Evolutionary computing methodologies for constrained parameter, combinatorial optimization: Solving the Sudoku puzzle

Three evolutionary computing algorithms are applied to a constrained parameter, combinatorial optimization problem; the Sudoku puzzle. These methodologies include, quantum simulated annealing, cultural genetic algorithm and a hybrid between simulated annealing and genetic algorithm. The results obtained from these techniques indicate that the most effective of these optimization techniques is quantum simulated annealing with an effective accuracy of solving 64 out of 100 simulations in under 6 000 iterations, with an average running time of approximately 40.2 seconds. While classical, logic based search algorithms tend to outperform these evolutionary computational algorithms (in both complexity and time) for simple, `unique-solution' problems, it is found that the evolutionary based algorithms surpasses these classical methodologies when solving higher dimensional, more complex puzzles.