基于二元不完全偏好的群决策方法

Aggregation approach to bivariate of incomplete preference information

分析了群决策过程中不完全区间数互反判断矩阵和不完全区间数互补判断矩阵的集成方法。基于决策者给出的允许偏差,定义了群决策满意度隶属函数,建立了求解群体偏好满意程度最大化的权重模型,提出了群体偏好一致性程度的衡量指标,研究了对一致性程度较差的决策者判断进行交互调整的方法。为解决模型存在多组最优解以及因决策者判断误差波动造成的决策误导问题,建立了群体偏好权重分布范围的估计模型。通过算例分析,证明了方法的有效性及实用性。

An aggregation approach to the incomplete complementary comparison matrix is analysed. First, the decision making is defined based on the allowable error interval reciprocal comparison matrix and the interval overall satisfaction membership function of the group given by the decision maker respectively. Following that, a weight model is suggested for finding the maximal degree of the group preference satisfaction. Based on the suggested model, the consistency degree of the group preference is defined. Then, an alternation approach is proposed to improve the judgment of the low consistency degree of the decision-maker. Finally, the weight value distributing region model to estimate the group preference is developed to solve the multi-optimization problem and reduce the misleading decision-making owing to the judgment error. The example indicates the effectiveness and practicablity of the approach.