3

I am trying to write a custom openAI Gym environment in which the agent takes 2-actions in each step, one of which is a discrete action and the other is continuous one. I am using Ray RLLib and using SAC algorithm as it supports both discrete and continuous action spaces. My question is does any one have an idea of how to make the agent take two-actions per step, giving one is continuous and the other is discrete ?

JoCode
  • 31
  • 2

1 Answers1

0

If action space is like MultiDiscrete([10,10]) it can take 2 different type of actions when it predict it (The Value 10 can be change btw)

Thiroshan Suthakar
  • 117
  • 2
  • 12