Theoretical analysis and empirical study results all show that there are situations in reality where observed data are not random variables. Thus, decision-making criteria based on probability theory are not suitable for people to make decisions. This paper proposes an uncertain dominance based on uncertainty theory to offer an alternative decision-making criterion for such situations. The paper first defines a new criterion of first- and second-order uncertain dominance, then proves some necessary conditions of them based on uncertainty theory. Some sufficient and necessary conditions of the first- and second-order uncertain dominance are given when uncertain variables are all normal or linear uncertain variables. In addition, the paper proves the link between the uncertain dominance criterion and the expected utility criterion and shows that the first-order uncertain dominance is suitable for all people to make decisions and the second-order uncertain dominance is suitable for risk-averse people to make decisions.

理论分析和实证研究结果都表明，现实中存在观测数据不是随机变量的情况。因此，基于概率论的决策标准不适合人们进行决策。本文提出了一种基于不确定性理论的不确定优势，为这种情况提供了一种替代决策标准。本文首先定义了一阶和​​二阶不确定支配的新判据，然后基于不确定性理论证明了它们的一些必要条件。给出了当不确定变量均为正态或线性不确定变量时，一阶和二阶不确定占优的一些充分必要条件。此外，