2
$\begingroup$

I have an environment that is computationally heavy (takes several seconds to get a reward and next state). This limits reinforcement capability, due to poor sampling of the problem. There is any strategy that could be used to address the problem (e.g. If I can use the environment in parallel, then I could use a multi-agent approach)

$\endgroup$
2
  • $\begingroup$ Mltiple environments, installing some form of reward shaping or prior knowledge into the problem? $\endgroup$ – FourierFlux Sep 22 '20 at 18:28
  • $\begingroup$ @FourierFlux In terms of multiple environments - then how you would integrate them? Average updates like in A3C alg. ? $\endgroup$ – Daniel Wiczew Sep 23 '20 at 14:53
0
$\begingroup$

You could use Ray RLlib. It has support for parallel environments, even over multiple GPUs and compute nodes.

$\endgroup$

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.