- [x] [Sokoban](https://en.wikipedia.org/wiki/Sokoban) (see DeepMind's [boxoban](https://github.com/deepmind/boxoban-levels)) - [x] [N-puzzle](https://en.wikipedia.org/wiki/15_puzzle) - [ ] [Rush Hour](https://en.wikipedia.org/wiki/Rush_Hour_(puzzle)) - [ ] Boxworld, as described in - [Relational deep reinforcement learning](https://arxiv.org/abs/1806.01830) §4.1 - [An investigation of model-free planning](https://arxiv.org/abs/1901.03559) §3 - [ ] Collect, as described in - [Value prediction network](https://arxiv.org/abs/1707.03497) §4.2