The document describes using a genetic algorithm to solve, rate, and generate Sudoku puzzles. It discusses representing Sudoku puzzles and their solutions as chromosomes. The genetic algorithm uses crossover and mutation operators to evolve puzzle solutions over generations. It also uses a fitness function based on puzzle difficulty metrics to evaluate and rank generated puzzles according to their difficulty level.
24. οΆ 1. T. Mantere, J. Koljonen, βSolving, Rating and
Generating Sudoku Puzzles with GA,β in 2007 IEEE
Congress on Evolutionary computation β CEC2007,
Singapore, 2007
οΆ 2. T. Mantere, J. Koljonen, βSolving and rating Sudoku
puzzles with genetic algorithms,β in Finnish Artificial
Intelligence Conference (STeP 2006), October 26-27,
FAIS, Espoo, Finland, 2006
οΆ 3. T. Mantere, J. Koljonen, βSolving and Analyzing
Sudokus with Cultural Algorithmsβ in 2008 IEEE
/
Sudoku