Spaceship Escape Problem

该数据集模拟了一个场景，其中宇航员（智能体）在飞船内部必须穿越迷宫，通过虫洞回家，同时避开黑洞并收集护盾。该问题被表示为一个具有明确高级任务和低级动作的马尔可夫决策过程。该数据集的规模为10x10的迷宫网格，包含多种状态。任务则是采用分层强化学习来导航迷宫并完成任务。

This dataset simulates a scenario where an astronaut (AI Agent) inside a spaceship must navigate a maze, travel through a wormhole to return home, avoid black holes, and collect shields. This problem is formulated as a Markov Decision Process with distinct high-level tasks and low-level actions. The dataset is based on a 10×10 maze grid containing multiple states. The task requires using hierarchical reinforcement learning to navigate the maze and complete the mission.