- About the Project
- Questions from Udacity
- Install Guide
- Dependencies
- Resources
This project is a part of AIND(Artificial Intelligence Nanodegree)'s assignment. This project is about building an agent to solve any 'Diagonal Sudoku' problem using 3 techniques, "eliminate", "only one choice", and "naked twins". The agent tries to solve problems by applying the techniques while performing Depth First Search.
A: First of all, naked twins problem is applicable when there are 2-digit boxes available. Second, when I come across a 2-digit box while searching all boxes, then I look up its row/column peers. Third, I go through all row/column peers and search for a box having the same 2-digit value. Fourth, if the box is found, I go through row/column peers and remove the 2-digit value from row/column peers. If a peer contain only one digit of the 2-digit value, that one digit will also be deleted.
Question 2 (Diagonal Sudoku) - How do we use constraint propagation to solve the diagonal sudoku problem?
A: There are only 2 diagonal lines in Sudoku, so I made a list of box belonging to the lines. Just like for normal Sudoku, the same techniques can be applied to Diagonal Sudoku but with slightly modified version. For eliminate technique, I eliminated all duplicate values of possible digits from boxes belonging to the same diagonal line. Also, I checked if there are boxes with the same 2-digit value so that I could perform naked twins technique for each diagonal lines. Only one choice technique is also modified to support diagonal sudoku to find a box which has the only unique value comparing to other boxes on each diagonal lines.
- This project requires Python 3.x.
- This project optionally requires Pygamehere for visualization purpose.
solution.py- Complete Diagonal Sudoku problem solving logic.test_solution.py- You can test your solution by runningpython -m unittest.PySudoku.py- This is code for visualizing your solution.visualize.py- This is code for visualizing your solution.