A new genetics-based diffusion model for social networks

A variety of stochastic diffusion models are used to simulate spreading processes in networks, but they focus on either one single object or exclusive objects spreading. In this paper, a genetics-based diffusion model (GDM) is introduced as a general model. It can simulate multiple objects with different relationships spreading in social networks. To simulate information diffusion, GDM regards an individual in a network as a `chromosome´, and a message that spreads in as a `gene´, and specifies a rule for the interactions between chromosomes to model the information interactions between individuals. We find that when modeling one single message spreading in networks, GDM would be exactly the same with the SI (Susceptible-Infected) case of independent cascade model. Besides, GDM can model many other cases of spreading processes, including competing processes. By applying GDM to simulating different cases of information diffusion process, we get many interesting results, including `break point´ in a diffusion process. The diffusion scale of a piece of information hardly increases before this point, but increases rapidly after it. So if we want to limit the diffusion scale of a piece of information, it would be very effective to block the propagation paths before its `break point´.