Towards the Group Formation through Social Norms

This paper examines the decentralized formation of groups within a multiagent normative society. In our case, a group is defined based on the set of social norms used by its members: all the agents using the same set of norms belong to the same social group. In this paper we explore different mechanisms that allow agents to recognize the others as members of a certain social group. Considering as the basic mechanism the one that makes agents interact with other agents without considering the previous interactions and with no communication, three new algorithms have been developed and tested to improve the efficiency of the basic one. These algorithms are: (1) the whitelisting algorithm, which works as a recomender of trusted neighbours; (2) the blacklisting algorithm, whose basic functioning is based on defaming the non-related agents inside a certain social group; and (3) the labelling algorithm, which basically publishes information of the interactions with different agents allowing the rest to access that information. Simulation results are shown, confirming that these algorithms improve the efficiency of the basic one. Finally, we present and discuss some of the weak points of the algorithms presented as well as future improvements.