Text this: Reducing Computational Time in Pixel-Based Path Planning for GMA-DED by Using Multi-Armed Bandit Reinforcement Learning Algorithm