Replication Package for Evaluating Fuzz Testing for Reinforcement Learning Models

This project provides a comprehensive evaluation of various reinforcement learning (RL) fuzzing approaches across the following benchmark environments:

• MountainCar: A foundational classic control task for evaluating exploration efficiency.
• BipedalWalker: A complex robotic locomotion task characterized by continuous state and action spaces.
• CARLA: A high-fidelity, photorealistic autonomous driving simulator for safety-critical scenario testing.