Multidimensional Consensus model on a Barabasi-Albert network

摘要

A Consensus Model according to Deffuant on a directed Barabasi-Albert networkwas simulated. Agents have opinions on different subjects. A multi-componentsubject vector was used. The opinions are discrete. The analysis regardsdistribution and clusters of agents which are on agreement in the opinions ofthe subjects. Remarkable results are on the one hand, that there mostly existsno absolute consens. It determines depending on the ratio of number of agentsto the number of subjects, whether the communication ends in a consens or apluralism. Mostly a second robust cluster remains, in its size depending on thenumber of subjects. Two agents agree either in (nearly) all or (nearly) nosubject. The operative parameter of the consens-formating-process is thetolerance in change of views of the group-members.