【摘要】 针对现有不确定语言信息集结算子难以兼顾数据个体间关联性与整体均衡性的问题，定义新的不确定语言幂Bonferroni平均算子和新的不确定语言加权幂Bonferroni平均算子。新算子通过Bonferroni平均算子的交叉运算来体现数据间的关联性，并引入幂平均算子的支撑度系数来刻画数据的整体均衡性特征。此外，新算子具有置换不变性、幂等性、介值性和单调性等性质，在此基础上提出一种多属性决策方法。选取某企业人才遴选问题进行算例分析，研究结果表明，该方法既能体现指标间关联性，也能削弱部分异常属性值的不利影响。新算子对不确定语言信息集结算子进行了有益的补充和完善，所提新方法对于解决复杂不确定语言环境下的多准则决策问题提供了方法论支撑。更多还原

【Abstract】 Multiple attribute decision making is a prominent area of research in normative decision theory. This topic has been widely discussed and studied. Considering that experts may have vague knowledge about the preference degree of objective things, uncertain linguistic variables are often used for decision makers as the evaluation information in some complex multiple attribute decision-making problems. For the sake of aggregating the uncertain linguistic evaluation information in multiple attribute decision-making problems, some uncertain linguistic information aggregation operators are developed and designed. However, it is worth noting that most of the uncertain linguistic information aggregation operators do not take into account the interrelationship between data and the overall balance of data simultaneously, which is unreasonable for the decision results in uncertain linguistic multiple attribute decision making problems. Therefore, it is necessary to pay attention to this issue. In order to better solve this issue, we combine uncertain linguistic variables, power averaging operator, and Bonferroni mean operator to define the uncertain linguistic power Bonferroni mean(ULPBM) operator and the uncertain linguistic weighted power Bonferroni mean(ULWPBM) operator, which is able to make the decision results more accurate. The ULPBM operator and ULWPBM operator have some characteristics in the process of aggregating uncertain linguistic variables. On the one hand, two new operators utilize the crossover operation of the Bonferroni mean(BM) operator to represent the interrelationship between data. On the other hand, two new operators use the support degree coefficient of the power averaging(PA) operator to denote the overall balance of data. At the same time, the commutativity, idempotency, boundary, and monotonicity of these operators are investigated, which validates that these operators are feasible. Based on the above analysis, the ULWPBM operator is used for solving uncertain linguistic multiple attribute decision making problems, and a novel uncertain linguistic multiple attribute decision making method with the ULWPBM operator is put forward. The talent selection is an important part for the development of enterprises, and the enterprises that own some excellent talents will make more profits and gain more space for survival. In this scenario, how to effectively find the most appropriate talent from several candidates has become a significant problem for every enterprise. However, there are a lot of factors that affect the capacity of talents, including ideology and morality, work attitude, work style, cultural level, knowledge structure, etc. Moreover, the experts often use uncertain linguistic variables when they evaluate the preference of these factors with regard to the enterprise talent selection. It is obvious that the enterprise talent selection is a classical uncertain linguistic multiple attribute decision-making problem. Therefore, we take an example with regard to the enterprise talent selection to prove the feasibility and effectiveness of the proposed method. The research results show that not only can the proposed method better solve the interrelationship between attributes, but the proposed method can also effectively alleviate the adverse influence of abnormal attribute values. The main contributions of this paper are that we define the ULPBM operator and ULWPBM operator, and then we put forward a novel uncertain linguistic multiple attribute decision making method with ULWPBM operator. It should be noted that two new operators are the supplement and improvement of uncertain linguistic information aggregation operators, and the proposed method can give some references for solving real-world complex decision-making problems. In the future, the power Bonferroni mean operator could be extended to hesitant fuzzy linguistic preference relations and probabilistic linguistic environments. In addition, the uncertain linguistic multiple attribute group decision making approach with ULWPBM operator is also an interesting research topic, which needs further investigation.更多还原