Grid-World
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/FRL2019/FRL
下载链接
链接失效反馈官方服务：
资源简介：
该数据集涉及的是网格世界领域，其中代理需要在由二进制值矩阵表示的环境中导航，以相遇彼此同时避开障碍物。环境内随机放置了障碍物，并设计了具有特定观察状态的两个代理，以促进导航。该数据集具有可变的大小（N_g = 8, 16, 32），其任务为强化学习。
This dataset pertains to the grid world domain, where agents are required to navigate in an environment represented by a binary-valued matrix to encounter each other while avoiding obstacles. Obstacles are randomly deployed within the environment, and two agents with specific observation states are engineered to facilitate navigation. This dataset supports variable sizes (N_g = 8, 16, 32) and is tailored for reinforcement learning tasks.
