Group Trust and Group Reputation

Trust and reputation play a vital role in the interactions of agents. Previous researches mostly concern individual agents, while group trust and group reputation are largely ignored. To this end, we present a Computational Model of Group trust and group reputation (CMGTGR) which as well as affects interactions of individual agents. The coverage of CMGTGR extends from individual agents to group agents. Moreover, we increase computational confidence level of trust through using preferences revision and decrease computational error of reputation through using direct observations of an agent. Experimental results show that the CMGTGR is capable of estimating group trust and group reputation and has an advantage over TRAVOS in computational deviation and computational error under inaccurate information.