for the decision making of the dynamic multiple-attribute system, an optimal index system is determined, and a new decision making model for "rewarding good and punishing bad" is proposed by using grey relative degree. The model not only uses the principle of "rewarding good and punishing bad" during the initialization, but also forms an optimal attribute system by the construction of the correlation matrix after exploiting the characteristics of the eigenvalue and eigenvector. Accordingly, the wight of each attribute in optimal attribute system is determined. This model has been applied to economic decising making. The results have shown the rationality of the model.