中文摘要：

考虑属性评估值、专家权重以及属性权重均以语义信息形式给出的纯语义多属性群决策问题，提出了一种基于二元语义的纯语义多属性群决策方法．采用二元语义表达模型处理专家给出的纯语义评估信息；提出了融合加权调和平均算子和有序加权平均算子的二元语义混合调和平均（2一tuplehybridharmonicaver—age，2THHA）算子，解决属性评估值以及属性权重的群决策问题．针对传统灰色关联分析方法仅考虑单一参考序列以及没有考虑属性间非线性补偿关系的问题，提出了改进灰色关联分析，同时考虑正理想参考序列和负理想参考序列计算候选方案在每个属性点上的灰色关联度；考虑属性间的非线性补偿，得到了方案的总体评估值．所提方法最后被应用于产品服务系统配置方案选择问题．

英文摘要：

Consider the multi - criteria group decision making （MCDM） problem with criteria values, expert weights and criteria weights all being expressed by pure linguistic information. A MCDM approach under pure linguistic information based on 2 - tuple linguistic model is given. The 2 - tuple linguistic model is used to deal with the pure linguistic information from experts. By combining the weighted harmonic average （WHA） operator with the ordered weighted average （OWA） operator, a 2 -tuple hybrid harmonic average （2THHA） operator is proposed to solve the group decision making problem of determining evaluation values and weights of criteria. As the traditional grey rela- tional analysis （GRA） method only considers a single reference sequence and does not consider the non - linear com- pensation relationships among criteria, an improved GRA, which can simultaneously take into account the positive and negative ideal reference sequences, is put forwarded. Considering the non - linear compensation relationship among criteria, the overall evaluation information is obtained. The proposed method is applied to solve the configuration design selecting problem to product service systems.