Risk-Constrained Reinforcement Learning with Percentile Risk Criteria

Abstract

In many sequential decision-making problems one is interested in minimizing an expected cumulative cost while taking into account \emph{risk}, i.e., increased awareness of events of small probability and high consequences. Accordingly, the objective of this paper is to present efficient reinforcement learning algorithms for risk-constrained Markov decision processes (MDPs), where risk is represented via a chance constraint or a constraint on the conditional value-at-risk (CVaR) of the cumulative cost. We collectively refer to such problems as percentile risk-constrained MDPs. Specifically, we first derive a formula for computing the gradient of the Lagrangian function for percentile risk-constrained MDPs. Then, we devise policy gradient and actor-critic algorithms that (1) estimate such gradient, (2) update the policy in the descent direction, and (3) update the Lagrange multiplier in the ascent direction. For these algorithms we prove convergence to locally optimal policies. Finally, we demonstrate the effectiveness of our algorithms in an optimal stopping problem and an online marketing application.

arXiv admin note: substantial text overlap with arXiv:1406.3339

References in corpus (2)

• Policy Gradients with Variance Related Risk Criteria
• Variance Adjusted Actor Critic Algorithms

Cited by in corpus (10)

• A Lyapunov-based Approach to Safe Reinforcement Learning
• Accelerated Primal-Dual Policy Optimization for Safe Reinforcement Learning
• Classification with Costly Features as a Sequential Decision-Making Problem
• Risk-Sensitive Reinforcement Learning: a Martingale Approach to Reward Uncertainty
• Efficient Off-Policy Safe Reinforcement Learning Using Trust Region Conditional Value at Risk
• Federated Learning with Superquantile Aggregation for Heterogeneous Data
• Device Heterogeneity in Federated Learning: A Superquantile Approach
• Reinforcement Learning by Guided Safe Exploration
• A Survey of Learning Criteria Going Beyond the Usual Risk
• Mean-Semivariance Policy Optimization via Risk-Averse Reinforcement Learning