This project includes a series of processes for generating a library of automated driving simulation scenarios using reinforcement learning methods. We ensure that the reward function in multi-agent reinforcement learning is effective by designing a function that simulates the complexity of the scene.
