研究语言偏好信息下的群决策问题.定义了反映群体共识的两个测度指标,分别反映群体内所有专家的一致性水平及专家的个人观点与群体观点的分歧程度;基于共识测度指标构建一种语言标度的颗粒优化模型,提出了求解语言标度颗粒最佳分界点的改进PSO算法,并给出一种对方案排序进行择优的群决策方法.最后,通过一个算例说明了所提出方法的可行性和有效性. The paper studies the problem of group decision-making under language preference information, and defines two measure indexes that reflect the group consensus, which respectively reflect the degree of consistency of all experts in the group and the degree of disagreement between the expert’s viewpoints and the group’s viewpoints. Based on the consensus measures, A new PSO algorithm is proposed to solve the optimal demarcation point of linguistic scale particles and a group decision method is given to optimize the ranking of programs.At last, an example is given to illustrate the proposed method Feasibility and effectiveness.