About the FPGA implementation of a genetic algorithm for solving Sudoku puzzles

In the last years, a number of papers present algorithms for solving and generating Sudoku puzzles. Some studies test and confirm the possibility to use the genetic algorithm (GA) for solving and rating the Sudoku puzzles. The results are that GA can solve Sudoku puzzles, but not very effectively. Other studies test the possibility of a FPGA implementation of a "Sudoku solver" algorithm and "brute-force" methods are used due to excessive resource requirement for the implementations of other algorithms. Our paper analyzes the possibility of a FPGA implementation of the GA for solving the puzzles. The paper presents some possible adaptations of the general GA algorithm, concerning the population and chromosome definitions, crossover and mutation phases and fitness function calculation for this FPGA implementation.