How to process local and global consensus? A large-scale group decision making model based on social network analysis with probabilistic linguistic information
How to process local and global consensus? A large-scale group decision making model based on social network analysis with probabilistic linguistic information
复制标题
如何处理本地和全球共识？
DOI：
10.1016/j.ins.2021.08.014
复制
发表时间：
2021-08-18
影响因子：
8.1
通讯作者：
Tang, Ming
中科院分区：
文献类型：
--
作者：
Liao, Huchang;Li, Xiaofang;Tang, Ming
随着社会和技术范式的快速发展，大规模群体决策成为一个新兴的话题。在传统的群决策方法中，通常假设所有的专家都是独立的。然而，随着社会化媒体的发展，专家之间往往会因为学术关系、工作关系或共同兴趣等原因而产生一定的联系和聚集。在这些情况下，专家不再是独立的个人。为了解决这个问题，本研究引入了一个基于社会网络分析的大规模群体决策模型。在该模型中，专家可以提供对其他专家的信任值。由于大规模群体决策问题的规模和复杂性，降维，即使用社区检测分类到本地社区的专家，被认为是必要的。基于这个过程，整个集团可以分为两个层次。第一层是包含所有社区的全球网络，第二层是社区内的本地网络。本研究发展一个模型，以解决大规模的群体决策问题，同时考虑本地和全球的共识，在两个层次。该模型允许专家使用概率语言偏好关系来表达他们的认知复杂性评价信息。一个说明性的例子来显示所提出的模型的实用性。(c)2021爱思唯尔公司All rights reserved.
With the rapid development of societal and technological paradigms, large-scale group decision making becomes an emerging topic. In conventional group decision making methods, it is often assumed that all experts are independent. However, with the expansion of social media, experts usually have some relationships and get together for some reasons such as the academic relationship, working relationship or common interests. In these cases, experts are no longer independent individuals. To address the issue, this study introduces a large-scale group decision making model based on the social network analysis. In this model, experts can provide trust values on other experts. Due to the scale and complexity of the large-scale group decision making problems, the dimensional reduction, which uses community detection to classify experts into local communities, is deemed essential. Based on this process, the whole group can be divided into two layers. The first layer is the global network containing all communities, and the second layer is the local network within a community. This study develops a model to address large-scale group decision-making problems considering the local and global consensus in two layers simultaneously. This model allows experts to use probabilistic linguistic preference relations to express their cognitive complex evaluation information. An illustrative example is presented to show the usefulness of the proposed model. (c) 2021 Elsevier Inc. All rights reserved.