Optimization of reward shaping function based on genetic algorithm applied to a cross validated deep deterministic policy gradient in a powered landing guidance problem.