Agent-based modeling of collective decision making in small groups: the role of a formal leader

Abstract

Agent-based modeling is a promising tool for studying collective behavior and has a high potential for increasing the efficiency of group work. Agent-based models can serve as a basis for supporting group work — selecting team members, recommending work protocols, and much more. The article proposes an agent-based model for studying the influence of the characteristics of a formal leader on the collective decision making. The task itself is modeled as the maximization of a utility function, carried out jointly by a group of agents. Each agent can explore the problem space independently and inform other agents about the results of such a search. A consolidated group decision is formed and refined on the basis of individual results obtained by agents, while the result of the formal leader of the group plays a special role. With the help of a computational experiment, the influence of the leader’s own search abilities (“intelligence”), his predisposition to communication (“talkativeness”), and the degree of influence on the effectiveness of the group decision are evaluated. The simulation results showed, in particular, that within the framework of the considered protocol, in the case when the degree of influence of the leader is limited, his low search abilities do not have a significant impact on the effectiveness of the solution found by the group. In the case of high influence, there is a risk of making ineffective decisions. The result itself is in good agreement with practice and is quite intuitive, but more importantly, it shows that fairly simple agent models allow us to study the group work, can be used to evaluate the protocols of such work, to form rules for committing collective action and, in general, to support the effective work of the groups.