Skip to yearly menu bar Skip to main content


Poster

Achieving Tractable Minimax Optimal Regret in Average Reward MDPs

Victor Boone · Zihan Zhang

[ ]
Wed 11 Dec 4:30 p.m. PST — 7:30 p.m. PST

Abstract: In recent years, significant attention has been directed towards learning average-reward Markov Decision Processes (MDPs).However, existing algorithms either encounter suffer of sub-optimal regret guarantees or computational inefficiencies.In this paper, we present the first *tractable* algorithm with minimax optimal regret of $\widetilde{\mathrm{O}}\left(\sqrt{\mathrm{sp}(h^*) S A T}\right)$ ($\widetilde{\mathrm{O}}(\cdot)$ hides logarithmic factors of $(S,A,T)$) where $\mathrm{sp}(h^*)$ is the span of the optimal bias function $h^*$, $S\times A$ is the size of the state-action space and $T$ the number of learning steps. Remarkably, our algorithm does not require prior information on $\mathrm{sp}(h^*)$. Our algorithm relies on a novel subroutine, **P**rojected **M**itigated **E**xtended **V**alue **I**teration (`PMEVI`), to compute bias-constrained optimal policies efficiently. This subroutine can be applied to various previous algorithms to obtain improved regret bounds.

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