Pith. sign in

REVIEW 2 cited by

A Policy Gradient Primal-Dual Algorithm for Constrained MDPs with Uniform PAC Guarantees

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2401.17780 v3 pith:FXSWNDEH submitted 2024-01-31 cs.LG

A Policy Gradient Primal-Dual Algorithm for Constrained MDPs with Uniform PAC Guarantees

classification cs.LG
keywords algorithmguaranteesoptimalpoliciesalgorithmscmdpconstrainedconvergence
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We study a primal-dual (PD) reinforcement learning (RL) algorithm for online constrained Markov decision processes (CMDPs). Despite its widespread practical use, the existing theoretical literature on PD-RL algorithms for this problem only provides sublinear regret guarantees and fails to ensure convergence to optimal policies. In this paper, we introduce a novel policy gradient PD algorithm with uniform probably approximate correctness (Uniform-PAC) guarantees, simultaneously ensuring convergence to optimal policies, sublinear regret, and polynomial sample complexity for any target accuracy. Notably, this represents the first Uniform-PAC algorithm for the online CMDP problem. In addition to the theoretical guarantees, we empirically demonstrate in a simple CMDP that our algorithm converges to optimal policies, while baseline algorithms exhibit oscillatory performance and constraint violation.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Primal-Dual Policy Optimization for Linear CMDPs with Adversarial Losses

    cs.LG 2026-05 unverdicted novelty 7.0

    A new primal-dual algorithm for adversarial linear CMDPs achieves the first sublinear regret and constraint violation bounds of order K to the 3/4 using weighted LogSumExp softmax policies with periodic mixing and reg...

  2. Near-Optimal Policy Identification in Robust Constrained Markov Decision Processes via Epigraph Form

    cs.LG 2024-08 unverdicted novelty 7.0

    Presents the first algorithm to identify an ε-optimal policy in robust constrained MDPs via epigraph form and bisection search with Õ(ε^{-4}) robust policy evaluations.