We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Policy-based Primal-Dual Methods for Convex Constrained Markov Decision Processes

Abstract: We study convex Constrained Markov Decision Processes (CMDPs) in which the objective is concave and the constraints are convex in the state-action visitation distribution. We propose a policy-based primal-dual algorithm that updates the primal variable via policy gradient ascent and updates the dual variable via projected sub-gradient descent. Despite the loss of additivity structure and the nonconvex nature, we establish the global convergence of the proposed algorithm by leveraging a hidden convexity in the problem under the general soft-max parameterization, and prove the $\mathcal{O}\left(T^{-1/3}\right)$ convergence rate in terms of both optimality gap and constraint violation. When the objective is strongly concave in the visitation distribution, we prove an improved convergence rate of $\mathcal{O}\left(T^{-1/2}\right)$. By introducing a pessimistic term to the constraint, we further show that a zero constraint violation can be achieved while preserving the same convergence rate for the optimality gap. This work is the first one in the literature that establishes non-asymptotic convergence guarantees for policy-based primal-dual methods for solving infinite-horizon discounted convex CMDPs.
Comments: 31 pages
Subjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
Cite as: arXiv:2205.10715 [cs.LG]
  (or arXiv:2205.10715v1 [cs.LG] for this version)

Submission history

From: Donghao Ying [view email]
[v1] Sun, 22 May 2022 02:50:16 GMT (47kb)

Link back to: arXiv, form interface, contact.