Basin of attraction of a three-state DSI model of misinformation propagation dynamics

Abstract

The Disinformed–Susceptible–Informed (DSI) model offers a simple framework for modeling information and opinion spread in complex systems. This paper presents a comparative analysis of two modeling approaches for the DSI system: differential equations, which describe continuous macroscopic dynamics, and lattice-based simulations, which capture localized agent-level interactions. Using time-series analysis and basin-of-attraction plots, we examine how the system evolves and stabilizes under both methods. Both methods capture system dynamics which favor more disinformed states, but lattice simulations stabilizes in more mixed states. In contrast, the differential equation approach typically results in a mixed steady state over a wider range of parameters. Our results show that local interactions can suppress the influence of certain parameters, such as α, and promote disinformed states even under conditions that would favor informed consensus in the ODE model. This dual-framework comparison contributes a novel perspective to the modeling of misinformation spread and demonstrates the importance of spatial and localized dynamics, suggesting that intervention strategies must be tailored to the structure of the social environment.