improve install scenario speed

DLR-TS / tsc

coupling TAPAS and SUMO

Eclipse Public License 2.0

10 stars 3 forks source link

improve install scenario speed #50

Open behrisch opened 1 year ago

behrisch commented 1 year ago

current stats for berlin_net:

net building: 3 minutes
taxi stops: 4 minutes
bidi taz: 22 minutes
gtfs2pt: 89 minutes
district mapping: 3 minutes per shape input
suburb mapping: 18 minutes

namdre commented 1 year ago

computeBidiDistricts.py might be made to run faster if getCandidates where to do vClass filtering. I imagine there are a lot of useless pedestrian paths within a 500m radius around any given edge that have an quadratic impact due to symmetrical=True.

behrisch commented 1 year ago

yes and no, the bidi taz are used for pedestrians as well but we might want to change that