arXiv:2605.07459 · v2

On the Complexity of Discounted Robust MDPs with $L_p$ Uncertainty Sets

首次提交 2026-05-08 17:06

版本更新 2026-11-06 06:42

雷达观测 2026-11-09 10:16

摘要

A basic model in sequential decision making is the Markov decision process (MDP), which is extended to Robust MDPs (RMDPs) by allowing uncertainty in transition probabilities and optimizing against the worst-case transition probabilities from the uncertainty sets. The class of $(s, a)$-rectangular RMDPs with $L_p$ uncertainty sets provides a flexible and expressive model for such problems. We study this class of RMDPs with a discounted-sum cost criterion and a constant discount factor. The existence of an efficient algorithm for this class is a fundamental theoretical question in optimization and sequential decision making. Previous results only establish a strongly polynomial-time algorithm for $L_\infty$ uncertainty sets. In this work, our main results are as follows: (a)~we show that for any compact uncertainty set, the policy iteration algorithm for RMDPs is strongly polynomial with oracle access to solutions of Robust Markov chains (RMCs); (b)~we present strongly polynomial-time bounds on the policy iteration algorithm for RMCs with $L_1$ and $L_\infty$ uncertainty sets; and (c)~we establish hardness results for RMCs with $L_p$ uncertainty sets for integer $p$ satisfying $1<p<\infty$. Finally, motivated by our theoretical bounds, we present experimental results showing how fast policy iteration converges for RMDPs with $L_1$ and $L_\infty$ uncertainty sets.