Skip to yearly menu bar Skip to main content


Poster

Online Reinforcement Learning in Stochastic Games

Chen-Yu Wei · Yi-Te Hong · Chi-Jen Lu

Pacific Ballroom #17

Keywords: [ Bandit Algorithms ] [ Online Learning ] [ Multi-Agent RL ] [ Reinforcement Learning ] [ Model-Based RL ] [ Markov Decision Processes ]


Abstract: We study online reinforcement learning in average-reward stochastic games (SGs). An SG models a two-player zero-sum game in a Markov environment, where state transitions and one-step payoffs are determined simultaneously by a learner and an adversary. We propose the \textsc{UCSG} algorithm that achieves a sublinear regret compared to the game value when competing with an arbitrary opponent. This result improves previous ones under the same setting. The regret bound has a dependency on the \textit{diameter}, which is an intrinsic value related to the mixing property of SGs. Slightly extended, \textsc{UCSG} finds an $\varepsilon$-maximin stationary policy with a sample complexity of $\tilde{\mathcal{O}}\left(\text{poly}(1/\varepsilon)\right)$, where $\varepsilon$ is the error parameter. To the best of our knowledge, this extended result is the first in the average-reward setting. In the analysis, we develop Markov chain's perturbation bounds for mean first passage times and techniques to deal with non-stationary opponents, which may be of interest in their own right.

Live content is unavailable. Log in and register to view live content