Constrained Model-based Reinforcement Learning with Robust Cross-Entropy Method

Abstract

This paper studies the constrained/safe reinforcement learning (RL) problem with sparse indicator signals for constraint violations. We propose a model-based approach to enable RL agents to effectively explore the environment with unknown system dynamics and environment constraints given a significantly small number of violation budgets. We employ the neural network ensemble model to estimate the prediction uncertainty and use model predictive control as the basic control framework. We propose the robust cross-entropy method to optimize the control sequence considering the model uncertainty and constraints. We evaluate our methods in the Safety Gym environment. The results show that our approach learns to complete the tasks with a much smaller number of constraint violations than state-of-the-art baselines. Additionally, we are able to achieve several orders of magnitude better sample efficiency when compared with constrained model-free RL approaches. The code is available at \url{https://github.com/liuzuxin/safe-mbrl}.

8 pages, 5 figures

References in corpus (14)

• Playing Atari with Deep Reinforcement Learning
• Trust Region Policy Optimization
• Safe Model-based Reinforcement Learning with Stability Guarantees
• Safe Exploration in Continuous Action Spaces
• Benchmarking Batch Deep Reinforcement Learning Algorithms
• Lyapunov-based Safe Policy Optimization for Continuous Control
• Constrained Policy Optimization
• Safe Reinforcement Learning in Constrained Markov Decision Processes
• Can Autonomous Vehicles Identify, Recover From, and Adapt to Distribution Shifts?
• Responsive Safety in Reinforcement Learning by PID Lagrangian Methods
• Variational Inference MPC for Bayesian Model-based Reinforcement Learning
• Scaling Up Robust MDPs by Reinforcement Learning
• Delay-Aware Multi-Agent Reinforcement Learning for Cooperative and Competitive Environments
• MAPPER: Multi-Agent Path Planning with Evolutionary Reinforcement Learning in Mixed Dynamic Environments

Cited by in corpus (4)

• Approximate Model-Based Shielding for Safe Reinforcement Learning
• Scalable Safety-Critical Policy Evaluation with Accelerated Rare Event Sampling
• Learning Off-Policy with Online Planning
• Risk Sensitive Model-Based Reinforcement Learning using Uncertainty Guided Planning