MATSim Code

Not only UmlaufDrivers need to traverse their first link, but all AbstractTransitDrivers. This just didn't fail because the non-Umlauf TransitDriver is not in wide use anymore. Remove it?

{@link TimeAllocationMutator} without an immediate {@link ReRoute} yields invalid routes.
Thus, agents tend to be stuck and cannot probe successfully for alternative departures.
Same argument holds true for individual transport (car), but implications are not that clear.

Hope, we get this somehow back into the core...

(some transport engineer)

First implementation of CA-parking-mobsim from Horni, Montini & Waraich.

functionality to compare the flow volumes of the ks2010 and matsim solutions added

made PersonDriverAgentImpl.chooseNextLinkId() more idempotent. Fixed a route in a test (the wrong route didn't matter in the test, because it only led to a stuckevent and that wasn't tested, but it came up when I slightly changed the protocol because when chooseNextLinkId is called twice in a row for a wrong route, it would crash instead of throwing a stuckevent. Removed one case of special treatment of transitdrivers on QLinkImpls for one-link-routes (everybody can do this now, even though persondriveragentimpl doesn't implement it).

advanced charging strategies

straightened control flow for handleTransit so that the QLink does not have to reach back into the QueueWithBuffer. Saves one interface method. Special treatment of one-link-routes (for TransitDrivers) into the TransitDriver (it will answer chooseNextLinkId with null now, consistent with other agents), even though the route is encoded startLink==endLink. Criterion if agent wants to end trip is now only chooseNextLinkId==null, for everybody. Corrected a not quite true comment.

simple analysis for virginia