Implementation of a genetic algorithm, using the DEAP framework, to solve a sudoku with 17 known numbers. The fitness function is proposed as a minimization problem, it consists in evaluating that the numbers are not repeated in the rows, columns or sectors, otherwise, the fitness is increased in one unit, so the ideal fitness for a solved sudoku is 0.0.