Social norm have often been used to promote cooperation and coordination and to avoid unnecessary con icts in multiagent systems. Furthermore, interaction between agents is ususally based on social network topology, that is, how individuals are connected. Thus, many studies are dedicated to the emergence of norms in these complex networks, but few of them focused on dynamically changing networks. This paper extends our previous study, in uence-based aggregative learning (IAL), which is framework to facilitate the emergence of social norms in static complex network, to apply to changing networks. Then, we experimentally indicate that the norm can emerge in the networks of agents by the extended IAL.

Sanction systems, such punishment and reward, play an important role to maintain cooperation in social dilemmas. Many empirical and experimental studies support the effect of the sanction systems. However, some studies show inconsistent results for the efficiency of punishment and reward. We conduct experiments to clarify the efficiency of punishment and reward. Our results show that punishment can promote cooperation highly than reward. Furthermore we show that players tend to overestimate the amount of punishment and underestimate the amount of reward. We consider that this misestimation of sanction leads the difference of cooperation ratio in the sanction systems.

Perceptual Crossing experiment is the method for analyzing human interaction with low dimensions. In this experiment, subjects need the cooperative interaction for the task. In this study, We introduce two different time delay patterns in this experiment, and evaluate the stability of human interaction to the time delay. It shows that although the time delay collapsed the cooperative human interaction human can manage to adapt to the time delay by moving slowly. Our results support that the behavioral social interaction between two is not fully sequential but more like synchronized.

We present open-source software, named OACIS (Organizing Assistant for Comprehensive and Interactive Simulations), for managing a large amount of simulation jobs and their results, which is useful for social modeling and simulations. One of the most fundamental difficulties in social simulations is that the models for social phenomena are not established as well as those for natural sciences. It is also difficult to determine the initial or boundary condition even if big data are available since they are usually incomplete. Thus, for simulating social phenomena, it is important to capture the global property of the phase space, which requires many simulation jobs of various parameters. We introduce functionality of OACIS, which is designed to resolve such difficulties and maximize the productivity of researchers. Finally, we discuss the future direction of OACIS.