Modeling and Controlling Polymorphic Opinion Dynamics in Social Networks With Heterogeneous Sensitivities

Social media platforms have amplified opinion polarization, complicating policymaking and public sentiment management. Existing models primarily address either consensus formation or binary polarization, failing to account for the diverse, multi-polar opinion distributions observed in real-world social interactions. This work introduces a dynamic opinion model integrating individual stubbornness and opinion sensitivity, enabling the simulation of multi-opinion coexistence within social networks. It is formally proven that, under low-impact events, opinions converge to a neutral consensus under both synchronous and asynchronous updates. For high-impact events, where theoretical analysis is challenging, this work employs simulations under asynchronous updating, which better reflect real conditions. The results indicate that mixed groups may exhibit multi-state coexistence, with homogenized social relationships leading to weak opinion cohesion and social heterogeneity fostering community formation. To regulate diverse opinions in social networks, this work introduces two governance strategies designed for events with different impacts. Notably, for events with limited impact, soft opinion guidance shapes group preferences more effectively than forceful interventions. These findings offer insights for refining policy communication strategies and advancing the broader understanding of opinion dynamics in diverse societies.