Study on the Effect of Utility Uncertainty on Behavior Diffusion in Autonomous and Rational Networks

For behavior diffusion in autonomous and rational networks, most understanding comes from models and ideas borrowed from epidemiology on empirical or synthetic biological networks. However, in rational and autonomous networks, agents make strategic choice about behaviors in contrast being randomly assigned an attribute (such as being infected). Specifically, an agent is assumed to behave rationally by selecting the behavior that brings a high utility. But, individual's utility contains some random element, so-called uncertainty, whose effect on the behavior diffusion is not examined. This paper thoroughly investigates the effect of utility uncertainty on the behavior diffusion in autonomous and rational networks. Specifically, we first introduce various games to model users' rational interactions. Then, through simplifying utility uncertainty with log it model, we propose a generic analytical framework based on mean field theory to formally analyze the effect of uncertainty on behavior diffusion. Our finding is that small uncertainty can facilitate the behavior diffusion in autonomous and rational networks.