Towards Painless Policy Optimization for Constrained MDPs

by   Arushi Jain, et al.

We study policy optimization in an infinite horizon, γ-discounted constrained Markov decision process (CMDP). Our objective is to return a policy that achieves large expected reward with a small constraint violation. We consider the online setting with linear function approximation and assume global access to the corresponding features. We propose a generic primal-dual framework that allows us to bound the reward sub-optimality and constraint violation for arbitrary algorithms in terms of their primal and dual regret on online linear optimization problems. We instantiate this framework to use coin-betting algorithms and propose the Coin Betting Politex (CBP) algorithm. Assuming that the action-value functions are ε_b-close to the span of the d-dimensional state-action features and no sampling errors, we prove that T iterations of CBP result in an O(1/(1 - γ)^3 √(T) + ε_b√(d)/(1 - γ)^2) reward sub-optimality and an O(1/(1 - γ)^2 √(T) + ε_b √(d)/1 - γ) constraint violation. Importantly, unlike gradient descent-ascent and other recent methods, CBP does not require extensive hyperparameter tuning. Via experiments on synthetic and Cartpole environments, we demonstrate the effectiveness and robustness of CBP.


Convergence and sample complexity of natural policy gradient primal-dual methods for constrained MDPs

We study sequential decision making problems aimed at maximizing the exp...

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

We study convex Constrained Markov Decision Processes (CMDPs) in which t...

A Primal Approach to Constrained Policy Optimization: Global Optimality and Finite-Time Analysis

Safe reinforcement learning (SRL) problems are typically modeled as cons...

Provably Efficient Safe Exploration via Primal-Dual Policy Optimization

We study the Safe Reinforcement Learning (SRL) problem using the Constra...

Safe and Near-Optimal Policy Learning for Model Predictive Control using Primal-Dual Neural Networks

In this paper, we propose a novel framework for approximating the explic...

The Geometry of Memoryless Stochastic Policy Optimization in Infinite-Horizon POMDPs

We consider the problem of finding the best memoryless stochastic policy...

Cautious Reinforcement Learning via Distributional Risk in the Dual Domain

We study the estimation of risk-sensitive policies in reinforcement lear...