Open codecamaru opened 6 months ago
Does someone know if this or something similar, like only modifying the reward function, is already possible?
If the features and reward function remain unchanged, there is no need to alter the environment; otherwise, modifications are required.
Is your feature request related to a problem? Please describe. I don't know if changing the environment to make state space composition is feasible so that users or developers can use different variables when modelling the MDP.
Describe the solution you'd like The user can select the variables to consider when establishing the reward and the ones to consider in the States.