search

google-deepmind / bsuite

bsuite is a collection of carefully-designed experiments that investigate core capabilities of a reinforcement learning (RL) agent
Apache License 2.0
1.51k stars 181 forks source link

The signature for `update` does not allow for sarsa or n-step methods? #34

Closed epignatelli closed 3 years ago

epignatelli commented 3 years ago

Hi There!

Thanks very much for bsuite, it is a great resource for reproducible research.

I have a question on the framework. I am setting up some pedagogic implementation of canonical rl algorithms, among which, sarsa.

Is there any design pattern you had in mind for n-step methods or any method that requires access to experience from longer transitions? I am currently solving the issue with sarsa by computing the next action with the select_action method in the update function. What about n-step methods or model-based methods?

epignatelli commented 3 years ago

This question doesn't even makes sense anymore to me.