An Attempt to solve Sudoku problem using a genetic algorithm in python `
- generate valid sudoku using graph coloring technique.
- Eliminate some of the values from the sudoku and generate different level of problem sudokus
- Pass problem Sudoku to the Genetic algorithm.
Initial Sample
- Initially, sample spaces are generated by picking the values allowed to a position, and if no valid value is possible then pick a random value.
Fitness calculation
- Fitness is calculated based on the incorrectness caused by each position. Instead of scoring the sample, each position of the samples is scored.
Selection
- Based on the probability (calculated from the score) of each position, positions are chosen for the next generation.
Numpy and Python are required to run the project.
python ga_solve_sudoku.py