Allocation Strategies Based on Possibilistic Rewards for the Multi-armed Bandit Problem: A Numerical Study and Regret Analysis | Publicación