Evaluating the Effectiveness of Q-Learning in Ant Colony Optimization for the Traveling Salesman Problem