If I remember correctly DQLearning is using discrete action space, limited number of possible actions which makes it possible to simply take the argmax_i(qvalue

If I remember correctly DQLearning is using discrete action space, limited number of possible actions which makes it possible to simply take the argmax_i(qvalue(state, action_i))