An Abstraction-based Method to Check Multi-Agent Deep Reinforcement-Learning Behaviors

TL;DR

This paper introduces an abstraction-based approach that combines formal verification with deep reinforcement learning to ensure agents adhere to safety constraints in multi-agent systems, applicable during training and testing.

Contribution

It presents a novel method that uses formal verification and abstract modeling to guarantee safety constraints in multi-agent deep RL, with automatic model generation and formal guarantees.

Findings

Effective in ensuring safety constraints in multi-agent environments

Reduces verification complexity through abstraction

Provides formal guarantees of safety compliance

Abstract

Multi-agent reinforcement learning (RL) often struggles to ensure the safe behaviours of the learning agents, and therefore it is generally not adapted to safety-critical applications. To address this issue, we present a methodology that combines formal verification with (deep) RL algorithms to guarantee the satisfaction of formally-specified safety constraints both in training and testing. The approach we propose expresses the constraints to verify in Probabilistic Computation Tree Logic (PCTL) and builds an abstract representation of the system to reduce the complexity of the verification step. This abstract model allows for model checking techniques to identify a set of abstract policies that meet the safety constraints expressed in PCTL. Then, the agents' behaviours are restricted according to these safe abstract policies. We provide formal guarantees that by using this method, the actions of the agents always meet the safety constraints, and provide a procedure to generate an abstract model automatically. We empirically evaluate and show the effectiveness of our method in a multi-agent environment.