CertRL: Formalizing Convergence Proofs for Value and Policy Iteration in Coq
This work addresses the need for verified correctness in reinforcement learning for safety-critical applications, though it is incremental as it builds on existing formal methods and algorithms.
The authors tackled the lack of formal convergence proofs for reinforcement learning algorithms in safety-critical settings by developing a Coq formalization of value and policy iteration for finite state Markov decision processes, resulting in a formal proof of Bellman's optimality principle and a library for further work.
Reinforcement learning algorithms solve sequential decision-making problems in probabilistic environments by optimizing for long-term reward. The desire to use reinforcement learning in safety-critical settings inspires a recent line of work on formally constrained reinforcement learning; however, these methods place the implementation of the learning algorithm in their Trusted Computing Base. The crucial correctness property of these implementations is a guarantee that the learning algorithm converges to an optimal policy. This paper begins the work of closing this gap by developing a Coq formalization of two canonical reinforcement learning algorithms: value and policy iteration for finite state Markov decision processes. The central results are a formalization of Bellman's optimality principle and its proof, which uses a contraction property of Bellman optimality operator to establish that a sequence converges in the infinite horizon limit. The CertRL development exemplifies how the Giry monad and mechanized metric coinduction streamline optimality proofs for reinforcement learning algorithms. The CertRL library provides a general framework for proving properties about Markov decision processes and reinforcement learning algorithms, paving the way for further work on formalization of reinforcement learning algorithms.