Skip to content

Exploring a variety of search techniques variants of the (modified) N-puzzle for CS534 (Artificial Intelligence). Team members are Neil Kale, Lauren Stanley, Ankit Mittal, Ashkan Ganj, and Rutwik Kulkarni.

License

Notifications You must be signed in to change notification settings

anki-mittal/npuzzle-search

 
 

Repository files navigation

npuzzle-search

Exploring a variety of search techniques variants of the (modified) N-puzzle for CS534 (Artificial Intelligence). Team members are Neil Kale, Lauren Stanley, Ankit Mittal, Ashkan Ganj, and Rutwik Kulkarni.

Alt text

Please run our program by calling npuzzle.py, in the code folder from the main npuzzle-search folder. The arguments are:

  • Board file name, with the board located in the main npuzzle-search folder.
  • Heuristic method
    • sliding will run A* using the manhattan distance heuristic detailed in Part 1
    • part2 will run the hill climbing search detailed in Part 3
    • greedy will run A* using the greedy search heuristic detailed in Part 3
  • Weighted tile boolean
    • false will use a strict Manhattan distance
    • true will use the weighted Manhattan distance

For example: python .\code\npuzzle.py board1.csv greedy false

Will solve board 1 using the greedy search heuristic with the strict Manhattan distance.

About

Exploring a variety of search techniques variants of the (modified) N-puzzle for CS534 (Artificial Intelligence). Team members are Neil Kale, Lauren Stanley, Ankit Mittal, Ashkan Ganj, and Rutwik Kulkarni.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 88.6%
  • Python 11.4%