Where to update actions in environment?

I'm creating a custom reinforcement learning environment by using the custom environment template (cart pole code). However, the cart pole environment has two actions at all times and therefore just defines them in the constructor. For my environment there can be varying actions based on the state and I was wondering where best to update that. I was thinking of doing it at the end of the setup function but I am not sure. Thanks

Answer 1

[Emmanouil Tzorakoleftherakis]

Reinforcement Learning Toolbox agents expect a static action space, so fixed number of options at each time step. To create a dynamic action space you would have to create your own custom agent.