基于证据推理和广义Shapley值的扩展概率语言多属性群决策方法

扩展概率语言词集通过语言变量概率分布的调整能够转化为多种语言信息表示模型，是语言变量、不确定语言信息、扩展犹豫模糊语言词集、分布语言评估信息、概率语言词集等的一般化，具有较强的通用性和实用性，是处理不确定性信息的重要工具。鉴于此，本文针对扩展概率语言环境下的多属性群决策问题，提出基于证据推理和广义Shapley值的多属性群决策方法。首先，提出扩展概率语言词集的定义和相关基础理论。其次，将广义Shapley值和证据推理相结合用于专家信息融合，并将广义Shapley值和TODIM方法相结合用于备选方案排序。再次，提出基于灰色关联法的权重确定模型来处理专家/属性权重部分未知的情况。最后，以绿色供应商选择为例进行分析，通过对比分析验证所提方法的有效性和优越性。

Multiple Attribute Group Decision-Making Method Based on Evidential Reasoning and Generalized Shapley for Extended Probabilistic Linguistic Term Set

Extended probabilistic linguistic term set (EPLTS) can be transformed into several linguistic information representation models by adjusting the probability distribution, such as uncertain linguistic variable, linguistic distribution assessment, extended hesitant fuzzy linguistic term set, probabilistic linguistic term set and so on. EPLTS can describe the original evaluation information more fully and objectively so that it is an important tool for dealing with uncertain information. In view of this, with respect to multiple attribute group decision making problems under extended probabilistic linguistic environment, this paper proposes an extended probabilistic linguistic group decision making method based on evidential reasoning and generalized Shapley value. Firstly, the concept and relative theories of extended probabilistic linguistic term set are given. Secondly, the combination of generalized Shapley value and evidential reasoning is used to aggregate evaluation information from multiple decision makers, and the combination of generalized Shapley value and classical TODIM method is utilized to rank multiple alternatives. Thirdly, the weight determination models based on grey relational method are proposed to calculate weights of decision makers and weights of attributes respectively. Finally, taking the green supplier selection as an example, the effectiveness and superiority of the proposed method are verified by comparative analysis with the existing methods.