Global Convergence of Multi-Agent Policy Gradient in Markov Potential Games
This work addresses multi-agent coordination in state-dependent settings for researchers in game theory and reinforcement learning, though it appears incremental as it adapts existing gradient dominance arguments to multi-agent learning.
The paper tackles the problem of multi-agent coordination in Markov games by introducing a novel definition of Markov Potential Games (MPG) that generalizes prior attempts, and proves fast convergence of independent policy gradient to Nash policies in these games.
Potential games are arguably one of the most important and widely studied classes of normal form games. They define the archetypal setting of multi-agent coordination as all agent utilities are perfectly aligned with each other via a common potential function. Can this intuitive framework be transplanted in the setting of Markov Games? What are the similarities and differences between multi-agent coordination with and without state dependence? We present a novel definition of Markov Potential Games (MPG) that generalizes prior attempts at capturing complex stateful multi-agent coordination. Counter-intuitively, insights from normal-form potential games do not carry over as MPGs can consist of settings where state-games can be zero-sum games. In the opposite direction, Markov games where every state-game is a potential game are not necessarily MPGs. Nevertheless, MPGs showcase standard desirable properties such as the existence of deterministic Nash policies. In our main technical result, we prove fast convergence of independent policy gradient to Nash policies by adapting recent gradient dominance property arguments developed for single agent MDPs to multi-agent learning settings.