Comparing probabilistic linguistic term sets (PLTSs) is quite essential in solving PLTS-expressed multi-attribute group decision-making problems (PLTS-MAGDM). Researchers have designed various comparison measures to obtain the rank of PLTSs. However, most of the existing PLTS comparison measures need additional tedious adjustments before conducting a specific computation. Besides, these measures do not adequately consider the effects of the semantics of the basic linguistic term set and the probabilistic distributions. This paper proposes a new preference degree for g-granularity probabilistic term sets (g-GPLTSs) to overcome the two shortcomings simultaneously by integrating the effect from basic linguistic terms and probabilistic distributions without any adjustment. Moreover, the g-GPLTS preference degree also shows the extended adaptability for comparing PLTSs with unbalanced semantics. Based on the newly proposed preference degree, we construct a useful min-conflict model to solve PLTS-MAGDM with a large number of experts expressing the three-way primary grading. Finally, an illustrative example concerning software supplier selections, followed by the comparative analysis, is presented to verify the feasibility and effectiveness of the proposed method.

比较概率语言术语集（PLTS）对于解决PLTS表达的多属性小组决策问题（PLTS-MAGDM）至关重要。研究人员设计了各种比较方法来获得PLTS的排名。但是，大多数现有PLTS比较措施在进行特定计算之前都需要进行其他繁琐的调整。此外，这些措施没有充分考虑基本语言术语集的语义和概率分布的影响。本文提出了一种新的偏好度克-granularity概率术语集（克-GPLTSs），以通过积分从基本语言项和概率分布的效果而没有任何调整同时克服这两个缺点。而且，g -GPLTS偏好度还显示了用于比较具有不平衡语义的PLTS的扩展适应性。基于新提出的偏好度，我们构建了一个有用的最小冲突模型，该模型可以使用大量表示三向一次评分的专家来解决PLTS-MAGDM。最后，给出了一个有关软件供应商选择的说明性例子，随后进行了比较分析，以验证所提出方法的可行性和有效性。