群决策中从众与迷信权威现象的动态网络分析  

Dynamic network analysis of conformity and blind faith in group decisions

作　　者：赵峰[1,2] 王书宁[1] 

机构地区：[1]清华大学自动化系,北京100084 [2]海军装备研究院,北京100161

出　　处：《清华大学学报（自然科学版）》2013年第11期1617-1622,共6页Journal of Tsinghua University(Science and Technology)

摘　　要：针对群决策中容易出现从众和迷信权威现象,将人员集合和观点集合抽象为观点-人员2-模网,并分别针对这2种现象构建初步的动态网络分析算法,进行由聚类分析、对应分析等方法组成的定性定量综合分析,探索群决策过程中人员间观点支撑关系的动态变化情况,并根据2种算法的混合实验,对群决策过程进行改进。仿真实验初步表明:从众现象中观点的入度分布符合幂律分布,迷信权威现象无明显统计规律;研讨中,尽可能靠后安排权威发言、靠前安排实践工作经验少者与活跃者,或是在权威发言必须靠前情况下,尽可能靠前安排活跃者发言,有助于减少从众与迷信权威现象的影响,提高群决策效率。Conformity and the effects of blind faith in authorities in group decisions were analyzed using a 2-mode network model for the opinions and the people. Two algorithms are given for the dynamic network analysis method for the conformity and the blind faith in authorites. The algorithms give qualitative and quantitative procedures which include correspondence and cluster analysis methods for static and dynamic network analyses. The model improves group decision procedures based on experiments using the two algorithms. Case studies show that the conformity statistics conform to a power law variation, while participant＇s blind faith in authorities does not have any obvious corelation. Authoritive speakers should be scheduled at the end of the meeting with people with less experience and more talkative people in the beginning of the meeting. If the authorities must speak first, the talkative people should still speak early in the meeting. This scheduling reduces the impact of conformity and blind faith and improves the efficiency.

关 键 词：群体决策 社会网络分析 从众 迷信权威 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]