Battlesnake Challenge Framework

该数据集是基于Battlesnake框架开发的多智能体强化学习框架，它包含了用于训练和评估的模拟健身房环境。此外，该框架还提供了一系列训练示例，并支持在竞争环境中对智能体进行评估。该数据集涵盖了不同规模的智能体数量（4、5和6个）和可变地图大小（7x7、11x11、19x19）。其任务是进行具有人类参与循环的多智能体强化学习。

This dataset is a multi-agent reinforcement learning (MARL) framework developed based on the Battlesnake framework, which includes simulated gym environments for training and evaluation. Additionally, the framework provides a series of training examples and supports agent evaluation in competitive environments. This dataset covers scenarios with varying agent counts (4, 5, and 6) and adjustable map sizes (7×7, 11×11, 19×19). Its targeted task is multi-agent reinforcement learning with human-in-the-loop participation.