To run the program, do the following:
- Using Github Desktop, select File > Clone Repository and paste the following forked url: https://github.com/winstreak1/sokoban-AI-updated
- In Github Desktop, open the program using pycharm.
- Select the file test_script.py.
- Beginning on line 7, change the boolean state to 'True' for the algorithm you wish to run and all other algorithms set to 'False'. ensure all files are in the same working directory and run the test scripts. Additional test cases can be added.
- Run the program.
The file solution.py contains notes on algorithms implemented in the program. Some notes may be from source of original program.
- Fung, A (angusfung), sokoban-AI, Github, accessed September 27, 2022, https://github.com/angusfung/sokoban-AI
AI of the game Sokoban, using basic search strategies {depth-first search, breadth-first search, best-first search, A-star search, uniform-cost search, custom search} and more advanced search variants thereof {anytime greedy best-first search, anytime weighted A-start search}, with heuristics {displaced blocks, manhattan distance, euclidean distance, L2-norm, etc.}, fully equipped with deadblock checking and dictionary lookup (for better amortized runtime).
Modular enough to add custom search strategies or heuristics in solution.py, and any more advanced techniques can be built on top of them.