1
$\begingroup$

I have to reproduce the algorithm of a scientific article and have chosen one that uses Reinforcement Learning.

However, I don't understand how to compute the transition probability function ($f$) used in the iteration:

Q-value iteration

The article says:

In order to run the Q-value iteration algorithm (Table 2), the transition probability function f was first computed.

My question is: how was it computed?

$\endgroup$
1
$\begingroup$

The transition probability function is a property of the environment. This is one of the things you need to run a simulation. How you compute it is entirely up to you. You could make one up.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you. Can I just take an arbitrary function or should I use some technique to evaluate a better function? $\endgroup$ – user35477 May 14 '14 at 17:51
  • $\begingroup$ I recommend looking at the standard environments at library.rl-community.org $\endgroup$ – Don Reba May 14 '14 at 19:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.