Multi-agent Simulation of New Product Diffusion: Based on Gravity Theory and Directed Network

As the penetration rate of the mobile Internet approaches saturation, the diffusion of new products has shifted from pursuing new user growth to enhancing user activity and interaction, that is, the directed and contingent nature of online social networks. This study combines the gravitational field theory and the evolutionary game model to design a mechanism for the formation of directed and contingent networks, and uses multi - Agent simulation to study the evolutionary laws and characteristics of consumer decision - making under different network structures. The research findings are as follows: (1) The number of Hub nodes has a negative impact on the diffusion stability in the introduction period. (2) The initial degree of nodes has a positive impact on the final diffusion proportion and a significantly positive impact on the diffusion speed in the introduction period. (3) A larger proportion of out - edges will intensify the fluctuations in group decision - making during the introduction period of new product diffusion, delay the time for product diffusion to reach a stable state, and lead to a reduction in the final diffusion proportion. (4) The initial adoption proportion is positively correlated with the diffusion convergence speed, and this effect shows a marginal diminishing return. This paper introduces the gravitational field theory and directed and contingent networks into the new product diffusion model, providing a new perspective for product diffusion research.