目前,对社会舆论演化的研究主要在静态网络中进行.但是,实际的社交关系网络是有向的动态自适应网络,舆论的演化受网络的影响;反之舆论的演化也可能会导致网络的动态变化.对此问题,提出了一种动态自适应网络的有界信任舆论演化算法,其主要功能是:如果2个个体的观点差值大于一定的信任水平,个体之间的有向连接就以一定的概率断开;如果观点差值小于一定的信任水平,新的连接将以一定的概率重新生成.研究的主要内容是:动态自适应网络的平均观点集合数、观点统一概率、最大集合人数比例、平均步数和平均度5种统计指标的宏观变化规律.实验结果表明,最终的观点集合有3种类型:信任水平较小时,形成多个观点集合;中等的信任水平时,以一定的概率达成统一的共识;信任水平较高时,总能达成统一的共识.动态自适应网络中观点演化的趋势和静态网络相似,但在动态自适应网络模型中,重连概率对观点演化有较大的影响.当重连概率为零时,静态网络模型的上述指标优于动态网络模型;当重连概率增加时,动态网络模型的统计指标逐渐优于静态网络模型.重连概率增加了网络的连边,增强了个体之间的交流,促进舆论的快速形成.实验结果与实际网络舆论演化的情况相符,能够在一定程度上反映和解释现实社会中舆论演化的情况.%Recently, the study on the evolution of social consensus is mainly carried out in static networks.But the actual social network is directional dynamic adaptive network, the evolution of opinion is affected by the networks, and in turn, the evolution of opinion also leads to dynamic changes of the networks.To overcome the problem, a bounded confidence consensus evolution algorithm in dynamic adaptive network is proposed.If the difference of opinions of two individuals is greater than a certain confidence level, directed connection between two individuals may break off with a certain probability;if the difference is less than a certain confidence level, a new connection can rebuild with a certain probability.We study the evolutional rule of five statistical indicators, that is, the average number of opinion clusters, the probability of consensus, the proportion of the largest cluster, the average number of running steps and average degree in a dynamic adaptive network.Simulation results show that, there are three types of opinion clusters: multiple opinion clusters with small confidence level, consensus with a certain probability with medium confidence level, and always consensus with large confidence level.The trend of the opinion evolution of the dynamic adaptive network model is similar with it of the static network model, but the probability of reconnection has a great influence on the opinion evolution of the dynamic adaptive network model.When the reconnection probability is zero, the above mentioned statistical indicators of the static network model is better than that of the dynamic adaptive network model.When the reconnection probability increases, the statistical indicators of the dynamic adaptive network model is better than that of the static network model.The results obtained are consistent with the consensus evolution of actual network, which can reflect and explain the consensus evolution in the real society to a certain extent.
展开▼
