* Fixing the DeepRMSA example

* Fixing the QoSConstrained example

Updating the file with the new metadata from the environment, i.e., episode_ prefix instead of the _since_reset suffix.

Updating the names of variables to reflect the terminology of RL, i.e., var_since_reset now reads episode_var.

Fixing an issue with the translation from action index into path and block ids in the DeepRMSA environment.

Updating and streamlining the execution of the RMSA and DeepRMSA environments. The RMSA environment now implements a multi-discrete action space.