Abstract: In recent years, with the rapid development of information, communication and technology, the arrival of the era of big data has brought a large number of complex and growing data, and the relationship between decision-makers has become increasingly close. In this context, decision makers interact based on social network relationships, and social network group decision-making will produce more information than traditional group decision-making. Factors such as incomplete information, randomness, and limited rationality of decision makers will lead to deviation between group decision-making results and the actual situation. Therefore, in view of the multi-attribute group decision-making problem in which the decision preference information in social network group decision-making is a qualitative concept of linguistic information, and there is a degree of trust between decision makers, this paper proposes a social network group decision-making method based on cloud model and PageRank algorithm, and develops a social network cloud clustering algorithm for the existing cloud clustering algorithm. The specific research contents are as follows: First, the uncertain expression of decision-maker preference information is studied. The golden section generation method based on cloud model can fully express the uncertainty of decision by transforming the language preference information given by decision makers. Then, the construction method of social network topology is studied. The whole social network is regarded as a topological graph, the decision-maker is regarded as a node, and the relationship between the decision-makers is regarded as an edge. The PageRank algorithm is used to calculate the trust degree of each decision-maker in the whole social network topological graph, and the trust degree is converted into the weight of the decision-maker for decision information aggregation. Secondly, the cloud clustering algorithm is studied. Cloud clustering is a clustering method based on cloud model. Most of the existing cloud clustering algorithms only consider the similarity of the initial preference information of decision makers. Due to the complexity of large group decision making in social network, it is obviously unreasonable to consider only the utility of a single factor, which is not suitable for the clustering of large group decision members in the context of social network. This paper improves the existing cloud clustering algorithm to apply to social network group decision-making. Finally, the effectiveness and rationality of this method are verified by an example of emergency decision-making and comparative analysis. The main steps of the method proposed in this paper are as follows: First, we translate language information into cloud model. Secondly, PageRank algorithm is used to calculate the weight of each decision maker based on the social network topology. Then, the social network cloud clustering algorithm is used to divide decision makers into several sub-clusters and calculate the weight of sub-clusters. Finally, the cloud model is integrated into the scheme cloud, and the random simulation technology is used to obtain the score of each scheme cloud, and the preference ranking of the schemes is given. To sum up, the contributions of this paper can be summarized into two points:(1)This paper quantifies the social relationship based on the directed relationship of individual decision-makers in the social network, then uses PageRank link analysis algorithm to simulate the social network of decision-makers to determine the individual's trust degree, and converts the calculated trust degree into weight, which is more in line with the objective reality, and enriches the research on the interaction of decision-makers in the social network. (2)This paper considers the complexity of decision groups in the context of social networks, integrates the utility of decision makers' trust and the utility of decision information, applies cloud clustering algorithm to social network group decision-making, and expands the universality of cloud clustering algorithm. Of course, the method proposed in this paper is not limited to the field of social network group decision-making, but also can be applied to a variety of decision-making scenarios, such as large group decision-making, emergency group decision-making, multi-attribute decision-making, and so on. At the same time, it can also provide some guidance for decision makers when making decisions, provide more scientific and reasonable basis for decision makers, and improve the quality and efficiency of decision-making, so as to improve the government, the overall operation level of enterprises and social organizations. In the follow-up research, finding a reasonable threshold of cloud clustering algorithm and the construction method of social network are also worthy of further research.

Key words: group decision making, cloud model, PageRank algorithm, trust relationship, social network

摘要： 近年来,随着信息、通信和技术的快速发展,决策者之间的关系也由此变得越来越紧密。在此背景下,决策者基于社会网络关系存在交互,社会网络群体决策会产生比传统群体决策更多的信息,信息不完全性、随机性以及决策者有限理性等因素都会导致群体决策结果与实际情况产生偏差。针对社会网络群体决策中决策偏好信息为定性概念的语言信息,且决策者之间存在信任度的群体决策问题,本文提出了一种基于云模型和PageRank算法的社会网络群决策方法并针对现有的云聚类算法拓展出了社会网络云聚类算法。首先,将语言信息转化为云模型;其次,基于社会网络拓扑图利用PageRank算法来计算每个决策者的权重;接着,使用社会网络云聚类算法将决策者分成几个子聚集并求出子聚集的权重;最后把云模型综合成方案云,利用随机模拟技术获取各个方案云的评分,给出方案的排序,并通过对比分析验证了该方法的有效性。

关键词: 群决策, 云模型, PageRank算法, 信任关系, 社会网络