Text this: Exploring Proximal Policy Optimization in ViZDoom: Training Agents for Complex Tasks with Hyperparameter Optimization