This paper considers a group decision-making mechanism for a group of social agents with ambiguous interactions. First, a fuzzy inference approach is introduced to describe bounded confidence based interaction rules among social agents for a certain object in a social network platform. Second, an influence graph is introduced to model the communication network topology associated with the group of agents. A fuzzy inference based opinion dynamics model is built when the defuzzified interaction weights among agents are used in the opinion update scheme for each agent. Third, the patterns of the collective final opinions are analyzed under the proposed fuzzy opinion dynamics model. The fuzziness of opinion gaps and interaction weights among agents are respectively investigated in the collective opinion evolution. Simulation results show the influence of fuzzy values of opinion gaps and interaction weights on the pattern of the collective final opinions and reveal the quantitative relationships of opinion gap and interaction weights between the original HK opinion dynamics model and the fuzzy HK opinion dynamics model.

本文考虑了一组具有模糊交互的社会代理的群体决策机制。首先，引入模糊推理方法来描述社交网络平台中某个对象的社交代理之间基于有界置信度的交互规则。其次，引入影响图来对与代理组关联的通信网络拓扑进行建模。当在每个代理的意见更新方案中使用代理之间的去模糊化交互权重时，建立了基于模糊推理的意见动态模型。第三，在提出的模糊意见动态模型下分析了集体最终意见的模式。在集体意见演化过程中分别研究了意见差距的模糊性和代理之间的交互权重。