Modeling collective adaptive agent design and its analysis in Barnga game

This paper explores the collective adaptive agent that adapts to a group in contrast with the individual adaptive agent that adapts to a single user. For this purpose, this paper starts by defining the collective adaptive situation through an analysis of the subject experiments in the playing card game, Barnga, and investigates the factors that lead the group to the collective adaptive situation. Intensive simulations using Barnga agents have revealed the following implications: (1) the leader who takes account of other players’ opinions contributes to guide players to the collective adaptation situation, and (2) an appropriate role balance among players (i.e., the leader, the claiming and quiet players, which make the most and least number of corrections of the leader’s decision) is required to derive the collective adaptive situation.