English

Linear Convergence of Entropy-Regularized Natural Policy Gradient with Linear Function Approximation

Machine Learning 2026-02-17 v4 Optimization and Control Machine Learning

Abstract

Natural policy gradient (NPG) methods with entropy regularization achieve impressive empirical success in reinforcement learning problems with large state-action spaces. However, their convergence properties and the impact of entropy regularization remain elusive in the function approximation regime. In this paper, we establish finite-time convergence analyses of entropy-regularized NPG with linear function approximation under softmax parameterization. In particular, we prove that entropy-regularized NPG with averaging satisfies the \emph{persistence of excitation} condition, and achieves a fast convergence rate of O~(1/T)\tilde{O}(1/T) up to a function approximation error in regularized Markov decision processes. This convergence result does not require any a priori assumptions on the policies. Furthermore, under mild regularity conditions on the concentrability coefficient and basis vectors, we prove that entropy-regularized NPG exhibits \emph{linear convergence} up to a function approximation error.

Keywords

Cite

@article{arxiv.2106.04096,
  title  = {Linear Convergence of Entropy-Regularized Natural Policy Gradient with Linear Function Approximation},
  author = {Semih Cayci and Niao He and R. Srikant},
  journal= {arXiv preprint arXiv:2106.04096},
  year   = {2026}
}
R2 v1 2026-06-24T02:56:35.743Z