Estimate time new edge would take

[narang-an]

Currently in mbta_env.py when a new edge is added it is assigned a random weight (aka travel time), update the _apply_action() function for action types 0 and 2 (adding a edge and rerouting an edge) so that the new edge has a better travel time estimate. Also assign the new edges to a line called "new" so we know what edges the agent added/changed.

[DanielCoder834]

You can update the mean commute time or make a separate function: (There is probably some other changes that need to be made like with the reward. This is just a template)

commute_times = self.commuters.get_commute_times()
valid_times = [t for t in commute_times.values() if t != float("inf")]
if valid_times: # Ensure it is not empty
     avg_commute_time = np.mean(valid_times)

There would need to be an update graph in the commuter population class:

this.graph = new_graph
self.station_list = list(new_graph.nodes())

From there:

....
# some actions have been taken
this.commuters.update(this._G)
....