adammain / ai-agent-sudoku

AI program solves Diagonal Sudokus using constraint propagation and search techniques. Additionally, taught the agent to use the Naked Twins advanced Sudoku strategy.

Sudoku Solver

Screenshot

Install

This project requires Python 3. Clone this repository

git clone https://github.com/adammain/ai-agent-sudoku

How to run

python3 sudoku.py

Input Format

81 characters long string row wise 9x9 with each character implying the value filled in that particular box. If box is unfilled put '.' as its place. For example: For this sudoku input:

4.....8.5.3..........7......2.....6.....8.4......1.......6.3.7.5..2.....1.4......

Algorithms detail

Constraint propagation

Used in functions elimination and naked_twin in utils.py to eliminate the possibility of presence of any digit if it is present in corresponding row, column or square.

Search

Implemented recursion techniques to create depth-first-search trees for solving hard sudokus that need guess work in any particular box to proceed.

Sudoku strategy

• Only choice rule
• Single possibility rule
• Naked Twin exclusion rule

MIT License

Copyright (c) 2017 Adam Main