Constrained Markov Decision Processes via Backward Value Functions

Harsh Satija, Philip Amortila, Joelle Pineau

Keywords: Reinforcement Learning, Constrained Markov Decision Processes, Deep Reinforcement Learning

TL;DR: We present an on-policy method for solving constrained MDPs that respects trajectory-level constraints by converting them into local state-dependent constraints, and works for both discrete and continuous high-dimensional spaces.

Abstract: Although Reinforcement Learning (RL) algorithms have found tremendous success in simulated domains, they often cannot directly be applied to physical systems, especially in cases where there are hard constraints to satisfy (e.g. on safety or resources). In standard RL, the agent is incentivized to explore any behavior as long as it maximizes rewards, but in the real world undesired behavior can damage either the system or the agent in a way that breaks the learning process itself. In this work, we model the problem of learning with constraints as a Constrained Markov Decision Process, and provide a new on-policy formulation for solving it. A key contribution of our approach is to translate cumulative cost constraints into state-based constraints. Through this, we define a safe policy improvement method which maximizes returns while ensuring that the constraints are satisfied at every step. We provide theoretical guarantees under which the agent converges while ensuring safety over the course of training. We also highlight computational advantages of this approach. The effectiveness of our approach is demonstrated on safe navigation tasks and in safety-constrained versions of MuJoCo environments, with deep neural networks.

Community Implementations: [ 4 code implementations](https://www.catalyzex.com/paper/arxiv:2008.11811/code)