Large-scale group decision-making (LSGDM) under uncertainty modelled by comparative linguistic expressions based on a hesitant fuzzy linguistic term set (HFLTS) has recently attracted the interest of many researchers and research, due to the necessity of its function in LSGDM, and the challenges it faces such as the managing of the scalability problem, uncertainty of experts’ opinions and dealing with polarized conflicting opinions. To smooth out such discrepancies and obtain agreed solutions Consensus Reaching Processes (CRPs) for LSGDM have been applied, in which experts are grouped into sub-groups according to the closeness of their opinions to deal with scalability. However, most CRPs for LSGDM are driven by a majority rule, in which larger sub-groups, where there might be internal disagreements, lead the consensus. In such processes, the internal disagreements can produce unsatisfactory solutions. Consequently, the majority view should be complemented by additional mechanisms that also measure the strength of the sub-groups’ opinions. A good measurement of such strength is the cohesion among the sub-group members. Therefore, in this paper, a new cohesion measure for HFLTS based on restricted equivalence functions for measuring the experts’ sub-group cohesiveness is introduced to drive the consensus process together the majority and thus reduce the impact of internal disagreements risen in majority driven CRPs. It is then integrated in a new cohesion-driven CRP approach based on LSGDM to deal with comparative linguistic expressions based on HFLTS. An experimental analysis on different large scale scenarios will show the performance and importance of cohesion in consensus based LSGDM.

基于不确定性模糊语言术语集（HFLTS）的比较语言表达式建模的不确定性下的大规模群体决策（LSGDM）最近吸引了许多研究人员和研究人员的关注，因为它必须在LSGDM中发挥作用，并且它面临的挑战包括可伸缩性问题的管理，专家意见的不确定性和对立的，相互矛盾的意见的处理。为了消除此类差异并获得商定的解决方案，已应用LSGDM的共识到达过程（CRP），其中专家根据其意见的接近程度将其分组，以处理可伸缩性。但是，大多数LSGDM的CRP是由多数规则驱动的，在该规则中，可能存在内部分歧的更大的子群体将导致共识。在这样的过程中 内部分歧可能会产生不令人满意的解决方案。因此，多数意见应辅之以也能衡量小组意见观点的其他机制。这种强度的一个很好的衡量标准是小组成员之间的凝聚力。因此，在本文中，引入了一种基于有限等价函数的HFLTS内聚度量，用于衡量专家的子组内聚性，以共同推动多数成员的共识过程，从而减少由多数驱动的CRP引起的内部分歧的影响。然后将其集成到基于LSGDM的新的内聚驱动CRP方法中，以处理基于HFLTS的比较语言表达。