ait-energy / matsim-drs Goto Github PK

This is probably due to us already using log4j2 instead of log4j.

Possible solutions:

• update to MATSim v15(-SNAPSHOT) because they switched to log4j2 as well
• downgrade our code to log4j v1 (means rewriting many log statements because the {} syntax is not supported by log4j v1)

Should have been fixed with 2f2ce9e

But for the full Upper Austria scenario with pickupWaitingSeconds=180 at iteration 10 we got:

• 17664/19588 successful pickups (that's still a 10% error rate)
• 1207 stuck riders

TODO confirm that this is still a problem.

Currently unmatched rider trips are teleported. Should we change this?

Previously we encountered the number of unmatched riders skyrocketing. Now that we got this under control with carpooling.riderMobilityGuaranteeMonetaryConstant this may be a non-issue.

The most logical idea is to replace the unmatched rider trip with the agent's original trip.

But there may be issues: We use subtourmodechoice to set the mode for all trips of a subtour to ride. Now if originally car was used for the whole subtour (e.g. consisting of 3 trips) we can not simply restore the car mode for a single trip if the other trips were matched - because the car won't be available

However: sbbrailraptor also ignores such situations, so maybe we could ignore it as well?

However, even if the rider arrives soon after the driver starts waiting for him, the driver will only depart when the full pickupWaitingSeconds are over.

Can we get notified in case the rider is ready to be picked up?

What I want:

• successful pickups during qsim
• unsuccessful pickups during qsim (i.e. nr of matched requests - successful pickups)
• nr of stuck riders
• nr of stuck drivers
• pickup waiting time stats (optional)

.. ideally in an ever-growing .csv / .png in the main output folder (not in the ITERS folder)