1 min readDec 18, 2019
Hello
Thanks for your comment.
Certainly the reward contributes to the decision making. But this is not what I meant in the sentence you quoted.
What I meant (and corrected) was the “reward function”. Similar to the transition probabilities which returns the next state (you can land an another state than the one you intended when you took an action), the reward function can return a different reward than the one intended by the “action”. Reward function can be stochastic too.
regards