Optimal control in opinion dynamics models: towards a unified framework

Abstract: Understanding how individuals change their opinions is essential to mitigate the harmful effect of fake news and stop opinion polarization. Building upon an extremely flexible opinion dynamics model that can capture the key microscopic mechanisms of social influence, the current paper elaborates a framework to find optimal control, which, thus, can be applied to a broad set of opinion dynamics settings. Using a combination of theoretical and computational approaches, I characterize the properties of optimal control and demonstrates how the elaborated framework can be applied to specific examples, which cover the classical situations that garner remarkable attention in the literature on opinion dynamics models: assimilative influence, bounded confidence, and dissimilative influence.