---
title: "A Survey on the Verification of Reinforcement Learning Policies"
authors:
  - "Luca Marzari"
  - "Ezio Bartocci"
  - "Enrico Marchesini"
doi: "10.24963/ijcai.2026/883"
url: "https://ijcai.org/proceedings/2026/883"
pdf: "https://ijcai.org/proceedings/2026/0883.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "7958-7966"
track: "Survey Track"
keywords:
  - "Agent-based and Multi-agent Systems: Formal verification, validation and synthesis"
  - "Machine Learning: Reinforcement learning"
---

# A Survey on the Verification of Reinforcement Learning Policies

Luca Marzari, Ezio Bartocci, Enrico Marchesini

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 7958-7966. DOI: [10.24963/ijcai.2026/883](https://doi.org/10.24963/ijcai.2026/883)

## Abstract

Reinforcement learning (RL) is increasingly applied in complex, safety-critical domains, yet the lack of rigorous behavioral guarantees for neural network-based policies remains a major barrier to deployment. Recent advances in policy expressiveness and scale have intensified this challenge, leading to a rapidly growing but conceptually fragmented body of work on RL policy verification. This survey provides a unifying perspective on RL verification methods. We introduce a taxonomy that clarifies relationships among existing approaches along three axes: verification paradigm (formal versus probabilistic), temporal scope (step-wise versus multi-step), and guarantees strength. Beyond taxonomy, we unify underlying theoretical foundations, make implicit assumptions and limitations explicit, and identify emerging directions.