中文摘要：

分析了以往供应商选择方法的不足,如供应商属性指标的权重不能得到客观合理的精确量化,供应商属性值的决策矩阵受到主观影响较大,且定性指标很难精确化到某一数值等.考虑到供应商评价的决策信息通常由定值型数据和区间型数据2种数据组成,提出了一种基于混合型数值决策信息的供应商选择方法,利用基于相离度和可能度的偏差最大化多属性决策的改进方法进行供应商选择评价,给出了算法的具体步骤,并用实际算例予以说明.

英文摘要：

In this paper, the disadvantages of former supplier evaluation method were analyzed. These lacks include that the weight of attributes can＇t be measured objectively and precision, the attribute value matrix is affected by subjective decision and the qualitative indexes is difficult to be precision. Considering to the decision information consisted by the fixed value and interzone value, a supplier evaluation method based on the mixed data decision information was put forward. Using improved maximum deviation method based on deviation degree and possibility degree to ewaluate supplier, the detail arithmetic steps and its application case were given. The results show that the new method can reflect the decision data reality; the uncertainty of decision data was eliminated clearly.