Create norm-based action representation

[bfaught3]

Represent action to the RL model as the pairwise norms matrix of the grid map that would be created after the selected action is taken.

Output:

• Actions outputted by the learning model now consist of the pairwise norms matrix of the grid map that would be created after the selected action is taken
• A method allows the ogm to output actions in the format described in the point above