MATSim Code

more useful failure messages

still fails... I can't understand why. Looking at the output events
file, the agents with 0-length trips really do not leave any link.
However, the count of leave link events has increased... Still
investigating.

adapt PseudoQSimEngine to new state of departure logic

add missing dependency

extracting WD listener

car- and optional pt-analysis separated , pt-driver excluded

current stage of ecab

do not handle 0-length car trips specially (MATSim-233)

this means that agents enter the vehicle, wait to the link and leave the
vehicle one second later. This should not change results much with
well-behaved scoring function (travel time passes from 0 to 1 sec,
assuming the vehicle is there and ready), makes the code shorter and is
more consistent (no need of the incorrect hack which triggered this
change anymore).

now three dilution methods

add the "standard" one (just prune the social network from anybody not
crossing the AOI

also proovide method for enums

attempts retaining less agents

retaining all "diluted" agents and their social contacts results in
keeping almost all the agents...
After dilution, attempt only adding alters of egos with leisure
activities, which themselves have leisure activities.