0

How many states could I work with on my ordinary home computer when I want to implement a reinforcement learning algorithm such as Q-Learning? 1 thousand, 1 million, more?

nbro
  • 15,395
  • 32
  • 113
  • 196
MMM
  • 373
  • 1
  • 4
  • 12
  • Hi Ben. Please, if you have questions related to reinforcement learning, the best place to ask them is [Artificial Intelligence Stack Exchange](https://ai.stackexchange.com/), unless your questions are only related to programming issues (which may be the case). So, please, next time, ask them there! – nbro Apr 15 '20 at 17:18
  • Also, this question is a bit unclear. What do you mean by "work with on my ordinary". Are you interested in memory requirements, or what? Your question isn't really clear. Please, next time, try to be more specific and clearer. – nbro Apr 15 '20 at 17:21
  • Oh ok, thank you. I can not delete this question anymore, because it already has an answer. – MMM Apr 15 '20 at 18:50

1 Answers1

1

It is highly unadvisable to run a lot of states. The reason is really simple - when there are a lot of states in the memory, by the time the GPU finds the state and its corresponding action, the game already changes to another state.

So the solution is to use something a bit more advanced than naive Q-learning. See Deep Q-learning and other popular variants of RL like A3C. They help to avoid this issue

neel g
  • 1,138
  • 1
  • 11
  • 25