Multiagent Online Learning in Time-Varying Games

We examine the long-run behavior of multiagent online learning in games that evolve over time. Specifically, we focus on a wide class policies based mirror descent, and show induced sequence play (a) converges to Nash equilibrium time-varying stabilize long run strictly monotone limit, (b) it stays asymptotically close evolving stage (assuming they are strongly monotone). Our results apply both gradient- payoff-based feedback—that is, when players only get observe payoffs their chosen actions. Funding: This research was partially supported by European Cooperation Science Technology COST Action [Grant CA16228] “European Network for Game Theory” (GAMENET). P. Mertikopoulos is grateful financial support French National Research Agency (ANR) framework “Investissements d’avenir” program ANR-15-IDEX-02], LabEx PERSYVAL ANR-11-LABX-0025-01], MIAI@Grenoble Alpes ANR-19-P3IA-0003], ALIAS ANR-19-CE48-0018-01].