Let's say we have N bandit machines with some distributions (assume some are gaussian, some are uniform, some are chi squared). We want to maximize rewards in X amount of time. I am aware that algorithms like Epsilon greedy, UCB can handle this. But is applying an RL algorithm like A2C the right choice to a plain vanilla Multi Armed Bandit problem (N independent machines that we have to maximize rewards from in X amount of time)? Is the notion even correct? Given that there is only always one state.



You must log in to answer this question.