Harsanyi's Equilibrium Selection for Distributed Sources Minimizing Age of Information

Efforts to minimize Age of Information (AoI) in communication networks, particularly within energy-constrained devices in the Internet of Things (IoT), have prompted extensive research into resource management techniques. This study explores the optimization of AoI over a finite horizon in the context of distributed IoT environments. We first frame the scenario of N distributed equivalent sources as a multi-agent coordination game, then we address the inefficiency of the resulting equilibria, quantified through the Price of Anarchy. We find the latter to be significant (higher than 1:5) already for few sources, and increasing in the number of players. Leveraging Harsanyi's theoretical framework for equilibrium selection, we argue for the importance of preplay communication for AoI efficiency, and suggest how this can be implemented in the IoT without resorting to full centralization.