What type of study is this?

August 17, 2025Journal of Artificial Intelligence ResearchOpen Access

Policy-based Primal-Dual Methods for Concave CMDP with Variance Reduction

View Full Paper

Ask AI

Bookmark

VR-PDPG demonstrates a convergence rate of O(T-1/3) for both optimality gap and constraint violation in concave cmdps.
The proposed method improves the convergence rate to O(T-1/2) with strong concavity of the objective in the occupancy measure.
In the sample-based setting, VR-PDPG achieves an O(ε-4) sample complexity for ε-global optimality.
Numerical experiments validate VR-PDPG's ability to achieve zero constraint violation without affecting the optimality gap.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Variance-Reduced Primal-Dual Policy Gradient Algorithm enhances convergence rate in concave cmdps, indicating significant improvements in sample complexity.

Cite This Study

Ying et al. (2025) studied this question.

synapsesocial.com/papers/68af453aad7bf08b1ead2a38 https://doi.org/10.1613/jair.1.18129

View Full Paper

Ask AI

Bookmark

Implication

Variance-Reduced Primal-Dual Policy Gradient Algorithm enhances convergence rate in concave cmdps, indicating significant improvements in sample complexity.

Policy-based Primal-Dual Methods for Concave CMDP with Variance Reduction

Key Points

Discussion

Implication

Cite This Study

Policy-based Primal-Dual Methods for Concave CMDP with Variance Reduction

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider