Coordination Games on Multiplex Networks: Consensus, Convergence, and Stability of Opinion Dynamics

Researchers Ruey-An Shiu and Parinaz Naghizadeh have published a new paper, "Coordination Games on Multiplex Networks: Consensus, Convergence, and Stability of Opinion Dynamics," that provides a formal model for how opinions form and spread across interconnected social networks. The study extends single-layer models by formulating opinion updates as a synchronous coordination game, where agents minimize a local cost to align with their neighbors. The team proposed two key coupling mechanisms: a merged model that aggregates influences from different network layers (like Twitter, Facebook, and Reddit) through weighted averages, and a switching model where interactions periodically alternate between these layers.

Using sophisticated random-walk and spectral graph analysis, the researchers derived mathematical conditions for when a group reaches consensus. A critical finding is that multilayer interactions can induce or accelerate global consensus even when no single social layer could achieve it alone. Conversely, individually coordinated layers may lose consensus once they become interconnected—a phenomenon with significant implications for understanding online polarization. Numerical experiments validated the theory, specifically highlighting the impact of layer weights and switching periods on convergence rates. The 12-page paper, featuring 11 figures, was submitted to arXiv in March 2026 (arXiv:2603.07633).

This research provides a crucial mathematical framework for tech companies and policymakers to understand information diffusion. The models help predict how echo chambers form or break, how misinformation spreads across platforms, and how interventions might steer large-scale public opinion. As social media platforms become increasingly intertwined through shared users and cross-posting, this work offers tools to analyze the stability and fragility of consensus in our digitally connected society.